research papers organizational learning

To read this content please select one of the options below:

Please note you do not have access to teaching notes, review of empirical research on leadership and organizational learning.

Journal of Knowledge Management

ISSN : 1367-3270

Article publication date: 29 May 2020

Issue publication date: 17 June 2020

This paper aims to investigate how the relationships between different leadership approaches and organizational learning have been examined in the literature, from which future research areas can be recommended.

Design/methodology/approach

This systematic literature review applies matrix method to examine major literature in leadership and organizational learning. A total of 57 peer-reviewed English publications from 45 journals were selected and analyzed.

The synthesis of these empirical studies revealed as follows: the relationship between leadership and organizational learning has been mostly quantitatively investigated in many countries and sectors; multiple leadership styles have been identified to ameliorate processes, levels and capabilities of organizational learning and transformational leadership still remains the most commonly used style; there are mediating mechanism and boundary conditions in the relationship between leadership and organizational learning.

Research limitations/implications

The literature search in this study was mainly focused on English articles only; therefore, some papers in other languages may have not been included.

Practical implications

This review offers an overall picture of the existing knowledge of organizational learning and leadership that will be fruitful for practitioners to understand and replicate these concepts.

Originality/value

There are little systematic literature reviews on the relationship between leadership and organizational learning. This paper is among the first systematic reviews to analyze how leadership has been associated with organizational learning and provide potential research directions.

• Organizational learning
• Leadership styles

Do, T.T. and Mai, N.K. (2020), "Review of empirical research on leadership and organizational learning", Journal of Knowledge Management , Vol. 24 No. 5, pp. 1201-1220. https://doi.org/10.1108/JKM-01-2020-0046

Emerald Publishing Limited

Copyright © 2020, Emerald Publishing Limited

Related articles

We’re listening — tell us what you think, something didn’t work….

Report bugs here

All feedback is valuable

Please share your general feedback

Join us on our journey

Platform update page.

Visit emeraldpublishing.com/platformupdate to discover the latest news and updates

Questions & More Information

Answers to the most commonly asked questions here

The effects of organizational learning culture and decentralization upon supply chain collaboration: analysis of covid-19 period

• Published: 01 September 2022
• Volume 16 , pages 511–530, ( 2023 )

Cite this article

• Alev Ozer Torgaloz 1 ,
• Mehmet Fatih Acar 2 &
• Cemil Kuzey   ORCID: orcid.org/0000-0003-0141-1744 3  

3982 Accesses

3 Citations

Explore all metrics

Crises cause delays in supply chain management with resulting changes to organizations’ internal structures. The COVID-19 pandemic has deeply affected the global supply chain and, with it, the organizational structure of companies. This research discusses supply chain collaboration (SCC) by considering two important organizational competencies: organizational learning culture (OLC) and decentralization (DC). It investigates the potential impact of these intangible resources upon SCC after the COVID-19 pandemic. The theoretical model was tested by variance-based Structural Equation Modeling (SEM) using results of a questionnaire which was completed by 245 respondents. In fact, this study explores which organizational capabilities determine the SCC level specifically within the current COVID-19 pandemic period. We believe that this contribution is significant, as the level of collaboration between companies can change during risk periods. The results show that OLC have significantly positive effects on SCC. Moreover, DC plays a critical role for the relationship between OLC and SCC. In other words, this study reveals the importance of DC to observe the positive effects of OLC on SCC. Unlike previous studies which explored SCC, this research demonstrates the importance of an organization’s inherent intangible resources in order to improve relationships with suppliers. The article ends with a discussion of the findings and their implications.

Similar content being viewed by others

Business Process Management and Supply Chain Collaboration: Critical Comparison of Four Thai Case Studies

Effect of Inter-Organizational Systems Use on Supply Chain Capabilities and Performance

Characteristics of Supply Chain Integration of Manufacturing Firms in Ethiopia

Avoid common mistakes on your manuscript.

1 Introduction

With a developing and increasing global trade environment in recent times, learning about organizational learning culture (OLC) and decentralization (DC), and subsequently applying them has become a major continuing necessity for organizations. In order to survive in such a business environment, companies must have an OLC (Egan et al. 2004 ). Organizational learning is the process of acquiring knowledge and, if well managed, it provides a competitive advantage to companies (Chang and Lee 2007 ). This can be achieved through observing developments within both the internal and external business environments. For this reason, organizational learning plays a critical role in helping organizations adapt to the changing environment comfortably. In today’s dynamic business world, organizations increasingly need to focus on learning faster and better than their competition in order to maintain their profitability and performance. As argued by Hosseini et al. ( 2020 ), culture indeed enables organizations to ensure sustainability. In this regard, the OLC can also be considered to be a guide for companies with which to achieve long-term financial profits. In other words, organizations with a learning culture are able not only to survive, but also to have greater financial profits (Eisenberg 2018 ). In addition, such a culture can assist organizations in creating advantages for their supply chain management by increasing flexibility and speed. Furthermore, a culture of learning improves the management capability of firms, leading to better business performance (Alegre et al. 2012 ; Andreou et al. 2016 ).

Organizations which survive are most often those that can transform themselves. As technology develops, so do the needs of customers. These changes reveal the necessity to continuously offer new products and services. Organizational learning, therefore, plays an important role at this point. Through integrating learning into their culture, companies are able to quickly meet the needs of the market by considering the expectations of their customers. As stated by Mirkamali et al. ( 2011 ), an OLC is crucial in order to offer new products and services to the market. Furthermore, digitization is a critical phenomenon that affects the operations and performance of businesses. It forms the basis of the fourth industrial revolution due to its enormous impact on the supply chain (World Economic Forum 2018; Truant et al., 2021 ). Digitization means the use of digital technologies to provide added value to an organization’s business model (Gartner 2018 ). In that point, the literature reveals that digitalization has a significant positive effect on learning processes and business continuity (Gupta et al. 2022 ; Ratchavieng and Srinet 2021 ). Therefore, organizations that want to have a learning culture should give importance to information management systems, as well.

Another factor that impacts organizational performance is structure. In this study, we consider DC to be a dimension of organizational structure. DC is chiefly defined as transferring decision-making and responsibilities to sub-workers or suppliers (Cullen and Perrewé 1981 ; Mehralizadeh 2005 ; Schiefelbein and Schiefelbein 2000 ). Many studies within the literature show the effect on performance arising from DC. Accordingly, under certain circumstances, central organizations are found to be more effective, whereas in other cases, DC can provide more successful results. (Nasirpour et al. 2010 ; Gaber 2003 ; Wang 2010 ; Andrews et al. 2007 ). It may be considered effective to mix these two structural systems for better management under certain circumstances (Hawkins 2000 ).

Organizations have experienced varying consequences in response to the COVID-19 pandemic, which began in December 2019. Within this context, companies have been forced to close for a certain period of time, working hours have changed, and many employees have been working remotely. This situation affected the control of all operations, making it much more difficult to organize efficiently. It is a matter of curiosity as to whether centralized companies will maintain their characteristics under the changed working conditions after the COVID-19 pandemic ends.

In recent years, the classical concept of supply chain management (SCM) has expanded, while SCC has been discussed frequently, both in business life and in academia (Ahmed et al.,  2020 ; Li and Chen 2019 ; Chiang et al. 2018 ). It is a strategic issue for organizations wishing to have a sustainable performance. SCC is a partnership process established by at least two independent organizations in order to carry out operations more effectively, and to ensure their mutual benefit (Cao and Zhang 2011 ; Parmigiani and Mitchell 2010 ). This cooperation can include not only firms, but also governments, universities, institutes, and R&D organizations (Lee et al. 2010 ; Mirc et al. 2017 ; Al-Tabbaa and Ankrah 2019 ). Meanwhile, different collaborations have been established in later years within the supply chain field, focusing upon green production (Ağan et al., 2016 ). Interest and awareness in SCC has increased critically; companies have found advantages such as sharing information comfortably and providing competitive advantages, while reducing overhead costs and inventory. (Soylu et al. 2006 ). The ultimate goal of SCC is to gain a competitive advantage within the related market (Soylu et al. 2006 ; Chen et al. 2017 ; Cao and Zhang 2011 ).

Many studies have discussed the SCC in the literature, however few studies have emphasized the importance of OLC and DC for supply chain operations (Chang and Lee 2007 ; Nasirpour et al. 2010 ). To fill this gap in the literature, we construct a theoretical model that the importance of OLC and DC for SCC. While establishing this structure, we consider the Resource Based View (RBV), which claims there are benefits for effective and joint usage of organizational resources, as a focal point. Strategic decisions are “the decisions that critically affect organizational well-being and survival” (Eisenhardt and Zbaracki 1992 ). Based on the RBV, organizations build competitive advantage through using the available resources strategically (Penrose 1959 ). The RBV accepts both tangible and intangible assets of an organization as resources contributing to creation of a competitive advantage (Wernerfelt 1984 ). While the view mainly assumes that organizations can develop a competitive advantage by creating strategic resources and capabilities (Dubey et al. 2018 ), the relevance of these resources varies depending on the conditions of such dynamic ecosystems, such as crises.

Having reviewed the related discussion, this study examines the effect of such organizational characteristics as OLC and DC on SCC. The main questions raised within this research are the following; (i) What is the relationship between SCC and the suggested organizational characteristics? (ii) Do these relationships change if experienced within small and medium sized enterprises (SMEs) or large firms? (iii) Do these relationships differ when the companies in a country are local or foreign? In this study, research questions were posed through cross-sectional questionnaires. The survey items were adapted from the literature. The managers of 245 businesses responded, which provided a comprehensive understanding of their companies’ inner workings and supply chains. The theoretical model for this research was tested with variance based structural equation modeling (SEM). With this study, our contribution to the literature will demonstrate the importance of OLC and DC for SCC and identify the precursors of SCC during the COVID-19 pandemic. This research provides these findings empirically with statistical analyses. We believe these results will inspire shareholders and managers to develop effective collaboration strategies in order to prevent further supply chain disruptions in the COVID-19 pandemic situation, or within similar risk environments.

The rest of the paper is organized as follows. Section 2 provides an extensive literature review on the studies regarding OLC, DC and SCC, with the theoretical development of our hypotheses building upon the research model. Sections 3 and 4 present background information about the methodology used and illustrate the results, respectively. The findings are discussed in Sect. 5. Finally, Sect. 6 provides concluding remarks and possible future research directions.

2 Literature review

Ensuring a learning culture that leads to performance improvement has become vital for the ability to compete in today’s changing business environment (Lau et al. 2019 ). This study provides evidence for the suggested relationships among OLC, DC and SCC, with explanations leading to the ways in which organization-level concepts and qualifications can be related to supply chain characteristics.

OLC is suggested in this study as providing the main infrastructure for an organization to establish cooperative behavior with its supply chain partners. Accordingly, having established a learning culture can help organizations to ensure a framework that facilitates exchanges of necessary information, adaptation strategies and supportive behavior, as learning culture makes organizations creative and problem-solving (Ojha et al. 2018 ). As organizational culture has a significant role in shaping the structure and decision-making processes, this study assumes that learning culture has an impact on the DC of the organization and it is DC that explains a further association between OLC and SCC. Accordingly, a learning culture can help the organization to become flexible, which contributes to the empowerment and delegation by the management. This can eventually add to the openness to the environment and therefore to become likely to collaborate with the inter-organizational partners.

Through this approach, the aim of this paper is to demonstrate that a collaborating supply chain, or indeed, much of the nature of any sub-system within an organization, has roots in organization-wide issues such as culture and decision making, which are learning culture and DC in this instance. To this end, this literature review section briefly introduces and discusses each construct within the model, followed by the hypotheses development section that provides explanations concerning the relevance of examining relationships among the constructs.

2.1 Organizational learning culture (OLC)

Organizational culture is about the norms and values established within an organization and helps to shape both employees’ and organization’s behavior. There are various types of organizational culture widely discussed within the organization literature, mainly as clan, adhocracy, hierarchy and market culture, according to the framework developed by Quinn and Rohrbaugh ( 1983 ). In addition, several other dimensions can be added to this classification, depending on which values an organization adopts. This means, on what basis an organization aims to shape its behavior can be converted into a culture in a long-run. For example, an organization can establish a culture out of the bureaucratization and adapt a bureaucratic culture. In this study, learning culture has been suggested, while aiming to explain SCC via a decentralized structure.

Organizational culture can be mainly defined as a pattern of basic assumptions developed while learning to cope with the problems of external adaptation (Schein 1992 ). Therefore, this study has integrated organizational culture as the main antecedent of SCC, which is also maintained with the aim to adapt to the supply chain environment in better terms. The OLC has been specifically chosen as the triggering construct of the model that contributes to SCC because of has two aspects. First, organizational culture facilitates learning for supply chain parties by developing relationships that help to exchange knowledge and experiences (Zhao et al. 2011 ). Secondly, the specific learning culture is identified from among other types of organizational cultures, as it is critical towards boosting the organizational performance (Arefin et al. 2020 ).

Watkins and Marsick ( 1993 ) referred to a learning organization as it being transformed continuously and enabling employee involvement in a collective manner. Furthermore, learning organization is considered to be the ability to adjust in response to new realities when demanded by change in the work environment Gephart et al. ( 1996 ). Combined, an approach that adapts learning behavior and habits as a culture becomes critical for organizations. This points to the significance of the learning culture and investigating its effects upon periods of change, such as in the case of the current pandemic. Škerlavaj et al. ( 2007 ) identified learning culture as a set of values supporting systematic approaches in order to achieve higher-level learning through information acquisition, interpreting this information, and instituting the resulting required behavioral changes. A learning culture encourages an organization and its employees to add to their knowledge and performance regularly (Chanani and Wibowo 2019 ). Specifically, according to Dodgson ( 1993 ), being a learning organization has become critical, especially for larger organizations, as they attempt to develop systems which are adaptable to change. This study asserts that learning culture is a turning point that creates awareness with regard to the need for change, and thus developing a capability for collaboration with suppliers. Organizational culture has been identified as a unique organizational resource within the literature (e.g. Zahra et al. 2004 ; Melville et al. 2004 ; Özçelik et al. 2016 ; Ahmed et al. 2018 ). Moreover, Tynjälä ( 2013 ) argued that the learning culture is vital in enabling organizations to remain competitive. Teece ( 2015 ) further regarded it as a source of competitive advantage. In their study, Jin and Hong ( 2007 ) considered SCC with a reference to organizational culture, yet via the six dimensions that describing the patterns of culture as suggested by Hofstede (1994). However, so far, no study has focused on the effect of the learning culture on the organization’s supply chain management, which this study considers.

In this study, with respect to a resource-based view, the OLC is considered to be a strategic resource being valuable, non-substitutable and unique to each organization, and which contributes to its competitive advantage. In this regard, this study maintains that an OLC is an important intangible resource for organizations.

2.2 Supply chain collaboration (SCC)

The business environment of today dictates a high level of integration and collaboration between processes, especially for manufacturing (Fatorachian and Kazemi 2018 ). More specifically, SCC has become a critical common norm for many organizations (Ramanathan and Gunasekaran 2014 ). The importance of studying collaboration is that it is highly related to several positive outcomes such as sharing risks (Parkhe 1993 ) and the general performance (Hewett and Bearden 2001 ) of organizations. According to Kanter ( 1994 ), ‘collaboration’ carries the implication of ‘creating value together’. Corsten and Felde ( 2005 ) identified collaboration as a joint effort within buyer-supplier relationships. Within this collaboration are included coordination, knowledge transfer and strategic alliance. Many organizations have realized the importance of external connections and have tried to become involved in networks where resources and information are mutually interchanged (Teece 1992 ). Collaboration provides a basis upon which to produce improvements via partners and can assist in reconsidering current processes (Dodgson 1993 ). Supply chain managers look externally to meet customer needs, to share risks, capitalize on partner expertise, and even to handle market operations (Ralston et al. 2017 ); this is the main explanation for the operation of SCC.

Qrunfleh and Tarafdar ( 2014 ) argued that there is a positive relationship connecting SCC with an organization’s performance. Specifically, during the current pandemic, supply chains have been largely destroyed (Tanner 2021 ). As organizations focus on recovery and reconstruction, specifically during a disaster, the relationship between organizations is critical (Altay et al. 2018 ). Therefore, we believe that it is timely to examine SCC in an exploratory way. So far, SCC has been associated mostly with trust (Walter 2003 ), commitment (Chen et al. 2011 ), dependence (Fynes et al. 2005 ), strategy (Angerhofer and Angelides 2006 ), technology (Lee et al. 2011 ) and information sharing (Cai et al. 2010 ) with a view of factors affecting SCC. However, this study focuses on the organizational resources contributing to SCC. With regard to the resource-based view, this study considers a positive outcome; collaboration, which is suggested to be improved with the help of resources as a learning culture and DC and through which awareness and understanding could have changed within the impact of the pandemic. We have focused upon collaborations where the objective is to share information and provide support when necessary, as has been evidenced in the case of this crisis. Analyses of this study have been conducted with respect to the pandemic period. As will be discussed in the following sections, the pandemic period is contended as being a critical turning point for the awareness of organizations with respect to their supply chain management and their relations within this focus.

2.3 Decentralization (DC)

New business trends such as working remotely can be assumed to have increased the importance of decentralized structures after COVID-19. There is extant literature support for the relevance of organizational structure and decision-making norms to organizational performance. Aiken and Hage ( 1966 ) defined DC as the extent to which freedom is given to members of an organization within which to perform their tasks without interruption by their supervisors. Similarly, Hempel et al. ( 2012 ) stated that DC concerns the dispersion of decision-making within an organization through giving employees greater autonomy as well as increasing the flow of information. Kochen and Deutsch ( 1980 ) considered that the discretionary power within DC would include not only decision making, but also other functions such as allocation, coordination and control. Vancil ( 1979 ) referred to a decentralized organization as one in which sub-managers have the responsibility for the performance of a subunit. Therefore, it can be stated that the shift of discretion and responsibility is transferred from accountability for the whole organization to accountability for smaller subunits (Hales 1999 ). This study considers DC consisting of being whether decisions within an organization are made by employees as teams in a cooperative manner, or simply by single authorities. In this regard, decisions on supply chain issues play a key role.

Within the related literature, DC has been considered mostly as ‘decentralized supply chains’ (e.g. Chen 1999 ; Fan et al. 2003 ; Gatignon et al. 2010 ). However, this study regards DC from the point of organizational structure with regard to decision making procedures and authority-delegation policies. DC can be seen from the point of the resource-based view, since it can facilitate specific situations when organizations have scarce resources, with respect to the management by exception (Mahmood et al. 2014 ). Accordingly, being flexible and delegating decision-making to subordinates can be pragmatic when, for example, the resource of time is tight, or information is limited. In such circumstances, the ability to operate flexibly in decision-making can ensure that an organization performs successfully. Similar observations can be made during difficult and/or challenging times. In this regard, this study considers DC to be an intangible resource that contributes to organizational competitiveness.

3 Theoretical background and hypotheses development

3.1 resource-based view.

This section describes the related theoretical framework which was applied in this study in order to explore the suggested associations. Similar to a study on supply chains by Dubey et al. ( 2018 ), this study draws upon a current theoretical framework that is based upon the resource-based view (RBV) and the relational view. Penrose ( 1959 ) described the resource-based view as gaining organizational character from a company’s own resources. The resource-based view argued in the main that organizations can improve their competitive advantage through creating strategic resources and capabilities (Dubey et al. 2018 ). With regard to this view, Wernerfelt ( 1984 ) stated that the intangible assets of an organization can also be considered to be resources that contribute to competitive advantage. This study regards the OLC and decentralized structure to be competitive intangible assets which help to develop an advantage. Barney ( 1986 ) accepted that organizational culture is a source of competitive advantage. In addition, Fiol ( 1991 ) regarded organizational culture as a competitive resource, underlining that several studies have regarded organizational culture as a means with which to achieve a desired organizational result. In a study by Clulow et al. ( 2003 ), it was revealed that intangible assets such as organizational culture are in fact critical resources for organizations. Moreover, a resource-based view has been largely incorporated within studies on organizational performance (Innocent 2015 ). In this study, SC is considered to be the desired organizational outcome, and that it contributes to organizational success with resulting survival.

Within the literature, there are several theories used to explain SCC, as relational view, socail exchange theory, transaction cost theory, agency theory, etc. (Soosay and Hyland 2015 ). However, all these theories mostly focus on the relationship between the organization and its partners; whereas this study considers organizational factors that may affect suppy chain collaboration. The reason why resource-based view was applied in this study is that both a decentralized organizational structure and a culture of learning are suggested as strategic assets of the organization contributing to competitive advantage. The resource-based view focuses on organizational internal resources, that are rare, imitable, valuable and nonsubstitutable (Barney 1991 ). Both organizational culture and structure have been identified by Brynjolfsson et al. ( 2002 ) as intangible assets for organizations. Besides, Moran and Meso ( 2008 ) regarded organizational culture as a strategic asset. Moreover, learning itself is also considered as an intangible strategic resource (Hult et al. 2003 ), and specifically for supply chains (Biotto et al. 2012 ; Willis et al. 2016 ). Organizations can use their cultural motives strategically to affect their external environments (Weber 2005 ). With respect to the DC, the resource-based view can help to reframe the relationship between strategy and organizational structure with a better integration (Moingeon et al. 1998 ). Considering the outcome of this study, SCC, Defee and Fugate ( 2010 ) argued that learning orientation of supply chain partners has a significant effect on the interactions between the partners.

Informed by the resource-based view, we explored how OLC, as a resource, developes SC, which, in fact, can be considered as a positive competitive advantage in times of change. Developing collaboration with a supplier can be seen to be within the framework of relationship management and is explained by way of the resource-based view.

3.2 Hypotheses Development

The structure (e.g. centralization vs. DC) and the decision-making patterns of organizations intertwine with their learning cultures. There are many studies within the literature (e.g. Janićijević 2013 ) that link culture and structure. Cullen and Perrewé ( 1981 ) defined DC as transferring decision-making and management responsibilities to sub-workers or suppliers. Centralization or DC are not the means to an end but contain the means to achieve the desired outcomes for organizations. Centralization, under specific circumstances, increases efficiency, yet in some other cases, DC will yield better results (Alhamad and Aladwan 2019 ). At certain times, a mixture of the two structural forms may be blended for better management (Hawkins 2000 ). Learning organizations have a tendency to adopt decentralized structures, since centralization hinders learning (Burns and Stalker 1961 ). In fact, the shared values that result from the culture guide employees in performing their strategic roles and responsibilities (Schilke and Cook 2015 ). Furthermore, post-bureaucratic organizations tend to have a learning culture which ensures flexibility and consensus, rather than hierarchy and authority (Lee and Edmondson 2017 ). On the other hand, there is evidence that OLC strengthens the competencies of employees (Potnuru et al. 2019 ), which in turn assists them to be prone to make decisions through assuming extra initiative. Similarly, according to Choi ( 2020 ), employees feel empowered through the help of an OLC. In accordance with the related research, the first hypothesis of this study is:

H1: There is a significant and positive association between OLC and DC

The decentralized processes of decision-making and enhanced communication strengthen an organization’s ability to respond quickly to changing conditions (Zammuto and O’Connor 1992 ). DC mainly functions as a way with which to respond to change effectively and contributes to organizational flexibility, critical to fast response to market conditions (Hill et al. 2000 ). It allows organizations to incorporate the capabilities of lower-level employees, whose contributions are most often neglected by more centralized decision-makers (Richardson et al. 2002 ). The structural characteristics of organizations significantly affect their effectiveness in ways such as adaptability (Ranson et al. 1980 ). The influence of mid-level managers, which is considered to be a sign of DC, affects the quality of decisions and also overall performance (Richardson et al. 2002 ).

There are abundant studies in the literature linking the effects of both centralization and DC to organizational performance (Nasirpour et al. 2010 ; Andrews et al. 2007 ). Centralization limits responsiveness (Argyris and Schön 1978 ). In fact, for an innovative organization, a structure-enabling information flow is required. (Chesbrough 2003 ). Decentralized structures allow organizations to be flexible in regard to their external environment (Ogbonna and Harris 2000 ). A decentralized structure facilitates a participative environment for spontaneous knowledge building (Hopper 1990 ). Better communication helps supply chain partners to effectively coordinate their participation (Zhang and Cao 2018 ), and this better communication is enabled by way of decentralized structures, within which information can easily flow. In decentralized structures, vertical communication is unnecessary, enabling decisions to be made by the most competent employees (Ellis et al. 2011 ), which increases risk awareness within organizations. The results of a study by Kandemir et al. ( 2002 ) revealed that DC contributes to organizational alliance orientation, which in turn leads to relationship commitment. This can also be valid with regard to suppliers and ensuing intentions to collaborate with them. Combined, this study contends that making decisions in a decentralized fashion would assist organizations to collaborate with their suppliers. Accordingly, the second hypothesis of this study is:

H2: There is a significant and positive association between DC and SCC

Lee et al. ( 2012 ) argued that to handle global competition within business environments, organizations must improve their learning approach as to how this can contribute to business opportunites. In fact, managing the supply chain effectively can be considered as a way to use these business opportunities. The learning motive within an interorganizational setting was defined by Hamel ( 1991 ) as a willingness to learn from all of the parties in a relationship. Senge ( 1992 ) identified a learning organization by pointing to cooperative learning within the institution. Similarly, learning processes suggested by Marks and Louis ( 1999 ) as were shown to be collaborative in nature. As suppliers are important factors within the environment of an organization, they can be included within these two definitions of organizational learning. At this point, learning together with a supplier can be considered, which can be the source of relevant information. In this regard, it would be critical to refer to ‘relationship learning’, in which information is shared mutually, jointly defined, then integrated into a shared memory (Selnes and Sallis 2003 ). In fact, organizational culture itself can encourage collaboration within partnerships (Gopal and Gosain 2010 ). An organizational culture with an external orientation can be considered to be the major driver in a successful collaboration (Mamillo 2015 ). Here, culture with an external orientation is defined by Cameron and Quinn ( 2011 ) as interacting with the actors beyond their boundaries; this type of culture can then be regarded as the learning culture. Kandemir et al. ( 2002 ) pointed out that learning culture facilitates an organizational orientation for alliances, which we can consider to be collaboration with suppiers in our case, since they also define it similarly. A study by Laaksonen et al. ( 2008 ) revealed that the learning of capabilities is critical, since effective interorganizational relationships develop evolutionarily over time. An organization which is accustomed to acquiring and sharing information, largely encouraged by a learning culture, would be expected to establish collaboration with its supplier. It can be argued that this point was revealed in a study by Islam et al. ( 2015 ), where learning orientation was shown to have a positive relationship with knowledge-sharing. In this regard, knowledge-sharing can be considered as referring to collaboration with suppliers. Being informed by this discussion, this study maintains that OLC aids the collaboration of an organization with it’s suppliers. Therefore, the third hypothesis of this study is:

H3: There is a significant and positive association between the OLC and SCC

This study aims to provide an explanation for the suggested association between the learning culture and collaboration with suppliers. We suggest that structural motives and decision-making styles can assist in connecting this culture with the resulting outcome. Accordingly, organizations with learning cultures would tend to establish more centralized structures, due to the nature of information sharing. This decentralized structure, in return, would provide a basis for organizations to have better cooperation within the organization first, followed by cooperating with other parties.

Similarly, in their research model, Lee et al. ( 2012 ) identified the three suggested variables as learning culture, DC and collaboration, providing explanations for associations among them. These three constructs are suggested in our study as being related in the sense that a decentralized structure provides a bridge between organizational learning adaption with the intention to collaborate with supplying parties. We believe that achieving SCC requires an organization-wide understanding of sharing and openness which can be maintained by the existence of a culture and decision-making structure. Accordingly, a learning culture can provide an open environment of inquiry and sharing, which can result in a decentralized approach to decision-making when employees are trained to learn, share, and accumulate experience. In this manner, an inner environment of sharing and cooperation can be reflected to the outer environment, which includes supply chain parties, allowing the organization in general to establish a relationship of collaboration within its supply chain. Having reviewed the relevant literature so far, no explanation has yet been found for an assumed relationship between learning culture and collaboration with suppliers. Sparked by this gap, this study contends the final hypothesis as:

H4: DC mediates the relationship between the OLC and SCC

In this study we have explored the importance of an OLC and DC upon SCC, but we tend to differ regarding analysis of the effects of organizational learning upon SCC directly, as well as taking into account the mediating effects of DC. Moreover, this research tests the strength of these relationships within the COVID-19 pandemic, being a major crisis time forcing organizations to increasingly focus upon their competitive abilities. We aim to contribute to the literature by providing empirical evidence regarding the effects of organizational learning and DC upon SCC within the circumstances of a pandemic. This contribution is important, because the level of collaboration between companies often changes during risky periods. The goal of this research is to show some of the changes evolving in relationships that have occurred between OLC, DC, and SCC during the COVID-19 pandemic.

4 Research methodology

A detailed description of the approaches that were employed for this study is discussed within this section. Diverse methodologies were performed in order to justify the selection of the models as well as to test the proposed research hypothesis. The research methodology incorporates various analysis approaches. First, the descriptive statistics of the items, sample distribution, data screening, and data collection are presented. Next, the description of the scales, Partial Least Square (PLS), factor loadings, and measurement models are examined. Then, multicollinearity analysis, confirmatory factor analysis, and measurement models are examined. Also, Structural Equation Modeling along with the mediation analysis, multi-group analysis, and comparison tests are performed.

4.1 Partial least squares structural equation modeling (PLS-SEM)

Various advantages of using the PLS-SEM approach have been reported in the literature. First, there are significant advantages over ordinary Covariance-Based Structural Equation Modeling (CB-SEM), when there is a relatively small sample size and accompanying restrictive assumptions. The PLS-SEM methodology was selected for this research over CB-SEM, since it has been highly recommended for a research study with either a small sample, non-normal data, or formative measures (Marcoulides and Saunders 2006 ; Ringle et al. 2012 ). Moreover, it is considered to be a soft-modeling approach because of its aspect of less rigid distributional assumptions about the research sample. Therefore, PLS-SEM was utilized as an alternative approach to the CB-SEM. As well, PLS-SEM was seen as a complementary tool for the CB-SEM since they compete against each other (Chin and Newsted 1999 ). Considering all the advantages, PLS-SEM is very robust to deviations from normality based on the Monte Carlo Simulation results (Cassel et al. 1999 ).

4.2 Proposed model

The proposed research model is shown in Fig.  1 . Direct as well as indirect paths are indicated in the model illustration. The blue full lines represent the direct paths, while the red dotted line indicates the indirect path in the illustration. The four proposed hypotheses are also noted on the model, together with firm size as well as the ownership, as the control variables. The control variable size has two categories within small and large firms. Small firms are designated as having less than 250 employees, while the large firms are described as having 250 or more employees. Finally, the control variable ownership incorporates two categories: domestic and foreign firms.

Proposed model

(H1, H2, and H3 indicate the direct paths while H4 indicates the mediation role of DC between OLC and SC)

Before the survey questions were distributed, the requisite information relative to the research project was provided to the participants, and the objective of the research was thoroughly explained. The participants were assured that there would be no right or wrong answer at the same time that they were also assured that their answers will be kept anonymous. These initial steps of providing the required information and informing the participants about the objective of the study helped to minimize any anxiety before responding to the survey. As a result of the preliminary results, 245 out of the 402 possible participants agreed to join the research, which yielded a response rate with a ratio of 61%. The survey was collected in early 2021, when the COVID-19 pandemic was at its peak worldwide. Before performing further analysis, the raw data research was subjected to the data preprocessing phase, which is a crucial step before testing the hypothesis (Hair et al. 2019 ).

This study depends on a unique dataset collected through a cross-sectional surveys. Based on the extant literature, scales of SCC, OLC, and DC are used as constructs. The revised survey was translated into Turkish. This translated questionnaire was then sent to a random sample of medium- and large-sized companies. In order to have homogeneous data, only manufacturing companies within Turkey were considered for this research. Approximately 1,000 firms were potentially identified for our survey. Professional online networks were employed to contact the managers of these companies. The respondents were located in different but related departments, such as supply chain, procurement, marketing, etc. A total of 245 complete questionnaires were actually returned. The raw dataset was subject to further data cleaning processes. The initial descriptive statistics revealed that there were no missing values in the raw data. As well, the univariate, as well as multivariate outlier detections, were performed; the results indicated that there were no significant outliers in the research sample. In the last stage, the final sample size available for further analysis was 245 records.

4.4 Descriptive statistics

There were two demographic variables in the research sample which were also used as control variables in the further section. The frequency analysis results are shown in Table  1 , which showed that 75.92% of the participants came from domestic firms while only 24.08% were from foreign firms. The firm size had two categories: small and large. 44.9% of the participants were from small firms, while 55.1% were from large firms. Moreover, the results indicated that 33.06% of the participants were purchasing managers, while 4.9% held quality control officer positions in their firms. Finally, the results revealed that 22.45% of the participant’s firms were in the service sector, 17.14% were in the metal sector, 12.25% were in the food and agriculture sector, 8.57% were in the automotive sector, and 6.53% were in the textile sector.

The summary of research items based upon descriptive statistics is provided in Table  2 . The total sample size was 24. The items of the OLC ranged between one and six, DC ranged between one and seven, and SC ranged between one and five. The mean values, as well as the standard deviations of the items, are provided within the results, which demonstrate that there is no significant variation around mean values.

The construct items in this research paper were modified/adapted arising from prior studies in the literature. The items were based upon multi-item scales; the original language of the items was in English. Before sending the survey questions to the participants, the items were subject to purification steps incorporating back-translation methodology (Brislin 1970 , 1986 ): the questions were first translated from English to Turkish by a bilingual expert in this field, then to ensure the content quality of the questions, the questions were translated back from Turkish to English by another bilingual expert in the field. Therefore, the accuracy of the translations was ensured through joint translation in order to alleviate possible contradictions.

The survey’s items incorporated the Likert scale of measurements, which was developed for each construct and are highly recommended within the related literature. The 7-item measurement for OLC used for this study is the scale developed by Yang ( 2003 ). This scale is a short version of the one developed by Marsick and Watkins ( 2003 ), and can be employed separately, thereby creating a single measurement of a learning culture. O LC was measured with the 7-point Likert scale, ranging from 1 (strongly disagree) to 7 (strongly agree). For DC, we used the 4-item scale developed by Zahra et al. ( 2004 ). The scale is 7 -points Likert, ranging from 1 (strongly disagree) to 7 (strongly agree) . Finally, we measured SC with the 4-item collaboration scale developed by Carr and Pearson ( 1999 ). The scale is 5-point Likert, where 1 means not at all and 5 means to a great extent.

The sample consisted of both Turkish and multinational corporations that operate within Turkey. We controlled for ownership type, measuring it with a categorical variable, where 0 represented a domestic company and 1 denoted a multinational one. Furthermore, we controlled for size then measured it again with a categorical variable, where 0 represented a small company and 1 denoted a large company. With regard to size, companies with less than 250 employees were categorized as small, while companies with over 250 employees were categorized as large.

4.5 Exploratory and confirmatory analysis

In this section, the PLS-based factor analysis was performed in order to investigate the factor loadings of the survey items. There were 15 items in the constructs: seven from OLC, four from DC, and four from Supplier Collaboration (SC). These indicated items were subject to the PLS-based factor analysis. No items were eliminated following the initial analysis, since the factor loadings of the corresponding items were relatively high. The results of the factor loadings are shown in Table  3 . Chin ( 1998 ) suggested 0.7 for the PLS-factor loadings. The results showed that the factor loadings of the items were near to or greater than 0.7, with OLC1 being 0.69 and SC4 being 0.67, both of which were very close to 0.70. Furthermore, according to Hair et al. ( 2019 ), the minimum factor loadings should be 0.35 when the sample size is 250, while the minimum factor loading should be 0.40 when the sample size is 200. Considering that the sample size of this research paper was 245, the minimum factor loading of constructs was 0.67, significantly greater than the suggested minimum value of 0.35. Thus, the factor loadings were satisfactory and could be used for further analysis.

As well, the discriminant validity was satisfied, since each item had a higher factor loading on its construct compared to the other constructs while each item also had a higher loading than the cross-loading on the row level as well as in the column level.

Regarding the reliability measures, the Cronbach’s alpha of OLC was 0.9, DC was 0.84, and SC was 0.77, all of which are greater than the suggested value of 0.70. Also, the composite reliability measures were 0.92 for OLC, 0.89 for DC, and 0.85 for SC, which are all greater than the suggested value of 0.70. The results of Cronbach’s alpha and composite reliability indicated that the reliability of the constructs was satisfied, since the values of the reliability measures were all significantly greater than the suggested threshold value of 0.70.

Before performing baseline analysis to test the research hypothesis, the question of multicollinearity between the independent latent variables was investigated. Variable Inflation Factors (VIF) were used to examine the multicollinearity issue. It was seen that OLC and DC were the independent variables in the model. The VIF’s results showed that the value of VIF of OLC was 3.05 and VIF of DC was 3.05, both of which were significantly less than the recommended threshold value of 10 (Hair et al. 2019 ). Therefore, there were no multicollinearity issues between the independent variables.

The confirmatory factor analysis (CFA) was applied to the research model, with three constructs following the factor analysis. The maximum likelihood (ML) method was selected during CFA. The construct’s validity was checked by using CFA. The standardized regression weights, t-statistics, p-value, and the fit measures of the model are provided in Table  4 . Accordingly, the results of the goodness of fit measures were \({{\chi }^{2}}_{(df=81)}=160.69, p<0.001\) ; \({\chi }^{2}/\) df= 1.98; the goodness of fit index was (GFI)= 0.92; the normed fit index was (NFI)= 0.93; the comparative fit index was (CFI)= 0.96; the relative fit index was (RFI) = 0.91; the incremental fit index was (IFI)=0.96; the Tucker-Lewis index was (TLI)=0.95; and the root mean square error of approximation was (RMSEA)= 0.055. The results of the fit measures revealed that the model fit indices were met. Therefore, sufficient evidence of good model fit was obtained (Hu and Bentler 1999 ). Furthermore, the standardized regression weights of the items were all statistically significant (p-value<0.001). Finally, the results showed that the convergent validity was satisfied, since the items are statistically significantly loaded to their respective constructs.

The relationship between a latent variable and its corresponding indicators can either be in the form of reflective or formative in the PLS-based SEM methodology. For this research model, a reflective modeling approach was chosen, since it has significant advantages over the formative modeling approach. The advantages of the reflective approach are: (i) variation in the construct causes variation in the item measure; (ii) dropping an item from the model does not affect the construct, and (iii) the indicators are correlated with the latent variable at a high level (Podsakoff et al. 2003 ).

Before testing the research hypotheses, the construct reliability and discriminant validity were examined. The results of the item-based reliability are given in Table  5 . The measurement model analysis included descriptive statistics, average variance extracted values, composite reliability, correlation coefficients, and the square root of average variance extracted value at the diagonal of the matrix of each construct. The obtained statistics investigated the discriminant validity, reliability, and convergent validity of the constructs (Hair et al. 2019 ).

The factor loadings (Table  3 ) showed that the individual item reliability was satisfied, since the factor loadings of all the included items were significantly greater than the suggested cut-off value (Chin 1998 ; Hair et al. 2019 ). The values of composite reliability were 0.922 (OLC), 0.894 (DC), and 0.854 (SC), significantly higher than the recommended cut-off value of 0.70 (Nunnally, 1987). Moreover, the values of Cronbach’s Alpha were 0.900 (OLC), 0.841 (DC), and 0.766 (SC), also higher than the threshold value of 0.70. Therefore, the constructs’ reliability was met.

Additionally, two important statistics were employed to examine the discriminant validity: (i) the Average Variance Extracted (AVE) and (ii) the square root of AVE values versus the correlation coefficients of the latent variables (Fornell and Larcker 1981 ). The results showed that the AVE values were higher than the recommended cut-off value of 0.50. Moreover, all the square roots of AVE values on the diagonal of the correlation matrix were higher than the correlation coefficients of constructs, when placed off-diagonal. As a result of these two important statistics, the proposed latent variables in the model were separated or different from the rest.

In summary, the measurement model with related analysis results showed that the individual item reliability measures were satisfied. Therefore, the PLS-based SEM analysis could be performed in the following section in order to test the proposed hypotheses.

4.5.1 Common method, Social Desirability, and response biases

We have addressed the issue of common method bias by using two crucial approaches. First, we examined the variance inflation factors (VIF): both VIF values of DC and OLC were 3.05, which is relatively lower than the suggested threshold value of 3.3 (Kock 2015 ). Therefore, the proposed research models were not negatively affected by the common method bias since the VIF values were less than the suggested value of 3.3. Second, we utilized Harman’s Single Factor approach (Harman 1960 ). Without a rotation, we included all the items of the survey into a single factor using the Exploratory Factor Analysis method. Following this analysis, we obtained 34.9% of variance, explained by the new common latent factor, which is significantly less than the recommended cut-off value of 50% (Harman 1960 ).

With the use of the method described by Armstrong and Overton ( 1977 ), the non-response bias was tested. The early respondents to the surveys were compared with the latter ones. Firstly, the comparison of the responses between the early and late respondents showed no statistically significant differences (p > 0.05). Secondly, when cross-checking a randomly selected group of 100 non-respondent firms with respondent firms, there were no significant differences identified for any organizational portion (such as number of employees, years of operation, etc.). Lastly, according to Rose et al. ( 2007 ) a relatively high response rate can prevent a non-response bias. In this research, the return rate was approximately 24.5%. It is thought that this ratio has provided an adequate representation for the total sample. The literature says that when several features of an object are defined at the same time, participants feel less pressure to defend the relevant one. The knowledge that each answer may be entered accurately prevents a social desirability bias (Tomassetti et al. 2016 ). Therefore, in advance of the survey, information was conveyed that there could be no right or wrong answers. In addition, there was no question about the identity of the participants.

4.6 Structural equation modeling-hypothesis testing

PLS-based SEM was employed in order to test the proposed hypotheses. The justification for using PLS-SEM is that the sample size of the research project was relatively small, so that there are significant advantages in terms of least restrictive assumption-free features (Chin 1998 ). Accordingly, Smart PLS v.2 with bootstrapping resampling was selected to test the statistical significance of the relationship between the latent variables (Ringle et al. 2005 ). Furthermore, bootstrapping with a recommended 5000 resamples was used in order to obtain the standard errors and t-statistics for the path coefficients during the analysis (Henseler et al. 2009 ). The path analysis results are presented in Table  6 .

The coefficient of determination ( R 2 ) shows the percentage of the explained variance in the dependent variable which can be explained by the independent variables. Namely, the values of R 2 were used to examine the explanatory power of a structural equation model. The threshold values of the coefficient of determination were 0.67-Substantial. 0.33-Moderate, and 0.19 - Weak (Höck and Ringle 2006 ). The results from PLS-SEM indicated that 67.66% of the variation in DC could be explained by the variation in OLC, while 59.9% of the variation in SC could be explained by the variation in OLC and DC.

The results of the path analysis showed a significant positive relationship between OLC and SC (β = 0.424, p < 0.01), between OLC and DC (β = 0.823, p < 0.01), and between DC and SC (β = 0.168, p < 0.05). Therefore, H1, H2, and H3 were supported.

4.6.1 Mediation analysis

Further analysis with mediation methodology was employed in order to investigate the mediating role of DC upon the relationship between OLC and supplier collaboration (SC). Baron and Kenny’s ( 1986 ) four steps as well as Hayes’s ( 2017 ) methodologies were followed for this analysis.

First, the indirect effects of OLC on SC via DC were tested by applying Hayes’ ( 2017 ) methodology with the module PROCESS 3.5 (Hayes 2019 ), which incorporates a bias-corrected bootstrapping approach in order to determine whether the indirect effect was significant. Preacher and Hayes ( 2008 ) suggested using 5000 bootstraps resamples in order to obtain a 95% confidence interval with the indirect effects.

The results are presented in Table  7 with effect sizes, bootstrap-based standard error, bootstrap-based lower limit confidence interval, and bootstrap-based upper limit confidence interval. The effect is considered statistically significant if the value of zero is not included between the lower limit confidence interval (LLCI) and upper limit confidence interval (ULCI). The results revealed that DC mediated the relationship between OLC and SC, since the lower and upper confidence interval did not include zero. Based upon mediation analysis, H4 was supported.

4.6.2 Robustness analysis

The robustness of the baseline model results using Hayes’ ( 2017 ) approach was further investigated by using Baron and Kenny’s (1986) four steps methodology. This is illustrated below, step by step. The independent variable (X) is OLC; the mediator variable (M) is DC; and the dependent variable (Y) is SC. The control variables (Size and Ownership) were also incorporated into the analysis.

Test for path c alone by conducting a simple regression analysis with X predicting Y.

Relationship between X and Y

Test for path a alone by conducting a simple regression analysis with X predicting M.

Relationship between X and M

Test for path b alone by conducting a simple regression analysis with M predicting Y.

Relationship between M and Y

Test for path c` (c-prime) alone by conducting a multiple regression analysis with X and M predicting Y.

Relationships of X and M to predict Y

After first confirming that a significant association existed from steps 1 through step 3, step 4 was used to determine whether there was either a full or partial mediation. Full mediation exists if X is no longer significant when M is controlled in step 4, while partial mediation exists if X and M both significantly predict Y in step 4.

These results are provided in Table  8 . OLC had a significant relationship with SC and DC which indicated that steps 1 and 2 were met. Also, DC had a significant relationship with SC, thus step 3 was also met. Finally, OLC and DC had a significant relationship with SC. The results showed that DC partially mediates the relationship between OLC and SC. The results from Tables  7 and 8 support each other, showing that DC mediated the relationship between OLC and SC.

The baseline PLS-based SEM model was subjected to further analysis. Multi-Group Analysis was performed in order to determine whether the coefficients were significantly different between small and large-sized firms as well as between foreign and domestic firms. First, the baseline model was subjected to multigroup analysis with firm size as the grouping variable (Small vs. Large.) The sample was split into two groups with small-sized firms that had less than 250 employees, while large-sized firms had 250 employees or more. The results are provided in Table  9 . Accordingly, the coefficients were not significantly different between the small and large-sized firms (p-values > 0.05).

c) Comparison between foreign and domestic ownership .

Similarly, the baseline PLS-based model was subjected to another multi-group analysis using ownership as the grouping variable. The firms were split into two groups: foreign or domestic firms. The analysis results are provided in Table  10 . The results revealed that the coefficient of the baseline research models was not significantly different between foreign and domestic firms (p-value > 0.05).

In the final section of the robustness analysis, the t-test for independent samples was employed for the latent variables by using size and ownership as grouping variables. The results are summarized in Table  11 . They indicated that the Mean OLC of the large size firms was significantly higher than the mean OLC of the small size firms; the mean DC of the large size firms was significantly higher than the mean DC of the small size firms, and the mean SC of the large size firms was not significantly higher than the mean SC of the small size firms at a 5% significance level, but it was weakly significantly higher at a 10% significance level.

Finally, the mean OLC of the foreign firms was significantly higher than the mean OLC of the domestic firms; the mean DC of the foreign firms was significantly higher than the mean DC of the domestic firms, and the mean SC of the foreign firms was significantly higher than the mean SC of the domestic firms at a 5% significance level.

5 Discussion

This study aimed to explore the organizational predecessors of SCC, based upon perceptual data collected from supply chain managers. In this context, the effects of OLC and DC, which are contended to be important organizational resources, were tested by the research. The findings of the study mainly revealed that both OLC and DC are precursors of SCC. It was also found that DC has a mediating effect upon the relationship between OLC and SCC. Variance-based PLS analysis indicated that the organizational structure determined the SCC level. Additionally, in the paired comparisons, the average of large firms was found to be significantly higher compared to that of small firms for all variables (OLC, DC, and SCC). The same result was also valid when comparing foreign companies to domestic ones. In other words, the average for foreign firms was statistically higher than that of domestic companies for all factors. In short, this work provides a number of theoretical and managerial contributions to supply chain research.

5.1 Implications for theory

The decision-making mechanisms of organizations are related to their learning culture. In the literature, the relationship between learning culture and organizational structure was discussed (Janićijević 2013 ). Although in some cases, centralization may increase efficiency, in general it is clear that DC has a positive impact upon organizational effectiveness (Alhamad and Aladwan 2019 ). It can be claimed that learning organizations prioritize decentralized structures because they consider centralization to be an obstacle to learning (Burns and Stalker 1961 ). As in the aforementioned studies, the within research showed that organizations which have a learning culture tend to result in a more decentralized structure.

Decentralized organizational structures enable companies to respond quickly to changing conditions (Zammuto and O’Connor 1992 ) and lead to organizational flexibility in decision-making processes (Hill et al. 2000 ). In other words, centralization may limit responsiveness (Argyris and Schön 1978 ). The extant studies claim that decentralized organizations communicate easily with the external environment such as suppliers and customers (Ogbonna and Harris 2000 ; Hopper 1990 ; Kandemir et al. 2002 ) also claimed that DC leads to cooperation between companies. As in the literature, the analyses of this research indicated that a decentralized structure encourages collaboration with suppliers.

Based upon a resource-based view, the current study proposes a model illustrating whether SCC is directly and/or indirectly affected by an OLC and DC. Our results showed that OLC and DC assist the achievement of an effective SCC level. This is especially important in times of crisis such as the COVID-19 pandemic. Company managers should select suppliers with whom they can work under extra-ordinary conditions. This can be made possible with advanced SCC conditions. While attempting to explain the relationship between OLC and SCC, we also noted that DC, as an organizational resource, is important since it mediates the link between OLC and SCC. As a result, it has become obvious that by employing a decentralized company structure, companies with an advanced learning culture are better able to collaborate with their suppliers. Issues such as the supply of raw materials and the timely and uninterrupted delivery of finished products are critically important in times of crisis. Due to the potential volatility in supply chains in times of crisis, cooperation with suppliers should be a major priority. Thus, companies must institute various institutional resources (such as OLC and DC) in advance.

Additional analyses with control variables were conducted in this study. Companies with less than 250 employees were classified as being small firms. When assuming company size as the control variable, no change is observed in the validity of the hypotheses. The same situation holds true when considering organizational ownership (foreign and domestic) as another control variable. Therefore, the effects of OLC and DC on SCC are valid in this discussion regarding supplier collaboration. The findings indicated that OLC and DC place critical pressures upon SCC. In addition, when OLC, DC and SCC were analyzed separately, it was seen that the averages of large firms were statistically higher than small firms. This result highlights the fact that the relationships of large companies with suppliers are therefore stronger, as compared to small companies. A similar implication is also valid when considering foreign companies. The averages for each factor (OLC, DC and SCC) of foreign companies were significantly higher than domestic companies. This outcome can be explained by the operational flexibility and financial strength of multinational companies. Due to having established many locations around the world, they may have more experience with crises. In addition, it is possible to argue that more investments are made in the internal structures of foreign companies, compared to domestic ones.

5.2 Implications for practice

The COVID-19 pandemic is an unprecedented crisis for supply chains. It has been revealed that there is a serious relationship between supply chain resilience and sustainability. For this reason, these two phenomena should be considered together (Sarkis 2021 ). The literature points out some concepts such as “collaborative management,” “proactive business continuity planning,” and “financial sustainability” as examples of the most important risk reduction strategies in times of crisis like the COVID-19 pandemic (Kumar et al. 2021 ). Therefore, organizations should give importance to cooperation with their suppliers to increase supply chain resilience. The common feature of the institutions that most successfully manage the COVID-19 crisis is having a strong collaboration with local supply chains. Thus, supporting local networks within the supply chains can help to build more vigorous resilience in the face of future shocks (Bassett et al. 2021 ). These findings show that SCC is vital for corporations to provide sustainability in their operations.

Knowledge sharing, being a part of the learning culture, strengthens cooperation among suppliers and/or organizations. (Selnes and Sallis 2003 ; Kandemir et al. 2002 ; Laaksonen et al. 2008 ; Islam et al. 2015 ) drew attention to the collaboration between companies in terms of improving the learning culture. Therefore, the literature reveals that, for organizations, it is advantageous to be in contact with other stakeholders that also prioritize learning. Similarly, the results of this research demonstrate that the learning culture is an important factor leading to collaboration between companies. In other words, companies which improve themselves through continuous learning and remain open to innovations communicate easily with their suppliers. Additionally, the literature shows that firms should seriously consider research and development projects (R&D), artificial intelligence applications and information management systems to foster their learning cultures (Armani et al. 2020 ; Modgil et al. 2021 ). As mentioned above; in times of a crisis such as COVID-19, the importance of the OLC increases because it brings much valuable knowledge to organizations and managers. Thus, during these times, organizations need to increase collaboration with their suppliers in order to avoid potential disruptions in their supply chain. Therefore, companies should integrate the OLC before problems surface. Moreover, in order to share knowledge, organizations which have a learning culture tend to set up decentralized structures. This situation leads organizations toward better internal cooperation and as a result, with other parties (Lee et al. 2012 ).

We believe that these findings will provide valuable implications showing how the interactions of organizational resources can build SCC. First of all, we can confidently claim that companies need to have a learning culture, as it enables SCC. This culture contributes to developing collaboration with suppliers. Therefore, it is important for organizations to invest in a learning culture. Moreover, it stimulates the development of a decentralized structure. This outcome points out that a company with a learning culture is prone to assign various responsibilities to employees at the lower levels. Furthermore, according to the results of this research, it can be claimed that when employees take initiative, this helps to build a strong collaboration with suppliers.

5.3 Limitations and Future research

As with all studies, this study has some limitations. The survey was collected within a specific country, so data from different cultures could be collected for future studies, making it possible to compare results among various countries. Meanwhile, the survey was collected from companies that were operating in varying sectors of manufacturing. It is possible to obtain more specific findings by considering the responses from a specific industry, such as automotive or food. In other words, if only one specific sector is analyzed, differing results may be revealed. Additionally, the questionnaire was often completed by supply chain and manufacturing experts. If the opinions of employees in departments such as marketing and finance are taken into account, the analyses may yield richer results. Finally, different moderators such as the use of information systems (IS) or supply chain orientation could be added to the conceptual model for a broader discussion.

Empirical tests can be included in future studies by adding other organizational resources such as information systems and knowledge storage, to the research model. In particular, the effect of trust upon suppliers with SCC may be examined. In addition, different research questions can be explored, such as: does DC trigger the trust of suppliers?; within which sectors did DC increase during the COVID-19 pandemic period?; under which conditions does DC display disadvantages with SCC?; what is the role of information systems within the relationship between OLC, DC and SCC? Answers to these questions in future studies can either be studied empirically with survey-based methodology or through in-depth interviews with professionals. Moreover, future studies could discuss how the SCC level varies by comparing different countries and/or cultures.

6 Conclusion

This research highlights the importance of organizational intangible resources to improve SCC. Following a detailed literature review, a hypothetical model based upon organizational constructs was developed. The model was tested with variance-based Partial Least Squares (PLS). The results showed the presence of positive effects by suggested intangible resources upon SCC, which in turn was regarded as a critical capability for organizations, especially in times of crisis and/or competitive environments.

The COVID-19 pandemic has seriously affected supply chains, resulting in dramatic cuts to the flow of products and services between companies. Global production in some industries has been interrupted, since the pandemic began in China, and many suppliers of international companies have been located there (Belhadi et al. 2020 ; Brusset and Teller 2017 ). In light of these difficult conditions, SCC has become a critical issue in the worlds of both academia and business. It can be predicted that when the results of the pandemic are finally analyzed, many organizations will prioritize collaboration with their suppliers in the future, since their experience has shown that this is critical for the unbroken flow of products and information. Because of this, this study examined the potential effect of organizational resources such as OLC and DC on SCC after the COVID-19 pandemic. Four main hypotheses have been developed accordingly. Two concern the direct positive effects of both OLC and DC on SCC. Moreover, it was suggested that OLC exerts positive effects upon DC. Finally, the research also indicated that DC mediated the relationship between OLC and SCC. The research model for this study was empirically validated, using data received by way of a survey which had 245 participants. According to the statistical results, both OLC and DC had positive and significant effects for SCC. In addition, learning culture was presented as the precursor for DC, which mediates the relationship between OLC and SCC. As well, firm size and ownership were considered as control variables for the study. Accordingly, the analyses for small and large companies and for domestic and foreign organizations were also evaluated. The results indicated that there was no significant difference concerning the validity and strength of the hypotheses. Finally, it was found that the average of foreign companies was statistically higher than domestic firms for OLC, DC and SCC variables. Likewise, the average of large firms was significantly higher than those of small organizations for all factors.

Ahmed A, Khuwaja FM, Brohi NA, Othman I, Bin L (2018) Organizational factors and organizational performance: A resource-based view and social exchange theory viewpoint. Int J Acad Res Bus Social Sci 8(3):579–599. https://doi.org/10.6007/IJARBSS/v8-i3/3951

Article   Google Scholar  

Ahmed W, Ashraf MS, Khan SA, Kusi-Sarpong S, Arhin FK, Kusi-Sarpong H, Najmi A (2020) Analyzing the impact of environmental collaboration among supply chain stakeholders on a firm’s sustainable performance. Oper Manage Res 13(1):4–21. https://doi.org/10.1007/s12063-020-00152-1

Aiken M, Hage J (1966) Organizational alienation: A comparative analysis. Am Sociol Rev 31:497–507. https://doi.org/10.2307/2090773

Alegre J, Pla-Barber J, Chiva R, Villar C (2012) Organisational learning capability, product innovation performance and export intensity. Technol Anal Strateg Manag 24(5):511–526. https://doi.org/10.1080/09537325.2012.674672

Alhamad BM, Aladwan R (2019) Balancing centralization and decentralization management at University of Bahrain. Qual Assur Educ 27(2):237–250. https://doi.org/10.1108/QAE-01-2018-0005

Al-Tabbaa O, Ankrah S (2019) ‘Engineered’ university‐industry collaboration: A social capital perspective. Eur Manage Rev 16(3):543–565. https://doi.org/10.1111/emre.12174

Altay N, Gunasekaran A, Dubey R, Childe SJ (2018) Agility and resilience as antecedents of supply chain performance under moderating effects of organizational culture within the humanitarian setting: A dynamic capability view. Prod Plann Control 29(14):1158–1174. https://doi.org/10.1080/09537287.2018.1542174

Andreou PC, Louca C, Petrou AP (2016) Organizational learning and corporate diversification performance. J Bus Res 69(9):3270–3284. https://doi.org/10.1016/j.jbusres.2016.02.022

Andrews RW, Boyne GA, Walker RM (2007) Centralization, organizational strategy, and public service performance. J Public Adm Res Theor 19(1):57–80. https://doi.org/10.1093/jopart/mum039

Angerhofer BJ, Angelides MC (2006) A model and a performance measurement system for collaborative supply chains. Decis Support Syst 42(1):283–301. https://doi.org/10.1016/j.dss.2004.12.005

Arefin MS, Hoque MR, Rasul T (2020) Organizational learning culture and business intelligence systems of health-care organizations in an emerging economy. J Knowl Manage 25(3):573–594. https://doi.org/10.1108/JKM-09-2019-0517

Argyris C, Schön DA (1978) Organizational learning. Addison-Wesley, Reading, MA. https://doi.org/10.5465/amr.1985.4279103

Book   Google Scholar  

Armani AM, Hurt DE, Hwang D, McCarthy MC, Scholtz A (2020) Low-tech solutions for the COVID-19 supply chain crisis. Nat Reviews Mater 5(6):403–406. https://doi.org/10.1038/s41578-020-0205-1

Armstrong JS, Overton TS (1977) Estimating nonresponse bias in mail surveys. J Mark Res 14(3):396–402. https://doi.org/10.1177/002224377701400320

Barney J (1986) Organizational culture – can it be a source of sustained competitive advantage? Acad Manage Rev 11:656–665. https://doi.org/10.5465/amr.1986.4306261

Barney J (1991) Firm resources and sustained competitive advantage. J Manag 17:99–120. https://doi.org/10.1177/014920639101700108

Baron RM, Kenny DA (1986) The moderator–mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. J Personal Soc Psychol 51(6):1173. https://doi.org/10.1037/0022-3514.51.6.1173

Bassett HR, Lau J, Giordano C, Suri SK, Advani S, Sharan S (2021) Preliminary lessons from COVID-19 disruptions of small-scale fishery supply chains. World Dev 143:105473. https://doi.org/10.1016/j.worlddev.2021.105473

Belhadi A, Kamble S, Jabbour CJC, Gunasekaran A, Ndubisi NO, Venkatesh M (2020) Manufacturing and service supply chain resilience to the COVID-19 outbreak: Lessons learned from the automobile and airline industries. Technol Forecast Soc Chang 163:120447. https://doi.org/10.1016/j.techfore.2020.120447

Biotto M, De Toni AF, Nonino F (2012) Knowledge and cultural diffusion along the supply chain as drivers of product quality improvement: The Illycaffe case study. Int J Logistics Manage 23(2):212–237. https://doi.org/10.1108/09574091211265369

Brislin RW (1986) The wording and translation of research instruments. In: Lonner WJ, Berry JW (eds) Field methods in cross-cultural research. Sage, Newbury Park, CA, pp 137–164

Google Scholar  

Brislin RW (1970) Back-translation for cross-cultural research. J Cross-Cult Psychol 1(3):185–216. https://doi.org/10.1177/135910457000100301

Brusset X, Teller C (2017) Supply chain capabilities, risks, and resilience. Int J Prod Econ 184:59–68. https://doi.org/10.1016/j.ijpe.2016.09.008

Brynjolfsson E, Hitt LM, Yang S (2002) Intangible assets: Computers and organizational capital. Brookings Papers on Economic Activity, 2002(1), 137–181

Burns T, Stalker GM (1961) The management of innovation. Tavistock Publishing, London

Cai S, Jun M, Yang Z (2010) Implementing supply chain information integration in China: The role of institutional forces and trust. J Oper Manag 28(3):257–268. https://doi.org/10.1016/j.jom.2009.11.005

Cameron KS, Quinn RE (2011) Diagnosing and changing organizational culture: Based on the competing values framework. Jossey Bass, San Francisco

Cao M, Zhang Q (2011) Supply chain collaboration: Impact on collaborative advantage and firm performance. J Oper Manag 29(3):163–180. https://doi.org/10.1016/j.jom.2010.12.008

Carr AS, Pearson JN (1999) Strategically managed buyer–supplier relationships and performance outcomes. J Oper Manag 17:497–519. https://doi.org/10.1016/S0272-6963(99)00007-8

Cassel C, Hackl P, Westlund AH (1999) Robustness of partial least-squares method for estimating latent variable quality structures. J Applied Statistics 26:435–446. https://doi.org/10.1080/02664769922322

Chanani UL, Wibowo UB (2019) A learning culture and continuous learning for a learning organization. KnE Social Sciences 591–598. https://doi.org/10.18502/kss.v3i17.4686

Chang S, Lee M (2007) A study on relationship among leadership, organizational culture, the operation of learning organization and employees’ job satisfaction. Learn Organ 14(2):155–185. https://doi.org/10.1108/09696470710727014

Chen F (1999) Decentralized supply chains subject to information delays. Manage Sci 45(8):1076–1090. https://doi.org/10.1287/mnsc.45.8.1076

Chen JV, Yen DC, Rajkumar TM, Tomochko NA (2011) The antecedent factors on trust and commitment in supply chain relationships. Comput Stand Interface 33(3):262–270. https://doi.org/10.1016/j.csi.2010.05.003

Chen L, Zhao X, Tang O, Price L, Zhang S, Zhu W (2017) Supply chain collaboration for sustainability: A literature review and future research agenda. Int J Prod Econ 194:73–87. https://doi.org/10.1016/j.ijpe.2017.04.005

Chesbrough HW (2003) Open innovation: the new imperative for creating and profiting from technology. Harvard Business Press, Boston

Chiang CY, Hanna M, Liu Z, Lu X (2018) Obtaining collaboration benefits: the role of collaboration-specific investment and absorptive capacity in China. Oper Manage Res 11(3–4):69–82. https://doi.org/10.1007/s12063-018-0133-z

Chin WW, Newsted PR (1999) Structural equation modeling analysis with small samples using partial least squares. Stat Strategies Small Sample Res 1:307–341

Chin WW (1998) The partial least squares approach for structural equation modeling. In: Marcoulides GA (ed) Modern methods for business research. Lawrence Erlbaum Associates, Mahwah, NJ, pp 295–336

Choi I (2020) Moving beyond mandates: Organizational learning culture, empowerment, and performance. Int J Public Adm 43(8):724–735. https://doi.org/10.1080/01900692.2019.1645690

Clulow VG, Gerstman J, Barry CA (2003) The resource-based view and sustainable competitive advantage: The case of a financial services firm. J Eur Industrial Train 27(5):220–232. https://doi.org/10.1108/03090590310469605

Corsten D, Felde J (2005) Exploring the performance effects of key-supplier collaboration: An empirical investigation into Swiss buyer‐supplier relationships. Int J Phys Distribution Logistics Manage 35(6):445–461. https://doi.org/10.1108/09600030510611666

Cullen JB, Perrewé PL (1981) Decision making configurations: An alternative to the centralization/decentralization conceptualization. J Manag 7(2):89–103. https://doi.org/10.1177/014920638100700207

Defee CC, Fugate BS (2010) Changing perspective of capabilities in the dynamic supply chain era. Int J Logistics Manage 21(2):180–206. https://doi.org/10.1108/09574091011071915

Dodgson M (1993) Learning, trust, and technological collaboration. Hum Relat 46(1):77–95. https://doi.org/10.1177/001872679304600106

Dubey R, Altay N, Gunasekaran A, Blome C, Papadopoulos T, Childe SJ (2018) Supply chain agility, adaptability and alignment: empirical evidence from the Indian auto components industry. Int J Oper Prod Manage 38(1):129–148. https://doi.org/10.1108/IJOPM-04-2016-0173

Egan TM, Yang B, Bartlett KR (2004) The effects of organizational Learning culture and job satisfaction on motivation to transfer learning and turnover intention. Hum Res Dev Q 15(3):279–301. https://doi.org/10.1002/hrdq.1104

Eisenhardt KM, Zbaracki MJ (1992) Strategic decision making. Strateg Manag J 13(S2):17–37. https://doi.org/10.1002/smj.4250130904

Eisenberg RS (2018) Shifting institutional roles in biomedical innovation in a learning healthcare system. J Institutional Econ 14(6):1139–1162. https://doi.org/10.1017/S1744137418000115

Ellis SC, Shockley J, Henry RM (2011) Making sense of supply disruption risk research: A conceptual framework grounded in enactment theory. J Supply Chain Manage 47(2):65–96. https://doi.org/10.1111/j.1745-493X.2011.03217.x

Fan M, Stallaert J, Whinston AB (2003) Decentralized mechanism design for supply chain organizations using an auction market. Inform Syst Res 14(1):1–22. https://doi.org/10.1287/isre.14.1.1.14763

Fatorachian H, Kazemi H (2018) A critical investigation of Industry 4.0 in manufacturing: theoretical operationalisation framework. Prod Plann Control 29(8):633–644. https://doi.org/10.1080/09537287.2018.1424960

Fiol CM (1991) Managing culture as a competitive resource: An identity-based view of sustainable competitive advantage. J Manag 17(1):191–211. https://doi.org/10.1177/014920639101700112

Fornell C, Larcker DF (1981) Evaluating structural equation models with unobservable variables and measurement error. J Mark Res 18(1):39–50. https://doi.org/10.1177/002224378101800104

Fynes B, Voss C, de Búrca S (2005) The impact of supply chain relationship quality on quality performance. Int J Production Economics 96(3):339–354. https://doi.org/10.1016/j.ijpe.2004.05.008

Gaber DA (2003) Building a system of autonomous institutions: Coordinaton and collaboration in British Columbia’s community college, university college, and institute system. Community Coll Rev 31(2):47–72. https://doi.org/10.1177/009155210303100203

Gartner (2018) IT glossary/digitalization. Retrieved from https://www.gartner.com/en/information-technology/glossary/digitalization , Accessed 21 April 2022

Gatignon A, Van Wassenhove LN, Charles A (2010) The Yogyakarta earthquake: Humanitarian relief through IFRC’s decentralized supply chain. Int J Prod Econ 126(1):102–110. https://doi.org/10.1016/j.ijpe.2010.01.003

Gephart MA, Marsick VJ, Van Buren ME, Spiro MS, Senge P (1996) Learning organizations come alive. Train Dev 50(12):34–46

Gopal A, Gosain S (2010) Research note - The role of organizational controls and boundary spanning in software development outsourcing: Implications for project performance. Inform Syst Res 21(4):960–982. https://doi.org/10.1287/isre.1080.0205

Gupta S, Tuunanen T, Kar AK, Modgil S (2022) Managing digital knowledge for ensuring business efficiency and continuity. Journal of Knowledge Management, Vol. ahead-of-print No. ahead-of-print. https://doi.org/10.1108/JKM-09-2021-0703

Hair JF Jr, Black WC, Babin BJ, Anderson RE (2019) Multivariate data analysis. Cengage Learning

Hales C (1999) Leading horses to water? The impact of decentralization on managerial behavior. J Manage Stud 36(6):831–851. https://doi.org/10.1111/1467-6486.00160

Hamel G (1991) Competition for competence and interpartner learning within international strategic alliances. Strateg Manag J 12(S1):83–103. https://doi.org/10.1002/smj.4250120908

Harman HH (1960) Modern factor analysis. University of Chicago Press, Chicago

Hawkins JN (2000) Centralization, decentralization, recentralization – educational reform in China. J Educational Adm 38(5):442–455. https://doi.org/10.1108/09578230010378340

Hayes AF (2017) Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. Guilford Publications

Hayes AF (2019) PROCESS v3. 4 [computer software]

Hempel PS, Zhang Z-X, Han Y (2012) Team empowerment and the organizational context decentralization and the contrasting effects of formalization. J Manag 38(2):475–501. https://doi.org/10.1177/0149206309342891

Henseler J, Ringle CM, Sinkovics RR (2009) The use of partial least squares path modeling in international marketing (ed. by Sinkovics, R. R. and P. N., G.), 277–319. Emerald, Bingley

Hewett K, Bearden WO (2001) Dependence, trust, and relational behavior on the part of foreign subsidiary marketing operations: Implications for managing global marketing operations. J Mark 65(4):51–66. https://doi.org/10.1509/jmkg.65.4.51.18380

Hill S, Martin R, Harris M (2000) Decentralization, integration and the post-bureaucratic organization: The case of R&D. J Manage Stud 37(4):563–586. https://doi.org/10.1111/1467-6486.00194

Höck M, Ringle CM (2006) Strategic networks in the software industry: An empirical analysis of the value continuum. IFSAM VIIIth World Congress, Berlin 2006

Hopper MD (1990) Rattling SABRE-new ways to compete on information. Harv Bus Rev 68(3):118–125

Hosseini SH, Hajipour E, Kaffashpoor A, Darikandeh A (2020) The mediating effect of organizational culture in the relationship of leadership style with organizational learning. J Hum Behav Social Environ 30(3):279–288. https://doi.org/10.1080/10911359.2019.1680473

Hu LT, Bentler PM (1999) Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Struct Equation Modeling: Multidisciplinary J 6(1):1–55. https://doi.org/10.1080/10705519909540118

Hult GTA, Ketchen DJ, Nichols EL (2003) Organizational learning as a strategic resource in supply management. J Oper Manag 21(5):541–556. https://doi.org/10.1016/j.jom.2003.02.001

Innocent O (2015) The performance of commercial banks: The role of organizational culture as a mediator and external environment as a moderator (Doctoral dissertation, Universiti Utara Malaysia)

Islam MZ, Jasimuddin SM, Hasan I (2015) Organizational culture, structure, technology infrastructure and knowledge sharing. Vine 45(1):67–88. https://doi.org/10.1108/VINE-05-2014-0037

Janićijević N (2013) The mutual impact of organizational culture and structure. Economic Annals 58(198):35–60. https://doi.org/10.2298/EKA1398035J

Jin Y, Hong P (2007) Coordinating global inter-firm product development. J Enterp Inform Manage 20(5):544–561. https://doi.org/10.1108/17410390710823699

Kandemir D, Ghauri P, Cavusgil ST (2002) The strategic role of organizational learning in relationship quality in strategic alliances. In: Contractor FJ, Lorange P (eds) Cooperative strategies and alliances. Elsevier Science, Oxford, pp 799–828

Kanter RM (1994) Collaborative advantage: the art of alliances. Harvard Business Rev 72(4):96–108

Kochen M, Deutsch KW (1980) Decentralization. Oelgeschlager, Gunn and Hain, Cambridge. Mass

Kock N (2015) Common method bias in PLS-SEM: A full collinearity assessment approach. Int J e-Collaboration 11(4):1–10. DOI: https://doi.org/10.4018/ijec.2015100101

Kumar A, Mangla SK, Kumar P, Song M (2021) Mitigate risks in perishable food supply chains: Learning from COVID-19. Technol Forecast Soc Chang 166:120643. https://doi.org/10.1016/j.techfore.2021.120643

Laaksonen T, Pajunen K, Kulmala HI (2008) Co-evolution of trust and dependence in customer–supplier relationships. Ind Mark Manage 37(8):910–920. https://doi.org/10.1016/j.indmarman.2007.06.007

Lau KW, Lee PY, Chung YY (2019) A collective organizational learning model for organizational development. Leadersh Organ Dev J 40(1):107–123. https://doi.org/10.1108/LODJ-06-2018-0228

Lee J, Veloso FM, Hounshell DA, Rubin ES (2010) Forcing technological change: a case of automobile emissions control technology development in the US. Technovation 30(4):249–264. https://doi.org/10.1016/j.technovation.2009.12.003

Lee J, Palekar US, Qualls W (2011) Supply chain efficiency and security: Coordination for collaborative investment in technology. Eur J Oper Res 210(3):568–578. https://doi.org/10.1016/j.ejor.2010.10.015

Lee MY, Edmondson AC (2017) Self-managing organizations: Exploring the limits of less-hierarchical organizing. Res Organizational Behav 37:35–58. https://doi.org/10.1016/j.riob.2017.10.002

Lee S, Kim BG, Kim H (2012) An integrated view of knowledge management for performance. J Knowl Manage 16(2). https://doi.org/10.1108/13673271211218807

Li S, Chen X (2019) The role of supplier collaboration and risk management capabilities in managing product complexity. Oper Manage Res 12(3):146–158. https://doi.org/10.1007/s12063-019-00144-w

Mahmood K, Ilyas M, Rehman CA (2014) Impact of knowledge management and decentralization on supply chain performance: A study of automobile sector of Pakistan. Int J Oper Logistics Manage 3(2):124–139

Mamillo D (2015) Supply chain collaboration under uncertainty in the Albanian beer market. Manage Dynamics Knowl Econ 3(1):99–117

Marcoulides GA, Saunders C (eds) (2006) Editor’s comments: PLS: A silver bullet?. MIS Quarterly, iii-ix. https://doi.org/10.2307/25148727

Marks HM, Louis KS (1999) Teacher empowerment and the capacity for organizational learning. Educational Adm Q 35:707–750. https://doi.org/10.1177/0013161X99355003

Marsick VJ, Watkins KE (2003) Demonstrating the value of an organization’s learning culture: the dimensions of the learning organization questionnaire. Adv Developing Hum Resour 5(2):132–151. https://doi.org/10.1177/1523422303005002002

Mehralizadeh Y (2005) New reforms in the management of the university: transition from centralized to decentralized (university-based management) in Iran. High Educ Policy 18(1):67–82. https://doi.org/10.1057/palgrave.hep.8300073

Melville N, Kraemer K, Gurbaxani V (2004) Review: information technology and organisational performance: An integrative model of IT business value. MIS Q 28:283–322. https://doi.org/10.2307/25148636

Mirc N, Rouzies A, Teerikangas S (2017) Do academics actually collaborate in the study of interdisciplinary phenomena? A look at half a century of research on mergers and acquisitions. Eur Manage Rev 14(3):333–357. https://doi.org/10.1111/emre.12110

Mirkamali SM, Thami FN, Alami F (2011) Examining the role of transformational leadership and job satisfaction in the organizational learning of an automotive manufacturing company. Procedia - Social and Behavioral Sciences 29:139–148

Modgil S, Singh RK, Hannibal C (2021) Artificial intelligence for supply chain resilience: Learning from COVID-19. The International Journal of Logistics Management, Vol. ahead-of-print No. ahead-of-print. https://doi.org/10.1108/IJLM-02-2021-0094

Moingeon B, Ramanantsoa B, Métais E, Orton JD (1998) Another look at strategy–structure relationships: The resource-based view. Eur Manag J 16(3):297–305. https://doi.org/10.1016/S0263-2373(98)00006-1

Moran TJ, Meso P (2008) A resource-based view of manufacturing strategy and implications to organizational culture and human resources. J Bus Econ Res (JBER) 6(11). https://doi.org/10.19030/jber.v6i11.2494

Nasirpour AA, Gohari MR, Moradi S (2010) The relationship of centralization, organizational culture and performance indexes in teaching hospitals affiliated to Tehran University of medical sciences. ActaMedica Iranica 48(5):326–331

Ogbonna E, Harris LC (2000) Leadership style, organizational culture and performance: empirical evidence from UK companies. Int J Hum Resour Manage 11(4):766–788. https://doi.org/10.1080/09585190050075114

Ojha D, Struckell E, Acharya C, Patel PC (2018) Supply chain organizational learning, exploration, exploitation, and firm performance: A creation-dispersion perspective. Int J Prod Econ 204:70–82. https://doi.org/10.1016/j.ijpe.2018.07.025

Özçelik G, Aybas M, Uyargil C (2016) High performance work systems and organizational values: Resource-based view considerations. Procedia-Social and Behavioral Sciences 235:332–341. https://doi.org/10.1016/j.sbspro.2016.11.040

Parkhe A (1993) Strategic alliance structuring: a game theoretic and transaction cost examination of interfirm cooperation. Acad Manag J 36(4):794–829. https://doi.org/10.5465/256759

Parmigiani A, Mitchell W (2010) The hollow corporation revisited: Can governance mechanisms substitute for technical expertise in managing buyer-supplier relationships? Eur Manage Rev 7(1):46–70. https://doi.org/10.1057/emr.2009.28

Penrose ET (1959) The theory of the growth of the firm. John Willey, New York

Podsakoff PM, MacKenzie SB, Lee JY, Podsakoff NP (2003) Common method biases in behavioral research: a critical review of the literature and recommended remedies. J Appl Psychol 88(5):879. DOI: https://doi.org/10.1037/0021-9010.88.5.879

Potnuru RKG, Sahoo CK, Sharma R (2019) Team building, employee empowerment and employee competencies: moderating role of organizational learning culture. Eur J Train Dev 43(1/2):39–60. https://doi.org/10.1108/EJTD-08-2018-0086

Preacher KJ, Hayes AF (2008) Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behav Res Methods 40(3):879–891. https://doi.org/10.3758/BRM.40.3.879

Ratchavieng A, Srinet S (2021) An innovative organization model for efficient industrial business operations in the digital era. RICE J Creative Entrepreneurship Manage 2(2):24–37. doi: https://doi.org/10.14456/rjcm.2021.28

Rose DS, Sidle SD, Griffith KH (2007) A penny for your thoughts: Monetary incentives improve response rates for company-sponsored employee surveys. Organizational Res Methods 10(2):225–240. https://doi.org/10.1177/1094428106294687

Qrunfleh S, Tarafdar M (2014) Supply chain information systems strategy: Impacts on supply chain performance and firm performance. Int J Prod Econ 147(January):340–350. https://doi.org/10.1016/j.ijpe.2012.09.018

Quinn C, Rohrbaugh J (1983) A spatial model of effectiveness criteria: Towards a competing values approach to organizational analysis. Manage Sci 29(3):336–377. https://doi.org/10.1287/mnsc.29.3.363

Ralston PM, Richey RG, Grawe SJ (2017) The past and future of supply chain collaboration: A literature synthesis and call for research. Int J Logistics Manage 28(2):508–530. https://doi.org/10.1108/IJLM-09-2015-0175

Ramanathan U, Gunasekaran A (2014) Supply chain collaboration: Impact of success in long-term partnerships. Int J Prod Econ 147(Part B 252–259. https://doi.org/10.1016/j.ijpe.2012.06.002

Ranson S, Hinings B, Greenwood R (1980) The structuring of organizational structures. Adm Sci Q 25(1):1–17. https://doi.org/10.2307/2392223

Richardson HA, Vandenberg RJ, Blum TC, Roman PM (2002) Does decentralization make a difference for the organization? An examination of the boundary conditions circumscribing decentralized decision-making and organizational financial performance. J Manag 28(2):217–244. https://doi.org/10.1177/014920630202800205

Ringle CM, Sarstedt M, Straub DW (eds) (2012) Editor’s comments: A critical look at the use of PLS-SEM in MIS Quarterly. MIS Quarterly, 36 (1): iii-xiv. https://doi.org/10.2307/41410402

Ringle C, Wende S, Will A (2005) Smart-PLS Version 2.0 M3. University of Hamburg

Sarkis J (2021) Supply chain sustainability: Learning from the COVID-19 pandemic. Int J Oper Prod Manage 41(1):63–73. https://doi.org/10.1108/IJOPM-08-2020-0568

Schein EH (1992) How can organizations learn faster? The challenge of entering the green room. MIT Sloan Management Review 34(2):85

Schiefelbein E, Schiefelbein P (2000) Three decentralization strategies in two decades: Chile 1981-2000. J Educational Adm 38(5):412–425. https://doi.org/10.1108/09578230010378313

Schilke O, Cook KS (2015) Sources of alliance partner trustworthiness: Integrating calculative and relational perspectives. Strateg Manag J 36(2):276–297. https://doi.org/10.1002/smj.2208

Selnes F, Sallis J (2003) Promoting relationship learning. J Mark 67(3):80–95. https://doi.org/10.1509/jmkg.67.3.80.18656

Senge P (1992) The fifth discipline: the art and practice of the learning organization, vol 3. Random House, Milsons Point

Škerlavaj M, Štemberger MI, Dimovski V (2007) Organizational learning culture—the missing link between business process change and organizational performance. Int J Prod Econ 106(2):346–367. https://doi.org/10.1016/j.ijpe.2006.07.009

Soosay CA, Hyland P (2015) A decade of supply chain collaboration and directions for future research. Supply Chain Management 20(6):613–630. https://doi.org/10.1108/SCM-06-2015-0217

Soylu A, Oruc C, Turkay M, Fujita K, Asakura T (2006) Synergy analysis of collaborative supply chain management in energy systems using multi-period MILP. Eur J Oper Res 174(1):387–403. https://doi.org/10.1016/j.ejor.2005.02.042

Tanner JF Jr (2021) The state of business-to-business marketing research. J Mark Theory Pract 1–9. https://doi.org/10.1080/10696679.2020.1860682

Teece DJ (1992) Competition, cooperation, and innovation – organizational arrangements for regimes of rapid technological progress. J Econ Behav Organ 18:1–25. https://doi.org/10.1016/0167-2681(92)90050-L

Teece DJ (2015) Intangible assets and a theory of heterogeneous firms. In: Bounfour A, Miyagawa T (eds) Intangibles, market failure and innovation performance. Springer, New York, pp 217–239

Chapter   Google Scholar  

Tomassetti AJ, Reeshad S, Dalal, Kaplan SA (2016) Is policy capturing really more resistant than traditional self-report techniques to socially desirable responding? Organizational Res Methods 19(2):255–285. https://doi.org/10.1177/1094428115627497

Truant E, Broccardo L, Dana LP (2021) Digitalisation boosts company performance: an overview of Italian listed companies. Technol Forecast Soc Chang 173:121173. https://doi.org/10.1016/j.techfore.2021.121173

Tynjälä P (2013) Toward a 3-P model of workplace learning: A literature review. Vocations and Learning 6(1):11–36. https://doi.org/10.1007/s12186-012-9091-z

Vancil RF (1979) Decentralization: managerial ambiguity by design. Dow Jones Irwin, Homewood, IL

Walter A (2003) Relationship-specific factors influencing supplier involvement in customer new product development. J Bus Res 56(9):721–733. https://doi.org/10.1016/S0148-2963(01)00257-0

Wang E (2010) Fiscal decentralization and revenue/expenditure disparities in China. Eurasian Geogr Econ 51(6):744–766. https://doi.org/10.2747/1539-7216.51.6.744

Watkins KE, Marsick VJ (1993) Sculpting the learning organization: Lessons in the art and science of systemic change. Jossey-Bass Inc., San Francisco

Weber K (2005) A toolkit for analyzing corporate cultural toolkits. Poetics 33(3–4):227–252. https://doi.org/10.1016/j.poetic.2005.09.011

Wernerfelt B (1984) A resource-based view of the firm. Strateg Manag J 5(2):171–180. https://doi.org/10.1002/smj.4250050207

Willis G, Genchev SE, Chen H (2016) Supply chain learning, integration, and flexibility performance: An empirical study in India. Int J Logistics Manage 27(3):755–769. https://doi.org/10.1108/IJLM-03-2014-0042

Yang B (2003) Identifying valid and reliable measures for dimensions of a learning culture. Adv Developing Hum Resour 5(2):152–162. https://doi.org/10.1177/1523422303005002003

Zahra SA, Hayton JC, Salvato C (2004) Entrepreneurship in family vs. non–family firms: A resource–based analysis of the effect of organizational culture. Entrepreneurship Theory and Practice 28(4):363–381. https://doi.org/10.1111/j.1540-6520.2004.00051.x

Zammuto RF, O’Connor EJ (1992) Gaining advanced manufacturing technologies’ benefits: The roles of organization design and culture. Acad Manage Rev 17(4):701–728. https://doi.org/10.5465/amr.1992.4279062

Zhang Q, Cao M (2018) Exploring antecedents of supply chain collaboration: Effects of culture and interorganizational system appropriation. Int J Prod Econ 195:146–157. https://doi.org/10.1016/j.ijpe.2017.10.014

Zhao X, Huo B, Selen W, Yeung JHY (2011) The impact of internal integration and relationship commitment on external integration. J Oper Manag 29(1):17–32. https://doi.org/10.1016/j.jom.2010.04.004

Ağan Y, Kuzey C, Acar MF, Açıkgöz A (2016) The relationships between corporate social responsibility, environmental supplier development, and firm performance. J Clean Prod 112:1872–1881. https://doi.org/10.1016/j.jclepro.2014.08.090

Download references

Author information

Authors and affiliations.

Department of Management, Izmir Demokrasi University, Üçkuyular, Gürsel Aksel Blv No:14, 35140, Karabağlar/İzmir, Turkey

Alev Ozer Torgaloz

Department of International Trade and Business, Izmir Katip Celebi University, Havaalanı Şosesi Cd. No:33/2, 35620, Balatçık, Atatürk Osb/Çiğli/İzmir, Turkey

Mehmet Fatih Acar

Arthur J. Bauernfeind College of Business, Murray State University, Murray, KY, USA

Cemil Kuzey

You can also search for this author in PubMed   Google Scholar

Corresponding author

Correspondence to Cemil Kuzey .

Ethics declarations

Declarations of interest.

All authors certify that they have no affiliations with or involvement in any organization or entity with any financial interest or non-financial interest in the subject matter or materials discussed in this manuscript

Additional information

Publisher’s note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Springer Nature or its licensor holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Torgaloz, A.O., Acar, M.F. & Kuzey, C. The effects of organizational learning culture and decentralization upon supply chain collaboration: analysis of covid-19 period. Oper Manag Res 16 , 511–530 (2023). https://doi.org/10.1007/s12063-022-00316-1

Download citation

Received : 10 November 2021

Revised : 21 July 2022

Accepted : 03 August 2022

Published : 01 September 2022

Issue Date : March 2023

DOI : https://doi.org/10.1007/s12063-022-00316-1

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Supply chain collaboration
• Organizational Learning Culture
• Decentralization
• Resource-based View
• Find a journal
• Publish with us
• Track your research

Academia.edu no longer supports Internet Explorer.

To browse Academia.edu and the wider internet faster and more securely, please take a few seconds to  upgrade your browser .

Enter the email address you signed up with and we'll email you a reset link.

• We're Hiring!
• Help Center

Organizational learning and knowledge in public service organizations: A systematic review of the literature

2009, International Journal of Management Reviews

Related Papers

Karol Olejniczak , Anna Domaradzka

The complexity of modern socio-economic issues turns our public policies into a continuous trial and error process. Every time public managers and stakeholders address a new policy issue they have to use research results, experience and insight to find out what works for whom and in what context. That requires from public organizations intense organizational learning. This challenge is especially demanding for administrations of countries that are undergoing modernization of their socio- economic systems. Poland is an example of such country. In order to face the challenge of effective organizational learning, we need to address three pressing questions: 1. How does learning work in our public organizations? 2. What promising practices can we implement to advance learning in public organizations? 3. What changes in public management are required to combine learning with the growing demands of performance and accountability? This book is an attempt to address those questions in a systematic and empirical manner. The answers presented in this volume are the result of a four-year empirical research project conducted in Polish ministries and study visits in public institutions of twelve countries of the Organization for Economic Co-operation and Development.

Joanna Samul , Foteini Asderaki

Background. The importance of learning in and by organizations has since long been recognized by academic and practitioner scientists for both business and public sector organizations. In particular in the last decades interest in organizational learning has been growing. There are important reasons for considering the acquisition of knowledge in public service organizations. Research aims. The paper reviews the state of research evidence in the field. The main question of this study is how important it is for employees in public sector to acquire the knowledge and use their full potential in the workplace and whether these organizations

OECD Journal on Budgeting

International Public Management …

Fred Thompson , G. Reschenthaler

Chris Gilson

PPG undertakes pure and applied research, policy evaluation and consultancy for government bodies, international organizations and major corporations active in the fields of policy evaluation, public management, budgeting and audit, and egovernment, survey or focus group research, public opinion, and the design of election systems.

Gerda Van Dijk , Christopher Thornhill

Teaching Public Administration

Ian C Elliott

It is broadly accepted that learning is important in supporting the delivery of change. Furthermore, during times of public sector budget cuts there is a growing interest in innovation and change. However, it is also generally accepted that austere times typically lead to a reduction in training and development budgets – particularly within the public sector. This paper explores the extent to which the assumed link between learning and change is contested during austerity. Fifty-one public sector managers from Wales and Scotland participated in the study, over a period of 13 months, from a wide range of public sector organisations across two devolved parts of the UK. The research explored issues surrounding motivations for learning, expectations of outcomes and perceptions around the drivers of change. It is found that public sector budget cuts are having a direct impact on learning and development opportunities across the UK public sector. Whilst budget cuts may lead to change, thi...

Global Encyclopedia of Public Administration, Public Policy, and Governance

Jonathan Rauh

Organizations with similar missions may learn from each other in an effort to gain additional knowledge, skills or abilities, all with a long-term goal of enhancing their reputation or gaining a competitive advantage. The incentive to seek these benefits stems from, among other things, competition over similar budgets and scopes of authority. Because learning often requires resources to be shifted to the learning process, unless the learned activities provide a benefit that is at least marginally greater than the cost of learning, the learning is not beneficial to the organization. In this paper the author develops a theoretical model of agency learning and the addition of new goals under dyadic collaboration, and provides a vignette as an illustration of the model.

Monica Kennedy

Canadian Public Administration/Administration publique du Canada

Louise Lemyre , Nancy Beauregard

RELATED PAPERS

Pedro Alejandro Toledo Barrera

Lifelong education: the XXI century

Elena Ignatovich

La Lettre de l’OCIM

sebastien minchin

Journal of Colloid and Interface Science

Ryan Workman

Journal of Geophysical Research

Warwick Smith

Tobin LeBlanc Haley

Bank and Policy

Dr. Nasir Mammadov

Arhivski Vjesnik

Ivana Prgin

Journal of the Mathematical Society of Japan

Ezzeddine Haouala

Nigerian Journal of Medical Rehabilitation

Olufemi Oyewole

Gastroenterology Review

Piotr Arkuszewski

BMC Complementary Medicine and Therapies

Dr. Iva Lloyd, ND

Neuro-Oncology

ozdil baskan

ANA LILIA GUZMAN CERRATO

Revista Electrónica de LEEME

jessica lopez perez

massimiliano stacchini

Yoennys Sánchez Martínez

Journal of the Medical Association of Thailand = Chotmaihet thangphaet

supaporn tengtrisorn

Journal of the American Veterinary Medical Association

Ronaldo Da Costa

الدويلة السورية الموازية: الفرقة المُدرَّعة الرابعة

Muhanad Abulhusn

International journal of molecular sciences

Luis Meseguer-Olmo, MD. PhD.

Ketan Tamboli

Anesthesiology

jean-jacques lehot

American Journal of Alzheimer's Disease & Other Dementiasr

Martin Elodie

Nordisk välfärdsforskning =

Anna Balkfors

See More Documents Like This

RELATED TOPICS

•   We're Hiring!
•   Help Center
• Find new research papers in:
• Health Sciences
• Earth Sciences
• Cognitive Science
• Mathematics
• Computer Science
• Academia ©2024

• SUGGESTED TOPICS
• The Magazine
• Newsletters
• Managing Yourself
• Managing Teams
• Work-life Balance
• The Big Idea
• Data & Visuals
• Reading Lists
• Case Selections
• HBR Learning
• Topic Feeds
• Account Settings
• Email Preferences

How Gamification Can Boost Employee Engagement

• Adrian R. Camilleri
• Ananta Neelim

Research on what works — and what to avoid.

Employee disengagement is a persistent problem, and attempts to inject excitement often fall flat. However, gamification — using elements of games to motivate — has serious potential when thoughtfully executed. This article explores the psychology behind gamification, successful examples, and how to leverage probabilistic rewards (like lotteries tied to performance) to increase employee motivation.

A core responsibility of every manager is to motivate and engage their employees. This is because disengaged employees show lower productivity and exhibit higher absenteeism. Traditional approaches to increasing employee engagement include giving workers more  autonomy , a higher sense of  belonging and purpose , and  additional growth opportunities . Nevertheless, employee disengagement in the workplace hit a nine-year high in 2022 in the U.S. ,  with approximately three-fourths of employees reporting being disengaged.

• AC Adrian Camilleri is an associate professor of marketing at the University of Technology Sydney (UTS) Business School. He uses experimental and survey research methods to understand, explain, and predict the cognitive processes underlying judgment and decision-making, and the application of this knowledge to environmental, financial, managerial, and consumption contexts.
• AN Ananta Neelim is a senior lecturer in economics at the University of Tasmania. His research specializes in employing behavioral economics to understand and tackle issues of diversity, discrimination, and workplace productivity.

Partner Center

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings

Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .

• Advanced Search
• Journal List
• Springer Nature - PMC COVID-19 Collection

Systematic Literature Review of E-Learning Capabilities to Enhance Organizational Learning

Michail n. giannakos.

1 Norwegian University of Science and Technology, Trondheim, Norway

Patrick Mikalef

Ilias o. pappas.

2 University of Agder, Kristiansand, Norway

E-learning systems are receiving ever increasing attention in academia, business and public administration. Major crises, like the pandemic, highlight the tremendous importance of the appropriate development of e-learning systems and its adoption and processes in organizations. Managers and employees who need efficient forms of training and learning flow within organizations do not have to gather in one place at the same time or to travel far away to attend courses. Contemporary affordances of e-learning systems allow users to perform different jobs or tasks for training courses according to their own scheduling, as well as to collaborate and share knowledge and experiences that result in rich learning flows within organizations. The purpose of this article is to provide a systematic review of empirical studies at the intersection of e-learning and organizational learning in order to summarize the current findings and guide future research. Forty-seven peer-reviewed articles were collected from a systematic literature search and analyzed based on a categorization of their main elements. This survey identifies five major directions of the research on the confluence of e-learning and organizational learning during the last decade. Future research should leverage big data produced from the platforms and investigate how the incorporation of advanced learning technologies (e.g., learning analytics, personalized learning) can help increase organizational value.

Introduction

E-learning covers the integration of information and communication technology (ICT) in environments with the main goal of fostering learning (Rosenberg and Foshay 2002 ). The term “e-learning” is often used as an umbrella term to portray several modes of digital learning environments (e.g., online, virtual learning environments, social learning technologies). Digitalization seems to challenge numerous business models in organizations and raises important questions about the meaning and practice of learning and development (Dignen and Burmeister 2020 ). Among other things, the digitalization of resources and processes enables flexible ways to foster learning across an organization’s different sections and personnel.

Learning has long been associated with formal or informal education and training. However organizational learning is much more than that. It can be defined as “a learning process within organizations that involves the interaction of individual and collective (group, organizational, and inter-organizational) levels of analysis and leads to achieving organizations’ goals” (Popova-Nowak and Cseh 2015 ) with a focus on the flow of knowledge across the different organizational levels (Oh 2019 ). Flow of knowledge or learning flow is the way in which new knowledge flows from the individual to the organizational level (i.e., feed forward) and vice versa (i.e., feedback) (Crossan et al. 1999 ; March 1991 ). Learning flow and the respective processes constitute the cornerstone of an organization’s learning activities (e.g., from physical training meetings to digital learning resources), they are directly connected to the psycho-social experiences of an organization’s members, and they eventually lead to organizational change (Crossan et al. 2011 ). The overall organizational learning is extremely important in an organization because it is associated with the process of creating value from an organizations’ intangible assets. Moreover, it combines notions from several different domains, such as organizational behavior, human resource management, artificial intelligence, and information technology (El Kadiri et al. 2016 ).

A growing body of literature lies at the intersection of e-learning and organizational learning. However, there is limited work on the qualities of e-learning and the potential of its qualities to enhance organizational learning (Popova-Nowak and Cseh 2015 ). Blockages and disruptions in the internal flow of knowledge is a major reason why organizational change initiatives often fail to produce their intended results (Dee and Leisyte 2017 ). In recent years, several models of organizational learning have been published (Berends and Lammers 2010 ; Oh 2019 ). However, detailed empirical studies indicate that learning does not always proceed smoothly in organizations; rather, the learning meets interruptions and breakdowns (Engeström et al. 2007 ).

Discontinuities and disruptions are common phenomena in organizational learning (Berends and Lammers 2010 ), and they stem from various causes. For example, organizational members’ low self-esteem, unsupportive technology and instructors (Garavan et al. 2019 ), and even crises like the Covid-19 pandemic can result in demotivated learners and overall unwanted consequences for their learning (Broadbent 2017 ). In a recent conceptual article, Popova-Nowak and Cseh ( 2015 ) emphasized that there is a limited use of multidisciplinary perspectives to investigate and explain the processes and importance of utilizing the available capabilities and resources and of creating contexts where learning is “attractive to individual agents so that they can be more engaged in exploring ways in which they can contribute through their learning to the ongoing renewal of organizational routines and practices” (Antonacopoulou and Chiva 2007 , p. 289).

Despite the importance of e-learning, the lack of systematic reviews in this area significantly hinders research on the highly promising value of e-learning capabilities for efficiently supporting organizational learning. This gap leaves practitioners and researchers in uncharted territories when faced with the task of implementing e-learning designs or deciding on their digital learning strategies to enhance the learning flow of their organizations. Hence, in order to derive meaningful theoretical and practical implications, as well as to identify important areas for future research, it is critical to understand how the core capabilities pertinent to e-learning possess the capacity to enhance organizational learning.

In this paper, we define e-learning enhanced organizational learning (eOL) as the utilization of digital technologies to enhance the process of improving actions through better knowledge and understanding in an organization. In recent years, a significant body of research has focused on the intersection of e-learning and organizational learning (e.g., Khandakar and Pangil 2019 ; Lin et al. 2019 ; Menolli et al. 2020 ; Turi et al. 2019 ; Xiang et al. 2020 ). However, there is a lack of systematic work that summarizes and conceptualizes the results in order to support organizations that want to move from being information-based enterprises to being knowledge-based ones (El Kadiri et al. 2016 ). In particular, recent technological advances have led to an increase in research that leverages e-learning capacities to support organizational learning, from virtual reality (VR) environments (Costello and McNaughton 2018 ; Muller Queiroz et al. 2018 ) to mobile computing applications (Renner et al. 2020 ) to adaptive learning and learning analytics (Zhang et al. 2019 ). These studies support different skills, consider different industries and organizations, and utilize various capacities while focusing on various learning objectives (Garavan et al. 2019 ). Our literature review aims to tease apart these particularities and to investigate how these elements have been utilized over the past decade in eOL research. Therefore, in this review we aim to answer the following research questions (RQs):

• RQ1: What is the status of research at the intersection of e-learning and organizational learning, seen through the lens of areas of implementation (e.g., industries, public sector), technologies used, and methodologies (e.g., types of data and data analysis techniques employed)?
• RQ2: How can e-learning be leveraged to enhance the process of improving actions through better knowledge and understanding in an organization?

Our motivation for this work is based on the emerging developments in the area of learning technologies that have created momentum for their adoption by organizations. This paper provides a review of research on e-learning capabilities to enhance organizational learning with the purpose of summarizing the findings and guiding future studies. This study can provide a springboard for other scholars and practitioners, especially in the area of knowledge-based enterprises, to examine e-learning approaches by taking into consideration the prior and ongoing research efforts. Therefore, in this paper we present a systematic literature review (SLR) (Kitchenham and Charters 2007 ) on the confluence of e-learning and organizational learning that uncovers initial findings on the value of e-learning to support organizational learning while also delineating several promising research streams.

The rest of this paper is organized as follows. In the next section, we present the related background work. The third section describes the methodology used for the literature review and how the studies were selected and analyzed. The fourth section presents the research findings derived from the data analysis based on the specific areas of focus. In the fifth section, we discuss the findings, the implications for practice and research, and the limitations of the selected methodological approach. In the final section, we summarize the conclusions from the study and make suggestions for future work.

Background and Related Work

E-learning systems.

E-learning systems provide solutions that deliver knowledge and information, facilitate learning, and increase performance by developing appropriate knowledge flow inside organizations (Menolli et al. 2020 ). Putting into practice and appropriately managing technological solutions, processes, and resources are necessary for the efficient utilization of e-learning in an organization (Alharthi et al. 2019 ). Examples of e-learning systems that have been widely adopted by various organizations are Canvas, Blackboard, and Moodle. Such systems provide innovative services for students, employees, managers, instructors, institutions, and other actors to support and enhance the learning processes and facilitate efficient knowledge flow (Garavan et al. 2019 ). Functionalities, such as creating modules to organize mini course information and learning materials or communication channels such as chat, forums, and video exchange, allow instructors and managers to develop appropriate training and knowledge exchange (Wang et al. 2011 ). Nowadays, the utilization of various e-learning capabilities is a commodity for supporting organizational and workplace learning. Such learning refers to training or knowledge development (also known in the literature as learning and development, HR development, and corporate training: Smith and Sadler-Smith 2006 ; Garavan et al. 2019 ) that takes place in the context of work.

Previous studies have focused on evaluating e-learning systems that utilize various models and frameworks. In particular, the development of maturity models, such as the e-learning capability maturity model (eLCMM), addresses technology-oriented concerns (Hammad et al. 2017 ) by overcoming the limitations of the domain-specific models (e.g., game-based learning: Serrano et al.  2012 ) or more generic lenses such as the e-learning maturity model (Marshall 2006 ). The aforementioned models are very relevant since they focus on assessing the organizational capabilities for sustainably developing, deploying, and maintaining e-learning. In particular, the eLCMM focuses on assessing the maturity of adopting e-learning systems and adds a feedback building block for improving learners’ experiences (Hammad et al. 2017 ). Our proposed literature review builds on the previously discussed models, lenses, and empirical studies, and it provides a review of research on e-learning capabilities with the aim of enhancing organizational learning in order to complement the findings of the established models and guide future studies.

E-learning systems can be categorized into different types, depending on their functionalities and affordances. One very popular e-learning type is the learning management system (LMS), which includes a virtual classroom and collaboration capabilities and allows the instructor to design and orchestrate a course or a module. An LMS can be either proprietary (e.g., Blackboard) or open source (e.g., Moodle). These two types differ in their features, costs, and the services they provide; for example, proprietary systems prioritize assessment tools for instructors, whereas open-source systems focus more on community development and engagement tools (Alharthi et al. 2019 ). In addition to LMS, e-learning systems can be categorized based on who controls the pace of learning; for example, an institutional learning environment (ILE) is provided by the organization and is usually used for instructor-led courses, while a personal learning environment (PLE) is proposed by the organization and is managed personally (i.e., learner-led courses). Many e-learning systems use a hybrid version of ILE and PLE that allows organizations to have either instructor-led or self-paced courses.

Besides the controlled e-learning systems, organizations have been using environments such as social media (Qi and Chau 2016 ), massive open online courses (MOOCs) (Weinhardt and Sitzmann 2018 ) and other web-based environments (Wang et al. 2011 ) to reinforce their organizational learning potential. These systems have been utilized through different types of technology (e.g., desktop applications, mobile) that leverage the various capabilities offered (e.g., social learning, VR, collaborative systems, smart and intelligent support) to reinforce the learning and knowledge flow potential of the organization. Although there is a growing body of research on e-learning systems for organizational learning due to the increasingly significant role of skills and expertise development in organizations, the role and alignment of the capabilities of the various e-learning systems with the expected competency development remains underexplored.

Organizational Learning

There is a large body of research on the utilization of technologies to improve the process and outcome dimensions of organizational learning (Crossan et al. 1999 ). Most studies have focused on the learning process and on the added value that new technologies can offer by replacing some of the face-to-face processes with virtual processes or by offering new, technology-mediated phases to the process (Menolli et al. 2020 ; Lau 2015 ) highlighted how VR capabilities can enhance organizational learning, describing the new challenges and frameworks needed in order to effectively utilize this potential. In the same vein, Zhang et al. ( 2017 ) described how VR influences reflective thinking and considered its indirect value to overall learning effectiveness. In general, contemporary research has investigated how novel technologies and approaches have been utilized to enhance organizational learning, and it has highlighted both the promises and the limitations of the use of different technologies within organizations.

In many organizations, alignment with the established infrastructure and routines, and adoption by employees are core elements for effective organizational learning (Wang et al. 2011 ). Strict policies, low digital competence, and operational challenges are some of the elements that hinder e-learning adoption by organizations (Garavan et al. 2019 ; Wang 2018 ) demonstrated the importance of organizational, managerial, and job support for utilizing individual and social learning in order to increase the adoption of organizational learning. Other studies have focused on the importance of communication through different social channels to develop understanding of new technology, to overcome the challenges employees face when engaging with new technology, and, thereby, to support organizational learning (Menolli et al. 2020 ). By considering the related work in the area of organizational learning, we identified a gap in aligning an organization’s learning needs with the capabilities offered by the various technologies. Thus, systematic work is needed to review e-learning capabilities and how these capabilities can efficiently support organizational learning.

E-learning Systems to Enhance Organizational Learning

When considering the interplay between e-learning systems and organizational learning, we observed that a major challenge for today’s organizations is to switch from being information-based enterprises to become knowledge-based enterprises (El Kadiri et al. 2016 ). Unidirectional learning flows, such as formal and informal training, are important but not sufficient to cover the needs that enterprises face (Manuti et al. 2015 ). To maintain enterprises’ competitiveness, enterprise staff have to operate in highly intense information and knowledge-oriented environments. Traditional learning approaches fail to substantiate learning flow on the basis of daily evidence and experience. Thus, novel, ubiquitous, and flexible learning mechanisms are needed, placing humans (e.g., employees, managers, civil servants) at the center of the information and learning flow and bridging traditional learning with experiential, social, and smart learning.

Organizations consider lack of skills and competences as being the major knowledge-related factors hampering innovation (El Kadiri et al. 2016 ). Thus, solutions need to be implemented that support informal, day-to-day, and work training (e.g., social learning, collaborative learning, VR/AR solutions) in order to develop individual staff competences and to upgrade the competence affordances at the organizational level. E-learning-enhanced organizational learning has been delivered primarily in the form of web-based learning (El Kadiri et al. 2016 ). More recently, the TEL tools portfolio has rapidly expanded to make more efficient joint use of novel learning concepts, methodologies, and technological enablers to achieve more direct, effective, and lasting learning impacts. Virtual learning environments, mobile-learning solutions, and AR/VR technologies and head-mounted displays have been employed so that trainees are empowered to follow their own training pace, learning topics, and assessment tests that fit their needs (Costello and McNaughton 2018 ; Mueller et al. 2011 ; Muller Queiroz et al. 2018 ). The expanding use of social networking tools has also brought attention to the contribution of social and collaborative learning (Hester et al. 2016 ; Wei and Ram 2016 ).

Contemporary learning systems supporting adaptive, personalized, and collaborative learning expand the tools available in eOL and contribute to the adoption, efficiency, and general prospects of the introduction of TEL in organizations (Cheng et al. 2011 ). In recent years, eOL has emphasized how enterprises share knowledge internally and externally, with particular attention being paid to systems that leverage collaborative learning and social learning functionalities (Qi and Chau 2016 ; Wang  2011 ). This is the essence of computer-supported collaborative learning (CSCL). The CSCL literature has developed a framework that combines individual learning, organizational learning, and collaborative learning, facilitated by establishing adequate learning flows and emerges effective learning in an enterprise learning (Goggins et al. 2013 ), in Fig.  1 .

Representation of the combination of enterprise learning and knowledge flows. (adapted from Goggins et al. 2013 )

Establishing efficient knowledge and learning flows is a primary target for future data-driven enterprises (El Kadiri et al. 2016 ). Given the involved knowledge, the human resources, and the skills required by enterprises, there is a clear need for continuous, flexible, and efficient learning. This can be met by contemporary learning systems and practices that provide high adoption, smooth usage, high satisfaction, and close alignment with the current practices of an enterprise. Because the required competences of an enterprise evolve, the development of competence models needs to be agile and to leverage state-of-the art technologies that align with the organization’s processes and models. Therefore, in this paper we provide a review of the eOL research in order to summarize the findings, identify the various capabilities of eOL, and guide the development of organizational learning in future enterprises as well as in future studies.

Methodology

To answer our research questions, we conducted an SLR, which is a means of evaluating and interpreting all available research relevant to a particular research question, topic area, or phenomenon of interest. A SLR has the capacity to present a fair evaluation of a research topic by using a trustworthy, rigorous, and auditable methodology (Kitchenham and Charters 2007 ). The guidelines used (Kitchenham and Charters 2007 ) were derived from three existing guides adopted by medical researchers. Therefore, we adopted SLR guidelines that follow transparent and widely accepted procedures (especially in the area of software engineering and information systems, as well as in e-learning), minimize potential bias (researchers), and support reproducibility (Kitchenham and Charters 2007 ). Besides the minimization of bias and support for reproducibility, an SLR allows us to provide information about the impact of some phenomenon across a wide range of settings, contexts, and empirical methods. Another important advantage is that, if the selected studies give consistent results, SLRs can provide evidence that the phenomenon is robust and transferable (Kitchenham and Charters 2007 ).

Article Collection

Several procedures were followed to ensure a high-quality review of the literature of eOL. A comprehensive search of peer-reviewed articles was conducted in February 2019 (short papers, posters, dissertations, and reports were excluded), based on a relatively inclusive range of key terms: “organizational learning” & “elearning”, “organizational learning” & “e-learning”, “organisational learning” & “elearning”, and “organisational learning” & “e-learning”. Publications were selected from 2010 onwards, because we identified significant advances since 2010 (e.g., MOOCs, learning analytics, personalized learning) in the area of learning technologies. A wide variety of databases were searched, including SpringerLink, Wiley, ACM Digital Library, IEEE Xplore, Science Direct, SAGE, ERIC, AIS eLibrary, and Taylor & Francis. The selected databases were aligned with the SLR guidelines (Kitchenham and Charters 2007 ) and covered the major venues in IS and educational technology (e.g., a basket of eight IS journals, the top 20 journals in the Google Scholar IS subdiscipline, and the top 20 journals in the Google Scholar Educational Technology subdiscipline). The search process uncovered 2,347 peer-reviewed articles.

Inclusion and Exclusion Criteria

The selection phase determines the overall validity of the literature review, and thus it is important to define specific inclusion and exclusion criteria. As Dybå and Dingsøyr ( 2008 ) specified, the quality criteria should cover three main issues – namely, rigor, credibility, and relevance – that need to be considered when evaluating the quality of the selected studies. We applied eight quality criteria informed by the proposed Critical Appraisal Skills Programme (CASP) and related works (Dybå and Dingsøyr 2008 ). Table ​ Table1 1 presents these criteria.

Quality criteria

Therefore, studies were eligible for inclusion if they were focused on eOL. The aforementioned criteria were applied in stages 2 and 3 of the selection process (see Fig.  2 ), when we assessed the papers based on their titles and abstracts, and read the full papers. From March 2020, we performed an additional search (stage 4) following the same process for papers published after the initial search period (i.e., 2010–February 2019). The additional search returned seven papers. Figure ​ Figure2 2 summarizes the stages of the selection process.

Stages of the selection process

Each collected study was analyzed based on the following elements: study design (e.g., experiment, case study), area (e.g., IT, healthcare), technology (e.g., wiki, social media), population (e.g., managers, employees), sample size, unit of analysis (individual, firm), data collections (e.g., surveys, interviews), research method, data analysis, and the main research objective of the study. It is important to highlight that the articles were coded based on the reported information, that different authors reported information at different levels of granularity (e.g., an online system vs. the name of the system), and that in some cases the information was missing from the paper. Overall, we endeavored to code the articles as accurately and completely as possible.

The coding process was iterative with regular consensus meetings between the two researchers involved. The primary coder prepared the initial coding for a number of articles and both coders reviewed and agreed on the coding in order to reach the final codes presented in the Appendix . Disagreements between the coders and inexplicit aspects of the reviewed papers were discussed and resolved in regular consensus meetings. Although this process did not provide reliability indices (e.g., Cohen’s kappa), it did provide certain reliability in terms of consistency of the coding and what Krippendorff ( 2018 ) stated as the reliability of “the degree to which members of a designated community concur on the readings, interpretations, responses to, or uses of given texts or data”, which is considered acceptable research practice (McDonald et al. 2019 ).

In this section, we present the detailed results of the analysis of the 47 papers. Analysis of the studies was performed using non-statistical methods that considered the variables reported in the Appendix . This section is followed by an analysis and discussion of the categories.

Sample Size and Population Involved

The categories related to the sample of the articles and included the number of participants in each study (size), their position (e.g., managers, employees), and the area/topic covered by the study. The majority of the studies involved employees (29), with few studies involving managers (6), civil servants (2), learning specialists (2), clients, and researchers. Regarding the sample size, approximately half of the studies (20) were conducted with fewer than 100 participants; some (12) can be considered large-scale studies (more than 300 participants); and only a few (9) can be considered small scale (fewer than 20 participants). In relation to the area/topic of the study, most studies (11) were conducted in the context of the IT industry, but there was also good coverage of other important areas (i.e., healthcare, telecommunications, business, public sector). Interestingly, several studies either did not define the area or were implemented in a generic context (sector-agnostic studies, n = 10), and some studies were implemented in a multi-sector context (e.g., participants from different sections or companies, n = 4).

Research Methods

When assessing the status of research for an area, one of the most important aspects is the methodology used. By “method” in the Appendix , we refer to the distinction between quantitative, qualitative, and mixed methods research. In addition to the method, in our categorization protocol we also included “study design” to refer to the distinction between survey studies (i.e., those that gathered data by asking a group of participants), experiments (i.e., those that created situations to record beneficial data), and case studies (i.e., those that closely studied a group of individuals).

Based on this categorization, the Appendix shows that the majority of the papers were quantitative (34) and qualitative (7), with few studies (6) utilizing mixed methods. Regarding the study design, most of the studies were survey studies (26), 13 were case studies, and fewer were experiments (8). For most studies, the individual participant (40) was the unit of analysis, with few studies having the firm as the unit of analysis, and only one study using the training session as a unit of analysis. Regarding the measures used in the studies, most utilized surveys (39), with 11 using interviews, and only a few studies using field notes from focus groups (2) and log files from the systems (2). Only eight studies involved researchers using different measures to triangulate or extend their findings. Most articles used structural equation modeling (SEM) (17) to analyze their data, with 13 studies employing descriptive statistics, seven using content analysis, nine using regression analysis or analyses of variances/covariance, and one study using social network analysis (SNA).

Technologies

Concerning the technology used, most of the studies (17) did not study a specific system, referring instead in their investigation to a generic e-learning or technological solution. Several studies (9) named web-based learning environments, without describing the functionalities of the identified system. Other studies focused on online learning environments (4), collaborative learning systems (3), social learning systems (3), smart learning systems (2), podcasting (2), with the rest of the studies using a specific system (e.g., a wiki, mobile learning, e-portfolios, Second Life, web application).

Research Objectives

The research objectives of the studies could be separated into six main categories. The first category focuses on the intention of the employees to use the technology (9); the second focuses on the performance of the employees (8); the third focuses on the value/outcome for the organization (4); the fourth focuses on the actual usage of the system (7); the fifth focuses on employees’ satisfaction (4); and the sixth focuses on the ability of the proposed system to foster learning (9). In addition to these six categories, we also identified studies that focused on potential barriers for eOL in organizations (Stoffregen et al. 2016 ), the various benefits associated with the successful implementation of eOL (Liu et al. 2012 ), the feasibility of eOL (Kim et al. 2014 ; Mueller et al. 2011 ), and the alignment of the proposed innovation with the other processes and systems in the organization (Costello and McNaughton 2018 ).

E-learning Capabilities in Various Organizations and for Various Objectives

The technology used has an inherent role for both the organization and the expected eOL objective. E-learning systems are categorized based on their functionalities and affordances. Based on the information reported in the selected papers, we ranked them based on the different technologies and functionalities (e.g., collaborative, online, smart). To do so, we focused on the main elements described in the selected paper; for instance, a paper that described the system as wiki-based or indicated that the system was Second Life was ranked as such, rather than being added to collaborative systems or social learning respectively. We did this because we wanted to capture all the available information since it gave us additional insights (e.g., Second Life is both a social and a VR system).

To investigate the connection between the various technologies used to enhance organizational learning and their application in the various organizations, we utilized the coding (see Appendix ) and mapped the various e-learning technologies (or their affordances) with the research industries to which they applied (Fig.  3 ). There was occasionally a lack of detailed information about the capabilities of the e-learning systems applied (e.g., generic, or a web application, or an online system), which limited the insights. Figure ​ Figure3 3 provides a useful mapping of the confluence of e-learning technologies and their application in the various industries.

Association of the different e-learning technologies with the industries to which they are applied in the various studies. Note: The size of the circles depicts the frequency of studies, with the smallest circle representing one study and the largest representing six studies. The mapping is extracted from the data in the Appendix , which outlines the papers that belong in each of the circles

To investigate the connection between the various technologies used to enhance organizational learning and their intended objectives, we utilized the coding of the articles (see Appendix ) and mapped the various e-learning technologies (or their affordances) with the intended objectives, as reported in the various studies (Fig.  4 ). The results in Fig.  4 show the objectives that are central in eOL research (e.g., performance, fostering learning, adoption, and usage) as well as those objectives on which few studies have focused (e.g., alignment, feasibility, behavioral change). In addition, the results also indicate the limited utilization of the various e-learning capabilities (e.g., social, collaborative, smart) to achieve objectives connected with those capabilities (e.g., social learning and behavioral change, collaborative learning, and barriers).

Association of the different e-learning technologies with the objectives investigated in the various studies. Note: The size of the circles depicts the frequency of studies, with the smallest circle representing one study and the largest representing five studies. The mapping is extracted from the data in the Appendix , which outlines the papers that belong in each of the circles

5. Discussion

After reviewing the 47 identified articles in the area of eOL, we can observe that all the works acknowledge the importance of the affordances offered by different e-learning technologies (e.g., remote collaboration, anytime anywhere), the importance of the relationship between eOL and employees’ satisfaction and performance, and the benefits associated with organizational value and outcome. Most of the studies agree that eOL provides employees, managers, and even clients with opportunities to learn in a more differentiated manner, compared to formal and face-to-face learning. However, how the organization adopts and puts into practice these capabilities to leverage them and achieve its goals are complex and challenging procedures that seem to be underexplored.

Several studies (Lee et al. 2015a ; Muller Queiroz et al. 2018 ; Tsai et al. 2010 ) focused on the positive effect of perceived managerial support, perceived usefulness, perceived ease of use, and other technology acceptance model (TAM) constructs of the e-learning system in supporting all three levels of learning (i.e., individual, collaborative, and organizational). Another interesting dimension highlighted by many studies (Choi and Ko 2012 ; Khalili et al. 2012 ; Yanson and Johnson 2016 ) is the role of socialization in the adoption and usage of the e-learning systems that offer these capabilities. Building connections and creating a shared learning space in the e-learning system is challenging but also critical for the learners (Yanson and Johnson 2016 ). This is consistent with the expectancy-theoretical explanation of how social context impacts on employees’ motivation to participate in learning (Lee et al. 2015a ; Muller Queiroz et al. 2018 ).

The organizational learning literature suggests that e-learning may be more appropriate for the acquisition of certain types of knowledge than others (e.g., procedural vs. declarative, or hard-skills vs. soft-skills); however, there is no empirical evidence for this (Yanson and Johnson 2016 ). To advance eOL research, there is a need for a significant move to address complex, strategic skills by including learning and development professionals (Garavan et al. 2019 ) and by developing strategic relationships. Another important element is to utilize e-learning technology that addresses and integrates organizational, individual, and social perspectives in eOL (Wang  2011 ). This is also identified in our literature review since we found only limited specialized e-learning systems in domain areas that have traditionally benefited from such technology. For instance, although there were studies that utilized VR environments (Costello and McNaughton 2018 ; Muller Queiroz et al. 2018 ) and video-based learning systems (Wei et al. 2013 ; Wei and Ram 2016 ), there was limited focus in contemporary eOL research on how specific affordances of the various environments that are used in organizations (e.g., Carnetsoft, Outotec HSC, and Simscale for simulations of working environments; or Raptivity, YouTube, and FStoppers to gain specific skills and how-to knowledge) can benefit the intended goals or be integrated with the unique qualities of the organization (e.g., IT, healthcare).

For the design and the development of the eOL approach, the organization needs to consider the alignment of individual learning needs, organizational objectives, and the necessary resources (Wang  2011 ). To achieve this, it is advisable for organizations to define the expected objectives, catalogue the individual needs, and select technologies that have the capacity to support and enrich learners with self-directed and socially constructed learning practices in the organization (Wang  2011 ). This needs to be done by taking into consideration that on-demand eOL is gradually replacing the classic static eOL curricula and processes (Dignen and Burmeister 2020 ).

Another important dimension of eOL research is the lenses used to approach effectiveness. The selected papers approached effectiveness with various objectives, such as fostering learning, usage of the e-learning system, employees’ performance, and the added organizational value (see Appendix ). To measure these indices, various metrics (quantitative, qualitative, and mixed) have been applied. The qualitative dimensions emphasize employees’ satisfaction and system usage (e.g., Menolli et al. 2020 ; Turi et al. 2019 ), as well as managers’ perceived gained value and benefits (e.g., Lee et al. 2015b ; Xiang et al. 2020 ) and firms’ perceived effective utilization of eOL resources (López-Nicolás and Meroño-Cerdán 2011 ). The quantitative dimensions focus on usage, feasibility, and experience at different levels within an organization, based on interviews, focus groups, and observations (Costello and McNaughton 2018 ; Michalski 2014 ; Stoffregen et al. 2016 ). However, it is not always clear the how eOL effectiveness has been measured, nor the extent to which eOL is well aligned with and is strategically impactful on delivering the strategic agenda of the organization (Garavan et al. 2019 ).

Research on digital technologies is developing rapidly, and big data and business analytics have the potential to pave the way for organizations’ digital transformation and sustainable development (Mikalef et al. 2018 ; Pappas et al. 2018 ); however, our review finds surprisingly limited use of big data and analytics in eOL. Despite contemporary e-learning systems adopting data-driven mechanisms, as well as advances in learning analytics (Siemens and Long 2011 ), the results of our analysis indicate that learner-generated data in the context of eOL are used in only a few studies to extract very limited insights with respect to the effectiveness of eOL and the intended objectives of the respective study (Hung et al. 2015 ; Renner et al. 2020 ; Rober and Cooper 2011 ). Therefore, eOL research needs to focus on data-driven qualities that will allow future researchers to gain deeper insights into which capabilities need to be developed to monitor the effectiveness of the various practices and technologies, their alignment with other functions of the organization, and how eOL can be a strategic and impactful vehicle for materializing the strategic agenda of the organization.

Status of eOL Research

The current review suggests that, while the efficient implementation of eOL entails certain challenges, there is also a great potential for improving employees’ performance as well as overall organizational outcome and value. There are also opportunities for improving organizations’ learning flow, which might not be feasible with formal learning and training. In order to construct the main research dimensions of eOL research and to look more deeply at the research objectives of the studies (the information we coded as objectives in the Appendix ), we performed a content analysis and grouped the research objectives. This enabled us to summarize the contemporary research on eOL according to five major categories, each of which is describes further below. As the research objectives of the published work shows, the research on eOL conducted during the last decade has particularly focused on the following five directions.

Research has particularly focused on how easy the technology is to use, on how useful it is, or on how well aligned/integrated it is with other systems and processes within the organization. In addition, studies have used different learning technologies (e.g., smart, social, personalized) to enhance organizational learning in different contexts and according to different needs. However, most works have focused on affordances such as remote training and the development of static courses or modules to share information with learners. Although a few studies have utilized contemporary e-learning systems (see Appendix ), even in these studies there is a lack of alignment between the capabilities of those systems (e.g., open online course, adaptive support, social and collaborative learning) and the objectives and strategy of the organization (e.g., organizational value, fostering learning).

The reviewed work has emphasized how different factors contribute to different levels of organizational learning, and it has focused on practices that address individual, collaborative, and organizational learning within the structure of the organization. In particular, most of the reviewed studies recognize that organizational learning occurs at multiple levels: individual, team (or group), and organization. In other words, although each of the studies carried out an investigation within a given level (except for Garavan et al. 2019 ), there is a recognition and discussion of the different levels. Therefore, the results align with the 4I framework of organizational learning that recognizes how learning across the different levels is linked by social and psychological processes: intuiting, interpreting, integrating, and institutionalizing (the 4Is) (Crossan et al. 1999 ). However, most of the studies focused on the institutionalizing-intuiting link (i.e., top-down feedback); moreover, no studies focused on contemporary learning technologies and processes that strengthen the learning flow (e.g., self-regulated learning).

There is a considerable amount of predominantly qualitative studies that focus on potential barriers to eOL implementation as well as on the risks and requirements associated with the feasibility and successful implementation of eOL. In the same vein, research has emphasized the importance of alignment of eOL (both in processes and in technologies) within the organization. These critical aspects for effective eOL are sometimes the main objectives of the studies (see Appendix ). However, most of the elements relating to the effectiveness of eOL were measured with questionnaires and interviews with employees and managers, and very little work was conducted on how to leverage the digital technologies employed in eOL, big data, and analytics in order to monitor the effectiveness of eOL.

In most of the studies, the main objective was to increase employees’ adoption, satisfaction, and usage of the e-learning system. In addition, several studies focused on the e-learning system’s ability to improve employees’ performance, increase the knowledge flow in the organization, and foster learning. Most of the approaches were employee-centric, with a small amount of studies focusing on managers and the firm in general. However, employees were seen as static entities within the organization, with limited work investigating how eOL-based training exposes employees to new knowledge, broadens their skills repertoire, and has tremendous potential for fostering innovation (Lin and Sanders 2017 ).

A considerable number of studies utilized the firm (rather than the individual employee) as the unit of analysis. Such studies focused on how the implementation of eOL can increase employee performance, organizational value, and customer value. Although this is extremely helpful in furthering knowledge about eOL technologies and practices, a more granular investigation of the different e-learning systems and processes to address the various goals and strategies of the organization would enable researchers to extract practical insights on the design and implementation of eOL.

Research Agenda

By conducting an SLR and documenting the eOL research of the last decade, we have identified promising themes of research that have the potential to further eOL research and practice. To do so, we define a research agenda consisting of five thematic areas of research, as depicted in the research framework in Fig.  5 , and we provide some suggestions on how researchers could approach these challenges. In this visualization of the framework, on the left side we present the organizations as they were identified from our review (i.e., area/topic category in the Appendix ) and the multiple levels where organizational learning occurs (Costello and McNaughton 2018 ). On the right side, we summarize the objectives as they were identified from our review (i.e., the objectives category in the Appendix ). In the middle, we depict the orchestration that was conducted and how potential future research on eOL can improve the orchestration of the various elements and accelerate the achievement of the intended objectives. In particular, our proposed research agenda includes five research themes discussed in the following subsections.

E-learning capabilities to enhance organizational research agenda

Theme 1: Couple E-learning Capabilities With the Intended Goals

The majority of the eOL studies either investigated a generic e-learning system using the umbrella term “e-learning” or did not provide enough details about the functionalities of the system (in most cases, it was simply defined as an online or web system). This indicates the very limited focus of the eOL research on the various capabilities of e-learning systems. In other words, the literature has been very detailed on the organizational value and employees’ acceptance of the technology, but less detailed on the capabilities of this technology that needs to be put into place to achieve the intended goals and strategic agenda. However, the capabilities of the e-learning systems and their use are not one-size-fits-all, and the intended goals (to obtain certain skills and competences) and employees’ needs and backgrounds play a determining role in the selection of the e-learning system (Al-Fraihat et al. 2020 ).

Only in a very few studies (Mueller et al. 2011 ; Renner et al. 2020 ) were the capabilities of the e-learning solutions (e.g., mobile learning, VR) utilized, and the results were found to significantly contribute to the intended goals. The intended knowledge can be procedural, declarative, general competence (e.g., presentation, communication, or leadership skills) or else, and its particularities and the pedagogical needs of the intended knowledge (e.g., a need for summative/formative feedback or for social learning support) should guide the selection of the e-learning system and the respective capabilities. Therefore, future research needs to investigate how the various capabilities offered by contemporary learning systems (e.g., assessment mechanisms, social learning, collaborative learning, personalized learning) can be utilized to adequately reinforce the intended goals (e.g., to train personnel to use a new tool, to improve presentation skills).

Theme 2: Embrace the Particularities of the Various Industries

Organizational learning entails sharing knowledge and enabling opportunities for growth at the individual, group, team, and organizational levels. Contemporary e-learning systems provide the medium to substantiate the necessary knowledge flow within organizations and to support employees’ overall learning. From the selected studies, we can infer that eOL research is either conducted in an industry-agnostic context (either generic or it was not properly reported) or there is a focus on the IT industry (see Appendix ). However, when looking at the few studies that provide results from different industries (Garavan et al. 2019 ; Lee et al. 2014 ), companies indicate that there are different practices, processes, and expectations, and that employees have different needs and perceptions with regards to e-learning systems and eOL in general. Such particularities influence the perceived dimensions of a learning organization. Some industries noted that eOL promoted the development of their learning organizations, whereas others reported that eOL did not seem to contribute to their development as a learning organization (Yoo and Huang 2016 ). Therefore, it is important that the implementation of organizational learning embraces the particularities of the various industries and future research needs to identify how the industry-specific characteristics can inform the design and development of organizational learning in promoting an organization’s goals and agenda.

Theme 3: Utilize E-learning Capabilities to Implement Employee-centric Approaches

For efficient organizational learning to be implemented, the processes and technologies need to recognize that learning is linked by social and psychological processes (Crossan et al. 1999 ). This allows employees to develop learning in various forms (e.g., social, emotional, personalized) and to develop elements such as self-awareness, self-control, and interpersonal skills that are vital for the organization. Looking at the contemporary eOL research, we notice that the exploration of e-learning capabilities to nurture the aforementioned elements and support employee-centric approaches is very limited (e.g., personalized technologies, adaptive assessment). Therefore, future research needs to collect data to understand how e-learning capabilities can be utilized in relation to employees’ needs and perceptions in order to provide solutions (e.g., collaborative, social, adaptive) that are employee-centric and focused on development, and that have the potential to move away from standard one-size-fits-all e-learning solutions to personalized and customized systems and processes.

Theme 4: Employ Analytics-enabled eOL

There is a lot of emphasis on measuring, via various qualitative and quantitative metrics, the effectiveness of eOL implemented at different levels in organizations. However, most of these metrics come from surveys and interviews that capture employees’ and managers’ perceptions of various aspects of eOL (e.g., fostering of learning, organizational value, employees’ performance), and very few studies utilize analytics (Hung et al. 2015 ; Renner et al. 2020 ; Rober and Cooper 2011 ). Given how digital technologies, big data, and business analytics pave the way towards organizations’ digital transformation and sustainable development (Mikalef et al. 2018 ; Pappas et al. 2018 ), and considering the learning analytics affordances of contemporary e-learning systems (Siemens and Long 2011 ), future work needs to investigate how learner/employee-generated data can be employed to inform practice and devise more accurate and temporal effectiveness metrics when measuring the importance and impact of eOL.

Theme 5: Orchestrate the Employees’ Needs, Resources, and Objectives in eOL Implementation

While considerable effort has been directed towards the various building blocks of eOL implementation, such as resources (intangible, tangible, and human skills) and employees’ needs (e.g., vision, growth, skills development), little is known so far about the processes and structures necessary for orchestrating those elements in order to achieve an organization’s intended goals and to materialize its overall agenda. In other words, eOL research has been very detailed on some of the elements that constitute efficient eOL, but less so on the interplay of those elements and how they need to be put into place. Prior literature on strategic resource planning has shown that competence in orchestrating such elements is a prerequisite to successfully increasing business value (Wang et al. 2012 ). Therefore, future research should not only investigate each of these elements in silos, but also consider their interplay, since it is likely that organizations with similar resources will exert highly varied levels in each of these elements (e.g., analytics-enabled, e-learning capabilities) to successfully materialize their goals (e.g., increase value, improve the competence base of their employees, modernize their organization).

Implications

Several implications for eOL have been revealed in this literature review. First, most studies agree that employees’ or trainees’ experience is extremely important for the successful implementation of eOL. Thus, keeping them in the design and implementation cycle of eOL will increase eOL adoption and satisfaction as well as reduce the risks and barriers. Another important implication addressed by some studies relates to the capabilities of the e-learning technologies, with easy-to-use, useful, and social technologies resulting in more efficient eOL (e.g., higher adoption and performance). Thus, it is important for organizations to incorporate these functionalities in the platform and reinforce them with appropriate content and support. This should not only benefit learning outcomes, but also provide the networking opportunities for employees to broaden their personal networks, which are often lost when companies move from face-to-face formal training to e-learning-enabled organizational learning.

Limitations

This review has some limitations. First, we had to make some methodological decisions (e.g., selection of databases, the search query) that might lead to certain biases in the results. However, tried to avoid such biases by considering all the major databases and following the steps indicated by Kitchenham and Charters ( 2007 ). Second, the selection of empirical studies and coding of the papers might pose another possible bias. However, the focus was clearly on the empirical evidence, the terminology employed (“e-learning”) is an umbrella term that covers the majority of the work in the area, and the coding of papers was checked by two researchers. Third, some elements of the papers were not described accurately, leading to some missing information in the coding of the papers. However, the amount of missing information was very small and could not affect the results significantly. Finally, we acknowledge that the selected methodology (Kitchenham and Charters 2007 ) includes potential biases (e.g., false negatives and false positives), and that different, equally valid methods (e.g., Okoli and Schabram 2010 ) might have been used and have resulted in slightly different outcomes. Nevertheless, despite the limitations of the selected methodology, it is a well-accepted and widely used literature review method in both software engineering and information systems (Boell and Cecez-Kecmanovic 2014 ), providing certain assurance of the results.

Conclusions and Future Work

We have presented an SLR of 47 contributions in the field of eOL over the last decade. With respect to RQ1, we analyzed the papers from different perspectives, such as research methodology, technology, industries, employees, and intended outcomes in terms of organizational value, employees’ performance, usage, and behavioral change. The detailed landscape is depicted in the Appendix and Figs.  3 and ​ and4; 4 ; with the results indicating the limited utilization of the various e-learning capabilities (e.g., social, collaborative) to achieve objectives connected with those capabilities (e.g., social learning and behavioral change, collaborative learning and overcoming barriers).

With respect to RQ2, we categorized the main findings of the selected papers into five areas that reflect the status of eOL research, and we have discussed the challenges and opportunities emerging from the current review. In addition, we have synthesized the extracted challenges and opportunities and proposed a research agenda consisting of five elements that provide suggestions on how researchers could approach these challenges and exploit the opportunities. Such an agenda will strengthen how e-learning can be leveraged to enhance the process of improving actions through better knowledge and understanding in an organization.

A number of suggestions for further research have emerged from reviewing prior and ongoing work on eOL. One recommendation for future researchers is to clearly describe the eOL approach by providing detailed information about the technologies and materials used, as well as the organizations. This will allow meta-analyses to be conducted and it will also identify the potential effects of a firm’s size or area on the performance and other aspects relating to organizational value. Future work should also focus on collecting and triangulating different types of data from different sources (e.g., systems’ logs). The reviewed studies were conducted mainly by using survey data, and they made limited use of data coming from the platforms; thus, the interpretations and triangulation between the different types of collected data were limited.

Biographies

is a Professor of Interaction Design and Learning Technologies at the Department of Computer Science of NTNU, and Head of the Learner-Computer Interaction lab (https://lci.idi.ntnu.no/). His research focuses on the design and study of emerging technologies in online and hybrid education settings, and their connections to student and instructor experiences and practices. Giannakos has co-authored more than 150 manuscripts published in peer-reviewed journals and conferences (including Computers & Education, Computers in Human Behavior, IEEE TLT, Behaviour & Information Technology, BJET, ACM TOCE, CSCL, Interact, C&C, IDC to mention few) and has served as an evaluator for the EC and the US-NSF. He has served/serves in various organization committees (e.g., general chair, associate chair), program committees as well as editor and guest editor on highly recognized journals (e.g., BJET, Computers in Human Behavior, IEEE TOE, IEEE TLT, ACM TOCE). He has worked at several research projects funded by diverse sources like the EC, Microsoft Research, The Research Council of Norway (RCN), US-NSF, the German agency for international academic cooperation (DAAD) and Cheng Endowment; Giannakos is also a recipient of a Marie Curie/ERCIM fellowship, the Norwegian Young Research Talent award and he is one of the outstanding academic fellows of NTNU (2017-2021).

is an Associate Professor in Data Science and Information Systems at the Department of Computer Science. In the past, he has been a Marie Skłodowska-Curie post-doctoral research fellow working on the research project “Competitive Advantage for the Data-driven Enterprise” (CADENT). He received his B.Sc. in Informatics from the Ionian University, his M.Sc. in Business Informatics for Utrecht University, and his Ph.D. in IT Strategy from the Ionian University. His research interests focus the on strategic use of information systems and IT-business value in turbulent environments. He has published work in international conferences and peer-reviewed journals including the Journal of Business Research, British Journal of Management, Information and Management, Industrial Management & Data Systems, and Information Systems and e-Business Management.

Ilias O. Pappasis

an Associate Professor of Information Systems at the Department of Information Systems, University of Agder (UiA), Norway. His research and teaching activities include data science and digital transformation, social innovation and social change, user experience in different contexts,as well as digital marketing, e-services, and information technology adoption. He has published articles in peer reviewed journals and conferences including Journal of Business Research, European Journal of Marketing, Computers in Human Behavior, Information & Management, Psychology & Marketing, International Journal of Information Management, Journal of Systems and Software. Pappas has been a Guest Editor for the journals Information & Management, Technological Forecasting and Social Change, Information Systems Frontiers, Information Technology & People, and Information Systems and e-Business Management. Pappas is a recipient of ERCIM and Marie Skłodowska-Curiefellowships.

Survey = survey study; Exp. = experiment; CaseSt = case study; ND = non-defined; MGM = management; Telec. = telecommunication; Bsn = business; Univ. = university; Cons. = consulting; Public = public sector; Ent. = enterprise; Web = Web-based; KRS = knowledge repository system; OERs = open educational resources; SL = Second Life, Mg, = managers; Empl = employees; Stud = students; Res. = researchers; Learn. = learning specialists; Individ. = individual; Surv. = surveys; Int. = interviews; FG = focus groups; Log = log files; Obs. = observations; Reg. = regression analysis; Descr. = descriptive statistics; A-VA = analysis of variances/covariance; CA = content analysis; ItU = intention to use; Sat. = satisfaction; OV = organizational value; Per. = performance; Flearn = foster learning; Benef. = benefits; Align. = alignment; Feas. = feasibility; Barr. = barriers; Beh. = behavioral change

Open Access funding provided by NTNU Norwegian University of Science and Technology (incl St. Olavs Hospital - Trondheim University Hospital).

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Contributor Information

Michail N. Giannakos, Email: on.untn@gliahcim .

Patrick Mikalef, Email: [email protected] .

Ilias O. Pappas, Email: [email protected] .

• Al-Fraihat D, Joy M, Sinclair J. Evaluating E-learning systems success: An empirical study. Computers in Human Behavior. 2020; 102 :67–86. doi: 10.1016/j.chb.2019.08.004. [ CrossRef ] [ Google Scholar ]
• Alharthi AD, Spichkova M, Hamilton M. Sustainability requirements for eLearning systems: A systematic literature review and analysis. Requirements Engineering. 2019; 24 (4):523–543. doi: 10.1007/s00766-018-0299-9. [ CrossRef ] [ Google Scholar ]
• Alsabawy AY, Cater-Steel A, Soar J. IT infrastructure services as a requirement for e-learning system success. Computers & Education. 2013; 69 :431–451. doi: 10.1016/j.compedu.2013.07.035. [ CrossRef ] [ Google Scholar ]
• Antonacopoulou E, Chiva R. The social complexity of organizational learning: The dynamics of learning and organizing. Management Learning. 2007; 38 :277–295. doi: 10.1177/1350507607079029. [ CrossRef ] [ Google Scholar ]
• Berends H, Lammers I. Explaining discontinuity in organizational learning: A process analysis. Organization Studies. 2010; 31 (8):1045–1068. doi: 10.1177/0170840610376140. [ CrossRef ] [ Google Scholar ]
• Boell SK, Cecez-Kecmanovic D. A hermeneutic approach for conducting literature reviews and literature searches. Communications of the Association for Information Systems. 2014; 34 (1):12. [ Google Scholar ]
• Bologa R, Lupu AR. Organizational learning networks that can increase the productivity of IT consulting companies. A case study for ERP consultants. Expert Systems with Applications. 2014; 41 (1):126–136. doi: 10.1016/j.eswa.2013.07.016. [ CrossRef ] [ Google Scholar ]
• Broadbent J. Comparing online and blended learner’s self-regulated learning strategies and academic performance. The Internet and Higher Education. 2017; 33 :24–32. doi: 10.1016/j.iheduc.2017.01.004. [ CrossRef ] [ Google Scholar ]
• Cheng B, Wang M, Moormann J, Olaniran BA, Chen NS. The effects of organizational learning environment factors on e-learning acceptance. Computers & Education. 2012; 58 (3):885–899. doi: 10.1016/j.compedu.2011.10.014. [ CrossRef ] [ Google Scholar ]
• Cheng B, Wang M, Yang SJ, Peng J. Acceptance of competency-based workplace e-learning systems: Effects of individual and peer learning support. Computers & Education. 2011; 57 (1):1317–1333. doi: 10.1016/j.compedu.2011.01.018. [ CrossRef ] [ Google Scholar ]
• Choi S, Ko I. Leveraging electronic collaboration to promote interorganizational learning. International Journal of Information Management. 2012; 32 (6):550–559. doi: 10.1016/j.ijinfomgt.2012.03.002. [ CrossRef ] [ Google Scholar ]
• Costello JT, McNaughton RB. Integrating a dynamic capabilities framework into workplace e-learning process evaluations. Knowledge and Process Management. 2018; 25 (2):108–125. doi: 10.1002/kpm.1565. [ CrossRef ] [ Google Scholar ]
• Crossan MM, Lane HW, White RE. An organizational learning framework: From intuition to institution. Academy of Management Review. 1999; 24 :522–537. doi: 10.2307/259140. [ CrossRef ] [ Google Scholar ]
• Crossan MM, Maurer CC, White RE. Reflections on the 2009 AMR decade award: Do we have a theory of organizational learning? Academy of Management Review. 2011; 36 (3):446–460. doi: 10.5465/amr.2010.0544. [ CrossRef ] [ Google Scholar ]
• Dee, J., & Leisyte, L. (2017). Knowledge sharing and organizational change in higher education. The Learning Organization, 24 (5), 355–365. 10.1108/TLO-04-2017-0034
• Dignen B, Burmeister T. Three pillars of organization and leadership in disruptive times. Cham: Springer; 2020. Learning and development in the organizations of the future; pp. 207–232. [ Google Scholar ]
• Dybå T, Dingsøyr T. Empirical studies of agile software development: A systematic review. Information and Software Technology. 2008; 50 (9–10):833–859. doi: 10.1016/j.infsof.2008.01.006. [ CrossRef ] [ Google Scholar ]
• El Kadiri S, Grabot B, Thoben KD, Hribernik K, Emmanouilidis C, Von Cieminski G, Kiritsis D. Current trends on ICT technologies for enterprise information systems. Computers in Industry. 2016; 79 :14–33. doi: 10.1016/j.compind.2015.06.008. [ CrossRef ] [ Google Scholar ]
• Engeström Y, Kerosuo H, Kajamaa A. Beyond discontinuity: Expansive organizational learning remembered. Management Learning. 2007; 38 (3):319–336. doi: 10.1177/1350507607079032. [ CrossRef ] [ Google Scholar ]
• Gal E, Nachmias R. Online learning and performance support in organizational environments using performance support platforms. Performance Improvement. 2011; 50 (8):25–32. doi: 10.1002/pfi.20238. [ CrossRef ] [ Google Scholar ]
• Garavan TN, Heneghan S, O’Brien F, Gubbins C, Lai Y, Carbery R, Grant K. L&D professionals in organisations: much ambition, unfilled promise. European Journal of Training and Development. 2019; 44 (1):1–86. [ Google Scholar ]
• Goggins SP, Jahnke I, Wulf V. Computer-supported collaborative learning at the workplace. New York: Springer; 2013. [ Google Scholar ]
• Hammad, R., Odeh, M., & Khan, Z. (2017). ELCMM: An e-learning capability maturity model. In Proceedings of the 15th International Conference (e-Society 2017) (pp. 169–178).
• Hester AJ, Hutchins HM, Burke-Smalley LA. Web 2.0 and transfer: Trainers’ use of technology to support employees’ learning transfer on the job. Performance Improvement Quarterly. 2016; 29 (3):231–255. doi: 10.1002/piq.21225. [ CrossRef ] [ Google Scholar ]
• Hung YH, Lin CF, Chang RI. Developing a dynamic inference expert system to support individual learning at work. British Journal of Educational Technology. 2015; 46 (6):1378–1391. doi: 10.1111/bjet.12214. [ CrossRef ] [ Google Scholar ]
• Iris R, Vikas A. E-Learning technologies: A key to dynamic capabilities. Computers in Human Behavior. 2011; 27 (5):1868–1874. doi: 10.1016/j.chb.2011.04.010. [ CrossRef ] [ Google Scholar ]
• Jia H, Wang M, Ran W, Yang SJ, Liao J, Chiu DK. Design of a performance-oriented workplace e-learning system using ontology. Expert Systems with Applications. 2011; 38 (4):3372–3382. doi: 10.1016/j.eswa.2010.08.122. [ CrossRef ] [ Google Scholar ]
• Joo YJ, Lim KY, Park SY. Investigating the structural relationships among organisational support, learning flow, learners’ satisfaction and learning transfer in corporate e-learning. British Journal of Educational Technology. 2011; 42 (6):973–984. doi: 10.1111/j.1467-8535.2010.01116.x. [ CrossRef ] [ Google Scholar ]
• Kaschig, A., Maier, R., Sandow, A., Lazoi, M., Barnes, S. A., Bimrose, J., … Schmidt, A. (2010). Knowledge maturing activities and practices fostering organisational learning: results of an empirical study. In European Conference on Technology Enhanced Learning (pp. 151–166). Berlin: Springer.
• Khalili A, Auer S, Tarasowa D, Ermilov I. International Conference on Knowledge Engineering and Knowledge Management. Berlin: Springer; 2012. SlideWiki: Elicitation and sharing of corporate knowledge using presentations; pp. 302–316. [ Google Scholar ]
• Khandakar MSA, Pangil F. Relationship between human resource management practices and informal workplace learning. Journal of Workplace Learning. 2019; 31 (8):551–576. doi: 10.1108/JWL-04-2019-0049. [ CrossRef ] [ Google Scholar ]
• Kim MK, Kim SM, Bilir MK. Investigation of the dimensions of workplace learning environments (WLEs): Development of the WLE measure. Performance Improvement Quarterly. 2014; 27 (2):35–57. doi: 10.1002/piq.21170. [ CrossRef ] [ Google Scholar ]
• Kitchenham, B., & Charters, S. (2007). Guidelines for performing systematic literature reviews in software engineering. Technical Report EBSE-2007-01, 2007 . https://citeseerx.ist.psu.edu/viewdoc/download;jsessionid=35909B1B280E2032BF116BDC9DCB71EA? .
• Krippendorff, K. (2018). Content analysis: an introduction to its methodology. Thousand Oaks: Sage Publications.
• Lai HJ. Examining civil servants’ decisions to use Web 2.0 tools for learning, based on the decomposed theory of planned behavior. Interactive Learning Environments. 2017; 25 (3):295–305. doi: 10.1080/10494820.2015.1121879. [ CrossRef ] [ Google Scholar ]
• Lau K. Organizational learning goes virtual? A study of employees’ learning achievement in stereoscopic 3D virtual reality. The Learning Organization. 2015; 22 (5):289–303. doi: 10.1108/TLO-11-2014-0063. [ CrossRef ] [ Google Scholar ]
• Lee J, Choi M, Lee H. Factors affecting smart learning adoption in workplaces: Comparing large enterprises and SMEs. Information Technology and Management. 2015; 16 (4):291–302. doi: 10.1007/s10799-014-0201-5. [ CrossRef ] [ Google Scholar ]
• Lee J, Kim DW, Zo H. Conjoint analysis on preferences of HRD managers and employees for effective implementation of m-learning: The case of South Korea. Telematics and Informatics. 2015; 32 (4):940–948. doi: 10.1016/j.tele.2015.04.010. [ CrossRef ] [ Google Scholar ]
• Lee J, Zo H, Lee H. Smart learning adoption in employees and HRD managers. British Journal of Educational Technology. 2014; 45 (6):1082–1096. doi: 10.1111/bjet.12210. [ CrossRef ] [ Google Scholar ]
• Lin CH, Sanders K. HRM and innovation: A multi-level organizational learning perspective. Human Resource Management Journal. 2017; 27 (2):300–317. doi: 10.1111/1748-8583.12127. [ CrossRef ] [ Google Scholar ]
• Lin CY, Huang CK, Zhang H. Enhancing employee job satisfaction via e-learning: The mediating role of an organizational learning culture. International Journal of Human–Computer Interaction. 2019; 35 (7):584–595. doi: 10.1080/10447318.2018.1480694. [ CrossRef ] [ Google Scholar ]
• Liu YC, Huang YA, Lin C. Organizational factors’ effects on the success of e-learning systems and organizational benefits: An empirical study in Taiwan. The International Review of Research in Open and Distributed Learning. 2012; 13 (4):130–151. doi: 10.19173/irrodl.v13i4.1203. [ CrossRef ] [ Google Scholar ]
• López-Nicolás C, Meroño-Cerdán ÁL. Strategic knowledge management, innovation and performance. International Journal of Information Management. 2011; 31 (6):502–509. doi: 10.1016/j.ijinfomgt.2011.02.003. [ CrossRef ] [ Google Scholar ]
• Manuti A, Pastore S, Scardigno AF, Giancaspro ML, Morciano D. Formal and informal learning in the workplace: A research review. International Journal of Training and Development. 2015; 19 (1):1–17. doi: 10.1111/ijtd.12044. [ CrossRef ] [ Google Scholar ]
• March JG. Exploration and exploitation in organizational learning. Organization Science. 1991; 2 (1):71–87. doi: 10.1287/orsc.2.1.71. [ CrossRef ] [ Google Scholar ]
• Marshall S. New Zealand Tertiary Institution E-learning Capability: Informing and Guiding eLearning Architectural Change and Development. Report to the ministry of education. NZ: Victoria University of Wellington; 2006. [ Google Scholar ]
• McDonald, N., Schoenebeck, S., & Forte, A. (2019). Reliability and inter-rater reliability in qualitative research: Norms and guidelines for CSCW and HCI practice. In Proceedings of the ACM on Human–Computer Interaction, 3(CSCW) (pp. 1–23).
• Menolli, A., Tirone, H., Reinehr, S., & Malucelli, A. (2020). Identifying organisational learning needs: An approach to the semi-automatic creation of course structures for software companies. Behaviour & Information Technology, 39 (11), 1140–1155.
• Michalski MP. Symbolic meanings and e-learning in the workplace: The case of an intranet-based training tool. Management Learning. 2014; 45 (2):145–166. doi: 10.1177/1350507612468419. [ CrossRef ] [ Google Scholar ]
• Mikalef P, Pappas IO, Krogstie J, Giannakos M. Big data analytics capabilities: A systematic literature review and research agenda. Information Systems and e-Business Management. 2018; 16 (3):547–578. doi: 10.1007/s10257-017-0362-y. [ CrossRef ] [ Google Scholar ]
• Mitić S, Nikolić M, Jankov J, Vukonjanski J, Terek E. The impact of information technologies on communication satisfaction and organizational learning in companies in Serbia. Computers in Human Behavior. 2017; 76 :87–101. doi: 10.1016/j.chb.2017.07.012. [ CrossRef ] [ Google Scholar ]
• Mueller J, Hutter K, Fueller J, Matzler K. Virtual worlds as knowledge management platform—A practice-perspective. Information Systems Journal. 2011; 21 (6):479–501. doi: 10.1111/j.1365-2575.2010.00366.x. [ CrossRef ] [ Google Scholar ]
• Muller Queiroz, A. C., Nascimento, M., Tori, A., Alejandro, R. Brashear, Veloso, T., de Melo, V., de Souza Meirelles, F., & da Silva Leme, M. I. (2018). Immersive virtual environments in corporate education and training. In AMCIS. https://aisel.aisnet.org/amcis2018/Education/Presentations/12/ .
• Navimipour NJ, Zareie B. A model for assessing the impact of e-learning systems on employees’ satisfaction. Computers in Human Behavior. 2015; 53 :475–485. doi: 10.1016/j.chb.2015.07.026. [ CrossRef ] [ Google Scholar ]
• Oh SY. Effects of organizational learning on performance: The moderating roles of trust in leaders and organizational justice. Journal of Knowledge Management. 2019; 23 :313–331. doi: 10.1108/JKM-02-2018-0087. [ CrossRef ] [ Google Scholar ]
• Okoli, C., & Schabram, K. (2010). A guide to conducting a systematic literature review of information systems research. Sprouts: Working Papers on Information Systems, 10 (26), 1–46.
• Pappas IO, Mikalef P, Giannakos MN, Krogstie J, Lekakos G. Big data and business analytics ecosystems: paving the way towards digital transformation and sustainable societies. Information Systems and e-Business Management. 2018; 16 :479–491. doi: 10.1007/s10257-018-0377-z. [ CrossRef ] [ Google Scholar ]
• Popova-Nowak IV, Cseh M. The meaning of organizational learning: A meta-paradigm perspective. Human Resource Development Review. 2015; 14 (3):299–331. doi: 10.1177/1534484315596856. [ CrossRef ] [ Google Scholar ]
• Qi, C., & Chau, P. Y. (2016). An empirical study of the effect of enterprise social media usage on organizational learning. In Pacific Asia Conference on Information Systems (PACIS'16). Proceedings , Paper 330. http://aisel.aisnet.org/pacis2016/330 .
• Renner B, Wesiak G, Pammer-Schindler V, Prilla M, Müller L, Morosini D, Cress U. Computer-supported reflective learning: How apps can foster reflection at work. Behaviour & Information Technology. 2020; 39 (2):167–187. doi: 10.1080/0144929X.2019.1595726. [ CrossRef ] [ Google Scholar ]
• Rober, M. B., & Cooper, L. P. (2011, January). Capturing knowledge via an” Intrapedia”: A case study. In 2011 44th Hawaii International Conference on System Sciences (pp. 1–10). New York: IEEE.
• Rosenberg MJ, Foshay R. E-learning: Strategies for delivering knowledge in the digital age. Performance Improvement. 2002; 41 (5):50–51. doi: 10.1002/pfi.4140410512. [ CrossRef ] [ Google Scholar ]
• Serrano, Á., Marchiori, E. J., del Blanco, Á., Torrente, J., & Fernández-Manjón, B. (2012). A framework to improve evaluation in educational games. The IEEE Global Engineering Education Conference (pp. 1–8). Marrakesh, Morocco.
• Siadaty, M., Jovanović, J., Gašević, D., Jeremić, Z., & Holocher-Ertl, T. (2010). Leveraging semantic technologies for harmonization of individual and organizational learning. In European Conference on Technology Enhanced Learning (pp. 340–356). Berlin: Springer.
• Siemens G, Long P. Penetrating the fog: Analytics in learning and education. EDUCAUSE Review. 2011; 46 (5):30. [ Google Scholar ]
• Škerlavaj M, Dimovski V, Mrvar A, Pahor M. Intra-organizational learning networks within knowledge-intensive learning environments. Interactive Learning Environments. 2010; 18 (1):39–63. doi: 10.1080/10494820802190374. [ CrossRef ] [ Google Scholar ]
• Smith PJ, Sadler-Smith E. Learning in organizations: Complexities and diversities. London: Routledge; 2006. [ Google Scholar ]
• Stoffregen JD, Pawlowski JM, Ras E, Tobias E, Šćepanović S, Fitzpatrick D, Friedrich H. Barriers to open e-learning in public administrations: A comparative case study of the European countries Luxembourg, Germany, Montenegro and Ireland. Technological Forecasting and Social Change. 2016; 111 :198–208. doi: 10.1016/j.techfore.2016.06.030. [ CrossRef ] [ Google Scholar ]
• Subramaniam, R., & Nakkeeran, S. (2019). Impact of corporate e-learning systems in enhancing the team performance in virtual software teams. In Smart Technologies and Innovation for a Sustainable Future (pp. 195–204). Berlin: Springer.
• Tsai CH, Zhu DS, Ho BCT, Wu DD. The effect of reducing risk and improving personal motivation on the adoption of knowledge repository system. Technological Forecasting and Social Change. 2010; 77 (6):840–856. doi: 10.1016/j.techfore.2010.01.011. [ CrossRef ] [ Google Scholar ]
• Turi JA, Sorooshian S, Javed Y. Impact of the cognitive learning factors on sustainable organizational development. Heliyon. 2019; 5 (9):e02398. doi: 10.1016/j.heliyon.2019.e02398. [ PMC free article ] [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Wang, M. (2011). Integrating organizational, social, and individual perspectives in Web 2.0-based workplace e-learning. Information Systems Frontiers, 13 (2), 191–205.
• Wang, M. (2018). Effects of individual and social learning support on employees’ acceptance of performance-oriented e-learning. In E-Learning in the Workplace (pp. 141–159). Springer. 10.1007/978-3-319-64532-2_13.
• Wang M, Ran W, Liao J, Yang SJ. A performance-oriented approach to e-learning in the workplace. Journal of Educational Technology & Society. 2010; 13 (4):167–179. [ Google Scholar ]
• Wang M, Vogel D, Ran W. Creating a performance-oriented e-learning environment: A design science approach. Information & Management. 2011; 48 (7):260–269. doi: 10.1016/j.im.2011.06.003. [ CrossRef ] [ Google Scholar ]
• Wang N, Liang H, Zhong W, Xue Y, Xiao J. Resource structuring or capability building? An empirical study of the business value of information technology. Journal of Management Information Systems. 2012; 29 (2):325–367. doi: 10.2753/MIS0742-1222290211. [ CrossRef ] [ Google Scholar ]
• Wang S, Wang H. Organizational schemata of e-portfolios for fostering higher-order thinking. Information Systems Frontiers. 2012; 14 (2):395–407. doi: 10.1007/s10796-010-9262-0. [ CrossRef ] [ Google Scholar ]
• Wei K, Ram J. Perceived usefulness of podcasting in organizational learning: The role of information characteristics. Computers in Human Behavior. 2016; 64 :859–870. doi: 10.1016/j.chb.2016.08.003. [ CrossRef ] [ Google Scholar ]
• Wei, K., Sun, H., & Li, H. (2013). On the driving forces of diffusion of podcasting in organizational settings: A case study and propositions. In PACIS 2013. Proceedings , 217. http://aisel.aisnet.org/pacis2013/217 .
• Weinhardt, J. M., & Sitzmann, T. (2018). Revolutionizing training and education? Three questions regarding massive open online courses (MOOCs). Human Resource Management Review, 29 (2), 218–225.
• Xiang, Q., Zhang, J., & Liu, H. (2020). Organisational improvisation as a path to new opportunity identification for incumbent firms: An organisational learning view. Innovation, 22 (4), 422–446. 10.1080/14479338.2020.1713001.
• Yanson R, Johnson RD. An empirical examination of e-learning design: The role of trainee socialization and complexity in short term training. Computers & Education. 2016; 101 :43–54. doi: 10.1016/j.compedu.2016.05.010. [ CrossRef ] [ Google Scholar ]
• Yoo SJ, Huang WD. Can e-learning system enhance learning culture in the workplace? A comparison among companies in South Korea. British Journal of Educational Technology. 2016; 47 (4):575–591. doi: 10.1111/bjet.12240. [ CrossRef ] [ Google Scholar ]
• Zhang X, Jiang S, Ordóñez de Pablos P, Lytras MD, Sun Y. How virtual reality affects perceived learning effectiveness: A task–technology fit perspective. Behaviour & Information Technology. 2017; 36 (5):548–556. doi: 10.1080/0144929X.2016.1268647. [ CrossRef ] [ Google Scholar ]
• Zhang X, Meng Y, de Pablos PO, Sun Y. Learning analytics in collaborative learning supported by Slack: From the perspective of engagement. Computers in Human Behavior. 2019; 92 :625–633. doi: 10.1016/j.chb.2017.08.012. [ CrossRef ] [ Google Scholar ]

Welcome to the MIT CISR website!

This site uses cookies. Review our Privacy Statement.

Making Talent a Competitive Advantage—Session Presentation

Why can't i view this research.

This presentation from MIT CISR's 2024 European International Executive Forum event is available only to logged-in members and attendees of the event.

See the list of MIT CISR members

Related publications.

Working Paper: Case Study

Developing a digital-first workforce: ai-driven skills enablement at johnson & johnson.

About the Researcher

Nick van der Meulen, Research Scientist, MIT Center for Information Systems Research (CISR)

Mit center for information systems research (cisr).

Founded in 1974 and grounded in MIT's tradition of combining academic knowledge and practical purpose, MIT CISR helps executives meet the challenge of leading increasingly digital and data-driven organizations. We work directly with digital leaders, executives, and boards to develop our insights. Our consortium forms a global community that comprises more than seventy-five organizations.

MIT CISR Associate Members

MIT CISR wishes to thank all of our associate members for their support and contributions.

MIT CISR's Mission Expand

MIT CISR helps executives meet the challenge of leading increasingly digital and data-driven organizations. We provide insights on how organizations effectively realize value from approaches such as digital business transformation, data monetization, business ecosystems, and the digital workplace. Founded in 1974 and grounded in MIT’s tradition of combining academic knowledge and practical purpose, we work directly with digital leaders, executives, and boards to develop our insights. Our consortium forms a global community that comprises more than seventy-five organizations.

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

• View all journals
• My Account Login
• Explore content
• About the journal
• Publish with us
• Sign up for alerts
• Open access
• Published: 26 March 2024

Predicting and improving complex beer flavor through machine learning

• Michiel Schreurs   ORCID: orcid.org/0000-0002-9449-5619 1 , 2 , 3   na1 ,
• Supinya Piampongsant 1 , 2 , 3   na1 ,
• Miguel Roncoroni   ORCID: orcid.org/0000-0001-7461-1427 1 , 2 , 3   na1 ,
• Lloyd Cool   ORCID: orcid.org/0000-0001-9936-3124 1 , 2 , 3 , 4 ,
• Beatriz Herrera-Malaver   ORCID: orcid.org/0000-0002-5096-9974 1 , 2 , 3 ,
• Christophe Vanderaa   ORCID: orcid.org/0000-0001-7443-5427 4 ,
• Florian A. Theßeling 1 , 2 , 3 ,
• Łukasz Kreft   ORCID: orcid.org/0000-0001-7620-4657 5 ,
• Alexander Botzki   ORCID: orcid.org/0000-0001-6691-4233 5 ,
• Philippe Malcorps 6 ,
• Luk Daenen 6 ,
• Tom Wenseleers   ORCID: orcid.org/0000-0002-1434-861X 4 &
• Kevin J. Verstrepen   ORCID: orcid.org/0000-0002-3077-6219 1 , 2 , 3  

Nature Communications volume  15 , Article number:  2368 ( 2024 ) Cite this article

41k Accesses

778 Altmetric

Metrics details

• Chemical engineering
• Gas chromatography
• Machine learning
• Metabolomics
• Taste receptors

The perception and appreciation of food flavor depends on many interacting chemical compounds and external factors, and therefore proves challenging to understand and predict. Here, we combine extensive chemical and sensory analyses of 250 different beers to train machine learning models that allow predicting flavor and consumer appreciation. For each beer, we measure over 200 chemical properties, perform quantitative descriptive sensory analysis with a trained tasting panel and map data from over 180,000 consumer reviews to train 10 different machine learning models. The best-performing algorithm, Gradient Boosting, yields models that significantly outperform predictions based on conventional statistics and accurately predict complex food features and consumer appreciation from chemical profiles. Model dissection allows identifying specific and unexpected compounds as drivers of beer flavor and appreciation. Adding these compounds results in variants of commercial alcoholic and non-alcoholic beers with improved consumer appreciation. Together, our study reveals how big data and machine learning uncover complex links between food chemistry, flavor and consumer perception, and lays the foundation to develop novel, tailored foods with superior flavors.

Similar content being viewed by others

BitterSweet: Building machine learning models for predicting the bitter and sweet taste of small molecules

Rudraksh Tuwani, Somin Wadhwa & Ganesh Bagler

Sensory lexicon and aroma volatiles analysis of brewing malt

Xiaoxia Su, Miao Yu, … Tianyi Du

Predicting odor from molecular structure: a multi-label classification approach

Kushagra Saini & Venkatnarayan Ramanathan

Introduction

Predicting and understanding food perception and appreciation is one of the major challenges in food science. Accurate modeling of food flavor and appreciation could yield important opportunities for both producers and consumers, including quality control, product fingerprinting, counterfeit detection, spoilage detection, and the development of new products and product combinations (food pairing) 1 , 2 , 3 , 4 , 5 , 6 . Accurate models for flavor and consumer appreciation would contribute greatly to our scientific understanding of how humans perceive and appreciate flavor. Moreover, accurate predictive models would also facilitate and standardize existing food assessment methods and could supplement or replace assessments by trained and consumer tasting panels, which are variable, expensive and time-consuming 7 , 8 , 9 . Lastly, apart from providing objective, quantitative, accurate and contextual information that can help producers, models can also guide consumers in understanding their personal preferences 10 .

Despite the myriad of applications, predicting food flavor and appreciation from its chemical properties remains a largely elusive goal in sensory science, especially for complex food and beverages 11 , 12 . A key obstacle is the immense number of flavor-active chemicals underlying food flavor. Flavor compounds can vary widely in chemical structure and concentration, making them technically challenging and labor-intensive to quantify, even in the face of innovations in metabolomics, such as non-targeted metabolic fingerprinting 13 , 14 . Moreover, sensory analysis is perhaps even more complicated. Flavor perception is highly complex, resulting from hundreds of different molecules interacting at the physiochemical and sensorial level. Sensory perception is often non-linear, characterized by complex and concentration-dependent synergistic and antagonistic effects 15 , 16 , 17 , 18 , 19 , 20 , 21 that are further convoluted by the genetics, environment, culture and psychology of consumers 22 , 23 , 24 . Perceived flavor is therefore difficult to measure, with problems of sensitivity, accuracy, and reproducibility that can only be resolved by gathering sufficiently large datasets 25 . Trained tasting panels are considered the prime source of quality sensory data, but require meticulous training, are low throughput and high cost. Public databases containing consumer reviews of food products could provide a valuable alternative, especially for studying appreciation scores, which do not require formal training 25 . Public databases offer the advantage of amassing large amounts of data, increasing the statistical power to identify potential drivers of appreciation. However, public datasets suffer from biases, including a bias in the volunteers that contribute to the database, as well as confounding factors such as price, cult status and psychological conformity towards previous ratings of the product.

Classical multivariate statistics and machine learning methods have been used to predict flavor of specific compounds by, for example, linking structural properties of a compound to its potential biological activities or linking concentrations of specific compounds to sensory profiles 1 , 26 . Importantly, most previous studies focused on predicting organoleptic properties of single compounds (often based on their chemical structure) 27 , 28 , 29 , 30 , 31 , 32 , 33 , thus ignoring the fact that these compounds are present in a complex matrix in food or beverages and excluding complex interactions between compounds. Moreover, the classical statistics commonly used in sensory science 34 , 35 , 36 , 37 , 38 , 39 require a large sample size and sufficient variance amongst predictors to create accurate models. They are not fit for studying an extensive set of hundreds of interacting flavor compounds, since they are sensitive to outliers, have a high tendency to overfit and are less suited for non-linear and discontinuous relationships 40 .

In this study, we combine extensive chemical analyses and sensory data of a set of different commercial beers with machine learning approaches to develop models that predict taste, smell, mouthfeel and appreciation from compound concentrations. Beer is particularly suited to model the relationship between chemistry, flavor and appreciation. First, beer is a complex product, consisting of thousands of flavor compounds that partake in complex sensory interactions 41 , 42 , 43 . This chemical diversity arises from the raw materials (malt, yeast, hops, water and spices) and biochemical conversions during the brewing process (kilning, mashing, boiling, fermentation, maturation and aging) 44 , 45 . Second, the advent of the internet saw beer consumers embrace online review platforms, such as RateBeer (ZX Ventures, Anheuser-Busch InBev SA/NV) and BeerAdvocate (Next Glass, inc.). In this way, the beer community provides massive data sets of beer flavor and appreciation scores, creating extraordinarily large sensory databases to complement the analyses of our professional sensory panel. Specifically, we characterize over 200 chemical properties of 250 commercial beers, spread across 22 beer styles, and link these to the descriptive sensory profiling data of a 16-person in-house trained tasting panel and data acquired from over 180,000 public consumer reviews. These unique and extensive datasets enable us to train a suite of machine learning models to predict flavor and appreciation from a beer’s chemical profile. Dissection of the best-performing models allows us to pinpoint specific compounds as potential drivers of beer flavor and appreciation. Follow-up experiments confirm the importance of these compounds and ultimately allow us to significantly improve the flavor and appreciation of selected commercial beers. Together, our study represents a significant step towards understanding complex flavors and reinforces the value of machine learning to develop and refine complex foods. In this way, it represents a stepping stone for further computer-aided food engineering applications 46 .

To generate a comprehensive dataset on beer flavor, we selected 250 commercial Belgian beers across 22 different beer styles (Supplementary Fig.  S1 ). Beers with ≤ 4.2% alcohol by volume (ABV) were classified as non-alcoholic and low-alcoholic. Blonds and Tripels constitute a significant portion of the dataset (12.4% and 11.2%, respectively) reflecting their presence on the Belgian beer market and the heterogeneity of beers within these styles. By contrast, lager beers are less diverse and dominated by a handful of brands. Rare styles such as Brut or Faro make up only a small fraction of the dataset (2% and 1%, respectively) because fewer of these beers are produced and because they are dominated by distinct characteristics in terms of flavor and chemical composition.

Extensive analysis identifies relationships between chemical compounds in beer

For each beer, we measured 226 different chemical properties, including common brewing parameters such as alcohol content, iso-alpha acids, pH, sugar concentration 47 , and over 200 flavor compounds (Methods, Supplementary Table  S1 ). A large portion (37.2%) are terpenoids arising from hopping, responsible for herbal and fruity flavors 16 , 48 . A second major category are yeast metabolites, such as esters and alcohols, that result in fruity and solvent notes 48 , 49 , 50 . Other measured compounds are primarily derived from malt, or other microbes such as non- Saccharomyces yeasts and bacteria (‘wild flora’). Compounds that arise from spices or staling are labeled under ‘Others’. Five attributes (caloric value, total acids and total ester, hop aroma and sulfur compounds) are calculated from multiple individually measured compounds.

As a first step in identifying relationships between chemical properties, we determined correlations between the concentrations of the compounds (Fig.  1 , upper panel, Supplementary Data  1 and 2 , and Supplementary Fig.  S2 . For the sake of clarity, only a subset of the measured compounds is shown in Fig.  1 ). Compounds of the same origin typically show a positive correlation, while absence of correlation hints at parameters varying independently. For example, the hop aroma compounds citronellol, and alpha-terpineol show moderate correlations with each other (Spearman’s rho=0.39 and 0.57), but not with the bittering hop component iso-alpha acids (Spearman’s rho=0.16 and −0.07). This illustrates how brewers can independently modify hop aroma and bitterness by selecting hop varieties and dosage time. If hops are added early in the boiling phase, chemical conversions increase bitterness while aromas evaporate, conversely, late addition of hops preserves aroma but limits bitterness 51 . Similarly, hop-derived iso-alpha acids show a strong anti-correlation with lactic acid and acetic acid, likely reflecting growth inhibition of lactic acid and acetic acid bacteria, or the consequent use of fewer hops in sour beer styles, such as West Flanders ales and Fruit beers, that rely on these bacteria for their distinct flavors 52 . Finally, yeast-derived esters (ethyl acetate, ethyl decanoate, ethyl hexanoate, ethyl octanoate) and alcohols (ethanol, isoamyl alcohol, isobutanol, and glycerol), correlate with Spearman coefficients above 0.5, suggesting that these secondary metabolites are correlated with the yeast genetic background and/or fermentation parameters and may be difficult to influence individually, although the choice of yeast strain may offer some control 53 .

Spearman rank correlations are shown. Descriptors are grouped according to their origin (malt (blue), hops (green), yeast (red), wild flora (yellow), Others (black)), and sensory aspect (aroma, taste, palate, and overall appreciation). Please note that for the chemical compounds, for the sake of clarity, only a subset of the total number of measured compounds is shown, with an emphasis on the key compounds for each source. For more details, see the main text and Methods section. Chemical data can be found in Supplementary Data  1 , correlations between all chemical compounds are depicted in Supplementary Fig.  S2 and correlation values can be found in Supplementary Data  2 . See Supplementary Data  4 for sensory panel assessments and Supplementary Data  5 for correlation values between all sensory descriptors.

Interestingly, different beer styles show distinct patterns for some flavor compounds (Supplementary Fig.  S3 ). These observations agree with expectations for key beer styles, and serve as a control for our measurements. For instance, Stouts generally show high values for color (darker), while hoppy beers contain elevated levels of iso-alpha acids, compounds associated with bitter hop taste. Acetic and lactic acid are not prevalent in most beers, with notable exceptions such as Kriek, Lambic, Faro, West Flanders ales and Flanders Old Brown, which use acid-producing bacteria ( Lactobacillus and Pediococcus ) or unconventional yeast ( Brettanomyces ) 54 , 55 . Glycerol, ethanol and esters show similar distributions across all beer styles, reflecting their common origin as products of yeast metabolism during fermentation 45 , 53 . Finally, low/no-alcohol beers contain low concentrations of glycerol and esters. This is in line with the production process for most of the low/no-alcohol beers in our dataset, which are produced through limiting fermentation or by stripping away alcohol via evaporation or dialysis, with both methods having the unintended side-effect of reducing the amount of flavor compounds in the final beer 56 , 57 .

Besides expected associations, our data also reveals less trivial associations between beer styles and specific parameters. For example, geraniol and citronellol, two monoterpenoids responsible for citrus, floral and rose flavors and characteristic of Citra hops, are found in relatively high amounts in Christmas, Saison, and Brett/co-fermented beers, where they may originate from terpenoid-rich spices such as coriander seeds instead of hops 58 .

Tasting panel assessments reveal sensorial relationships in beer

To assess the sensory profile of each beer, a trained tasting panel evaluated each of the 250 beers for 50 sensory attributes, including different hop, malt and yeast flavors, off-flavors and spices. Panelists used a tasting sheet (Supplementary Data  3 ) to score the different attributes. Panel consistency was evaluated by repeating 12 samples across different sessions and performing ANOVA. In 95% of cases no significant difference was found across sessions ( p  > 0.05), indicating good panel consistency (Supplementary Table  S2 ).

Aroma and taste perception reported by the trained panel are often linked (Fig.  1 , bottom left panel and Supplementary Data  4 and 5 ), with high correlations between hops aroma and taste (Spearman’s rho=0.83). Bitter taste was found to correlate with hop aroma and taste in general (Spearman’s rho=0.80 and 0.69), and particularly with “grassy” noble hops (Spearman’s rho=0.75). Barnyard flavor, most often associated with sour beers, is identified together with stale hops (Spearman’s rho=0.97) that are used in these beers. Lactic and acetic acid, which often co-occur, are correlated (Spearman’s rho=0.66). Interestingly, sweetness and bitterness are anti-correlated (Spearman’s rho = −0.48), confirming the hypothesis that they mask each other 59 , 60 . Beer body is highly correlated with alcohol (Spearman’s rho = 0.79), and overall appreciation is found to correlate with multiple aspects that describe beer mouthfeel (alcohol, carbonation; Spearman’s rho= 0.32, 0.39), as well as with hop and ester aroma intensity (Spearman’s rho=0.39 and 0.35).

Similar to the chemical analyses, sensorial analyses confirmed typical features of specific beer styles (Supplementary Fig.  S4 ). For example, sour beers (Faro, Flanders Old Brown, Fruit beer, Kriek, Lambic, West Flanders ale) were rated acidic, with flavors of both acetic and lactic acid. Hoppy beers were found to be bitter and showed hop-associated aromas like citrus and tropical fruit. Malt taste is most detected among scotch, stout/porters, and strong ales, while low/no-alcohol beers, which often have a reputation for being ‘worty’ (reminiscent of unfermented, sweet malt extract) appear in the middle. Unsurprisingly, hop aromas are most strongly detected among hoppy beers. Like its chemical counterpart (Supplementary Fig.  S3 ), acidity shows a right-skewed distribution, with the most acidic beers being Krieks, Lambics, and West Flanders ales.

Tasting panel assessments of specific flavors correlate with chemical composition

We find that the concentrations of several chemical compounds strongly correlate with specific aroma or taste, as evaluated by the tasting panel (Fig.  2 , Supplementary Fig.  S5 , Supplementary Data  6 ). In some cases, these correlations confirm expectations and serve as a useful control for data quality. For example, iso-alpha acids, the bittering compounds in hops, strongly correlate with bitterness (Spearman’s rho=0.68), while ethanol and glycerol correlate with tasters’ perceptions of alcohol and body, the mouthfeel sensation of fullness (Spearman’s rho=0.82/0.62 and 0.72/0.57 respectively) and darker color from roasted malts is a good indication of malt perception (Spearman’s rho=0.54).

Heatmap colors indicate Spearman’s Rho. Axes are organized according to sensory categories (aroma, taste, mouthfeel, overall), chemical categories and chemical sources in beer (malt (blue), hops (green), yeast (red), wild flora (yellow), Others (black)). See Supplementary Data  6 for all correlation values.

Interestingly, for some relationships between chemical compounds and perceived flavor, correlations are weaker than expected. For example, the rose-smelling phenethyl acetate only weakly correlates with floral aroma. This hints at more complex relationships and interactions between compounds and suggests a need for a more complex model than simple correlations. Lastly, we uncovered unexpected correlations. For instance, the esters ethyl decanoate and ethyl octanoate appear to correlate slightly with hop perception and bitterness, possibly due to their fruity flavor. Iron is anti-correlated with hop aromas and bitterness, most likely because it is also anti-correlated with iso-alpha acids. This could be a sign of metal chelation of hop acids 61 , given that our analyses measure unbound hop acids and total iron content, or could result from the higher iron content in dark and Fruit beers, which typically have less hoppy and bitter flavors 62 .

Public consumer reviews complement expert panel data

To complement and expand the sensory data of our trained tasting panel, we collected 180,000 reviews of our 250 beers from the online consumer review platform RateBeer. This provided numerical scores for beer appearance, aroma, taste, palate, overall quality as well as the average overall score.

Public datasets are known to suffer from biases, such as price, cult status and psychological conformity towards previous ratings of a product. For example, prices correlate with appreciation scores for these online consumer reviews (rho=0.49, Supplementary Fig.  S6 ), but not for our trained tasting panel (rho=0.19). This suggests that prices affect consumer appreciation, which has been reported in wine 63 , while blind tastings are unaffected. Moreover, we observe that some beer styles, like lagers and non-alcoholic beers, generally receive lower scores, reflecting that online reviewers are mostly beer aficionados with a preference for specialty beers over lager beers. In general, we find a modest correlation between our trained panel’s overall appreciation score and the online consumer appreciation scores (Fig.  3 , rho=0.29). Apart from the aforementioned biases in the online datasets, serving temperature, sample freshness and surroundings, which are all tightly controlled during the tasting panel sessions, can vary tremendously across online consumers and can further contribute to (among others, appreciation) differences between the two categories of tasters. Importantly, in contrast to the overall appreciation scores, for many sensory aspects the results from the professional panel correlated well with results obtained from RateBeer reviews. Correlations were highest for features that are relatively easy to recognize even for untrained tasters, like bitterness, sweetness, alcohol and malt aroma (Fig.  3 and below).

RateBeer text mining results can be found in Supplementary Data  7 . Rho values shown are Spearman correlation values, with asterisks indicating significant correlations ( p  < 0.05, two-sided). All p values were smaller than 0.001, except for Esters aroma (0.0553), Esters taste (0.3275), Esters aroma—banana (0.0019), Coriander (0.0508) and Diacetyl (0.0134).

Besides collecting consumer appreciation from these online reviews, we developed automated text analysis tools to gather additional data from review texts (Supplementary Data  7 ). Processing review texts on the RateBeer database yielded comparable results to the scores given by the trained panel for many common sensory aspects, including acidity, bitterness, sweetness, alcohol, malt, and hop tastes (Fig.  3 ). This is in line with what would be expected, since these attributes require less training for accurate assessment and are less influenced by environmental factors such as temperature, serving glass and odors in the environment. Consumer reviews also correlate well with our trained panel for 4-vinyl guaiacol, a compound associated with a very characteristic aroma. By contrast, correlations for more specific aromas like ester, coriander or diacetyl are underrepresented in the online reviews, underscoring the importance of using a trained tasting panel and standardized tasting sheets with explicit factors to be scored for evaluating specific aspects of a beer. Taken together, our results suggest that public reviews are trustworthy for some, but not all, flavor features and can complement or substitute taste panel data for these sensory aspects.

Models can predict beer sensory profiles from chemical data

The rich datasets of chemical analyses, tasting panel assessments and public reviews gathered in the first part of this study provided us with a unique opportunity to develop predictive models that link chemical data to sensorial features. Given the complexity of beer flavor, basic statistical tools such as correlations or linear regression may not always be the most suitable for making accurate predictions. Instead, we applied different machine learning models that can model both simple linear and complex interactive relationships. Specifically, we constructed a set of regression models to predict (a) trained panel scores for beer flavor and quality and (b) public reviews’ appreciation scores from beer chemical profiles. We trained and tested 10 different models (Methods), 3 linear regression-based models (simple linear regression with first-order interactions (LR), lasso regression with first-order interactions (Lasso), partial least squares regressor (PLSR)), 5 decision tree models (AdaBoost regressor (ABR), extra trees (ET), gradient boosting regressor (GBR), random forest (RF) and XGBoost regressor (XGBR)), 1 support vector regression (SVR), and 1 artificial neural network (ANN) model.

To compare the performance of our machine learning models, the dataset was randomly split into a training and test set, stratified by beer style. After a model was trained on data in the training set, its performance was evaluated on its ability to predict the test dataset obtained from multi-output models (based on the coefficient of determination, see Methods). Additionally, individual-attribute models were ranked per descriptor and the average rank was calculated, as proposed by Korneva et al. 64 . Importantly, both ways of evaluating the models’ performance agreed in general. Performance of the different models varied (Table  1 ). It should be noted that all models perform better at predicting RateBeer results than results from our trained tasting panel. One reason could be that sensory data is inherently variable, and this variability is averaged out with the large number of public reviews from RateBeer. Additionally, all tree-based models perform better at predicting taste than aroma. Linear models (LR) performed particularly poorly, with negative R 2 values, due to severe overfitting (training set R 2  = 1). Overfitting is a common issue in linear models with many parameters and limited samples, especially with interaction terms further amplifying the number of parameters. L1 regularization (Lasso) successfully overcomes this overfitting, out-competing multiple tree-based models on the RateBeer dataset. Similarly, the dimensionality reduction of PLSR avoids overfitting and improves performance, to some extent. Still, tree-based models (ABR, ET, GBR, RF and XGBR) show the best performance, out-competing the linear models (LR, Lasso, PLSR) commonly used in sensory science 65 .

GBR models showed the best overall performance in predicting sensory responses from chemical information, with R 2 values up to 0.75 depending on the predicted sensory feature (Supplementary Table  S4 ). The GBR models predict consumer appreciation (RateBeer) better than our trained panel’s appreciation (R 2 value of 0.67 compared to R 2 value of 0.09) (Supplementary Table  S3 and Supplementary Table  S4 ). ANN models showed intermediate performance, likely because neural networks typically perform best with larger datasets 66 . The SVR shows intermediate performance, mostly due to the weak predictions of specific attributes that lower the overall performance (Supplementary Table  S4 ).

Model dissection identifies specific, unexpected compounds as drivers of consumer appreciation

Next, we leveraged our models to infer important contributors to sensory perception and consumer appreciation. Consumer preference is a crucial sensory aspects, because a product that shows low consumer appreciation scores often does not succeed commercially 25 . Additionally, the requirement for a large number of representative evaluators makes consumer trials one of the more costly and time-consuming aspects of product development. Hence, a model for predicting chemical drivers of overall appreciation would be a welcome addition to the available toolbox for food development and optimization.

Since GBR models on our RateBeer dataset showed the best overall performance, we focused on these models. Specifically, we used two approaches to identify important contributors. First, rankings of the most important predictors for each sensorial trait in the GBR models were obtained based on impurity-based feature importance (mean decrease in impurity). High-ranked parameters were hypothesized to be either the true causal chemical properties underlying the trait, to correlate with the actual causal properties, or to take part in sensory interactions affecting the trait 67 (Fig.  4A ). In a second approach, we used SHAP 68 to determine which parameters contributed most to the model for making predictions of consumer appreciation (Fig.  4B ). SHAP calculates parameter contributions to model predictions on a per-sample basis, which can be aggregated into an importance score.

A The impurity-based feature importance (mean deviance in impurity, MDI) calculated from the Gradient Boosting Regression (GBR) model predicting RateBeer appreciation scores. The top 15 highest ranked chemical properties are shown. B SHAP summary plot for the top 15 parameters contributing to our GBR model. Each point on the graph represents a sample from our dataset. The color represents the concentration of that parameter, with bluer colors representing low values and redder colors representing higher values. Greater absolute values on the horizontal axis indicate a higher impact of the parameter on the prediction of the model. C Spearman correlations between the 15 most important chemical properties and consumer overall appreciation. Numbers indicate the Spearman Rho correlation coefficient, and the rank of this correlation compared to all other correlations. The top 15 important compounds were determined using SHAP (panel B).

Both approaches identified ethyl acetate as the most predictive parameter for beer appreciation (Fig.  4 ). Ethyl acetate is the most abundant ester in beer with a typical ‘fruity’, ‘solvent’ and ‘alcoholic’ flavor, but is often considered less important than other esters like isoamyl acetate. The second most important parameter identified by SHAP is ethanol, the most abundant beer compound after water. Apart from directly contributing to beer flavor and mouthfeel, ethanol drastically influences the physical properties of beer, dictating how easily volatile compounds escape the beer matrix to contribute to beer aroma 69 . Importantly, it should also be noted that the importance of ethanol for appreciation is likely inflated by the very low appreciation scores of non-alcoholic beers (Supplementary Fig.  S4 ). Despite not often being considered a driver of beer appreciation, protein level also ranks highly in both approaches, possibly due to its effect on mouthfeel and body 70 . Lactic acid, which contributes to the tart taste of sour beers, is the fourth most important parameter identified by SHAP, possibly due to the generally high appreciation of sour beers in our dataset.

Interestingly, some of the most important predictive parameters for our model are not well-established as beer flavors or are even commonly regarded as being negative for beer quality. For example, our models identify methanethiol and ethyl phenyl acetate, an ester commonly linked to beer staling 71 , as a key factor contributing to beer appreciation. Although there is no doubt that high concentrations of these compounds are considered unpleasant, the positive effects of modest concentrations are not yet known 72 , 73 .

To compare our approach to conventional statistics, we evaluated how well the 15 most important SHAP-derived parameters correlate with consumer appreciation (Fig.  4C ). Interestingly, only 6 of the properties derived by SHAP rank amongst the top 15 most correlated parameters. For some chemical compounds, the correlations are so low that they would have likely been considered unimportant. For example, lactic acid, the fourth most important parameter, shows a bimodal distribution for appreciation, with sour beers forming a separate cluster, that is missed entirely by the Spearman correlation. Additionally, the correlation plots reveal outliers, emphasizing the need for robust analysis tools. Together, this highlights the need for alternative models, like the Gradient Boosting model, that better grasp the complexity of (beer) flavor.

Finally, to observe the relationships between these chemical properties and their predicted targets, partial dependence plots were constructed for the six most important predictors of consumer appreciation 74 , 75 , 76 (Supplementary Fig.  S7 ). One-way partial dependence plots show how a change in concentration affects the predicted appreciation. These plots reveal an important limitation of our models: appreciation predictions remain constant at ever-increasing concentrations. This implies that once a threshold concentration is reached, further increasing the concentration does not affect appreciation. This is false, as it is well-documented that certain compounds become unpleasant at high concentrations, including ethyl acetate (‘nail polish’) 77 and methanethiol (‘sulfury’ and ‘rotten cabbage’) 78 . The inability of our models to grasp that flavor compounds have optimal levels, above which they become negative, is a consequence of working with commercial beer brands where (off-)flavors are rarely too high to negatively impact the product. The two-way partial dependence plots show how changing the concentration of two compounds influences predicted appreciation, visualizing their interactions (Supplementary Fig.  S7 ). In our case, the top 5 parameters are dominated by additive or synergistic interactions, with high concentrations for both compounds resulting in the highest predicted appreciation.

To assess the robustness of our best-performing models and model predictions, we performed 100 iterations of the GBR, RF and ET models. In general, all iterations of the models yielded similar performance (Supplementary Fig.  S8 ). Moreover, the main predictors (including the top predictors ethanol and ethyl acetate) remained virtually the same, especially for GBR and RF. For the iterations of the ET model, we did observe more variation in the top predictors, which is likely a consequence of the model’s inherent random architecture in combination with co-correlations between certain predictors. However, even in this case, several of the top predictors (ethanol and ethyl acetate) remain unchanged, although their rank in importance changes (Supplementary Fig.  S8 ).

Next, we investigated if a combination of RateBeer and trained panel data into one consolidated dataset would lead to stronger models, under the hypothesis that such a model would suffer less from bias in the datasets. A GBR model was trained to predict appreciation on the combined dataset. This model underperformed compared to the RateBeer model, both in the native case and when including a dataset identifier (R 2  = 0.67, 0.26 and 0.42 respectively). For the latter, the dataset identifier is the most important feature (Supplementary Fig.  S9 ), while most of the feature importance remains unchanged, with ethyl acetate and ethanol ranking highest, like in the original model trained only on RateBeer data. It seems that the large variation in the panel dataset introduces noise, weakening the models’ performances and reliability. In addition, it seems reasonable to assume that both datasets are fundamentally different, with the panel dataset obtained by blind tastings by a trained professional panel.

Lastly, we evaluated whether beer style identifiers would further enhance the model’s performance. A GBR model was trained with parameters that explicitly encoded the styles of the samples. This did not improve model performance (R2 = 0.66 with style information vs R2 = 0.67). The most important chemical features are consistent with the model trained without style information (eg. ethanol and ethyl acetate), and with the exception of the most preferred (strong ale) and least preferred (low/no-alcohol) styles, none of the styles were among the most important features (Supplementary Fig.  S9 , Supplementary Table  S5 and S6 ). This is likely due to a combination of style-specific chemical signatures, such as iso-alpha acids and lactic acid, that implicitly convey style information to the original models, as well as the low number of samples belonging to some styles, making it difficult for the model to learn style-specific patterns. Moreover, beer styles are not rigorously defined, with some styles overlapping in features and some beers being misattributed to a specific style, all of which leads to more noise in models that use style parameters.

Model validation

To test if our predictive models give insight into beer appreciation, we set up experiments aimed at improving existing commercial beers. We specifically selected overall appreciation as the trait to be examined because of its complexity and commercial relevance. Beer flavor comprises a complex bouquet rather than single aromas and tastes 53 . Hence, adding a single compound to the extent that a difference is noticeable may lead to an unbalanced, artificial flavor. Therefore, we evaluated the effect of combinations of compounds. Because Blond beers represent the most extensive style in our dataset, we selected a beer from this style as the starting material for these experiments (Beer 64 in Supplementary Data  1 ).

In the first set of experiments, we adjusted the concentrations of compounds that made up the most important predictors of overall appreciation (ethyl acetate, ethanol, lactic acid, ethyl phenyl acetate) together with correlated compounds (ethyl hexanoate, isoamyl acetate, glycerol), bringing them up to 95 th percentile ethanol-normalized concentrations (Methods) within the Blond group (‘Spiked’ concentration in Fig.  5A ). Compared to controls, the spiked beers were found to have significantly improved overall appreciation among trained panelists, with panelist noting increased intensity of ester flavors, sweetness, alcohol, and body fullness (Fig.  5B ). To disentangle the contribution of ethanol to these results, a second experiment was performed without the addition of ethanol. This resulted in a similar outcome, including increased perception of alcohol and overall appreciation.

Adding the top chemical compounds, identified as best predictors of appreciation by our model, into poorly appreciated beers results in increased appreciation from our trained panel. Results of sensory tests between base beers and those spiked with compounds identified as the best predictors by the model. A Blond and Non/Low-alcohol (0.0% ABV) base beers were brought up to 95th-percentile ethanol-normalized concentrations within each style. B For each sensory attribute, tasters indicated the more intense sample and selected the sample they preferred. The numbers above the bars correspond to the p values that indicate significant changes in perceived flavor (two-sided binomial test: alpha 0.05, n  = 20 or 13).

In a last experiment, we tested whether using the model’s predictions can boost the appreciation of a non-alcoholic beer (beer 223 in Supplementary Data  1 ). Again, the addition of a mixture of predicted compounds (omitting ethanol, in this case) resulted in a significant increase in appreciation, body, ester flavor and sweetness.

Predicting flavor and consumer appreciation from chemical composition is one of the ultimate goals of sensory science. A reliable, systematic and unbiased way to link chemical profiles to flavor and food appreciation would be a significant asset to the food and beverage industry. Such tools would substantially aid in quality control and recipe development, offer an efficient and cost-effective alternative to pilot studies and consumer trials and would ultimately allow food manufacturers to produce superior, tailor-made products that better meet the demands of specific consumer groups more efficiently.

A limited set of studies have previously tried, to varying degrees of success, to predict beer flavor and beer popularity based on (a limited set of) chemical compounds and flavors 79 , 80 . Current sensitive, high-throughput technologies allow measuring an unprecedented number of chemical compounds and properties in a large set of samples, yielding a dataset that can train models that help close the gaps between chemistry and flavor, even for a complex natural product like beer. To our knowledge, no previous research gathered data at this scale (250 samples, 226 chemical parameters, 50 sensory attributes and 5 consumer scores) to disentangle and validate the chemical aspects driving beer preference using various machine-learning techniques. We find that modern machine learning models outperform conventional statistical tools, such as correlations and linear models, and can successfully predict flavor appreciation from chemical composition. This could be attributed to the natural incorporation of interactions and non-linear or discontinuous effects in machine learning models, which are not easily grasped by the linear model architecture. While linear models and partial least squares regression represent the most widespread statistical approaches in sensory science, in part because they allow interpretation 65 , 81 , 82 , modern machine learning methods allow for building better predictive models while preserving the possibility to dissect and exploit the underlying patterns. Of the 10 different models we trained, tree-based models, such as our best performing GBR, showed the best overall performance in predicting sensory responses from chemical information, outcompeting artificial neural networks. This agrees with previous reports for models trained on tabular data 83 . Our results are in line with the findings of Colantonio et al. who also identified the gradient boosting architecture as performing best at predicting appreciation and flavor (of tomatoes and blueberries, in their specific study) 26 . Importantly, besides our larger experimental scale, we were able to directly confirm our models’ predictions in vivo.

Our study confirms that flavor compound concentration does not always correlate with perception, suggesting complex interactions that are often missed by more conventional statistics and simple models. Specifically, we find that tree-based algorithms may perform best in developing models that link complex food chemistry with aroma. Furthermore, we show that massive datasets of untrained consumer reviews provide a valuable source of data, that can complement or even replace trained tasting panels, especially for appreciation and basic flavors, such as sweetness and bitterness. This holds despite biases that are known to occur in such datasets, such as price or conformity bias. Moreover, GBR models predict taste better than aroma. This is likely because taste (e.g. bitterness) often directly relates to the corresponding chemical measurements (e.g., iso-alpha acids), whereas such a link is less clear for aromas, which often result from the interplay between multiple volatile compounds. We also find that our models are best at predicting acidity and alcohol, likely because there is a direct relation between the measured chemical compounds (acids and ethanol) and the corresponding perceived sensorial attribute (acidity and alcohol), and because even untrained consumers are generally able to recognize these flavors and aromas.

The predictions of our final models, trained on review data, hold even for blind tastings with small groups of trained tasters, as demonstrated by our ability to validate specific compounds as drivers of beer flavor and appreciation. Since adding a single compound to the extent of a noticeable difference may result in an unbalanced flavor profile, we specifically tested our identified key drivers as a combination of compounds. While this approach does not allow us to validate if a particular single compound would affect flavor and/or appreciation, our experiments do show that this combination of compounds increases consumer appreciation.

It is important to stress that, while it represents an important step forward, our approach still has several major limitations. A key weakness of the GBR model architecture is that amongst co-correlating variables, the largest main effect is consistently preferred for model building. As a result, co-correlating variables often have artificially low importance scores, both for impurity and SHAP-based methods, like we observed in the comparison to the more randomized Extra Trees models. This implies that chemicals identified as key drivers of a specific sensory feature by GBR might not be the true causative compounds, but rather co-correlate with the actual causative chemical. For example, the high importance of ethyl acetate could be (partially) attributed to the total ester content, ethanol or ethyl hexanoate (rho=0.77, rho=0.72 and rho=0.68), while ethyl phenylacetate could hide the importance of prenyl isobutyrate and ethyl benzoate (rho=0.77 and rho=0.76). Expanding our GBR model to include beer style as a parameter did not yield additional power or insight. This is likely due to style-specific chemical signatures, such as iso-alpha acids and lactic acid, that implicitly convey style information to the original model, as well as the smaller sample size per style, limiting the power to uncover style-specific patterns. This can be partly attributed to the curse of dimensionality, where the high number of parameters results in the models mainly incorporating single parameter effects, rather than complex interactions such as style-dependent effects 67 . A larger number of samples may overcome some of these limitations and offer more insight into style-specific effects. On the other hand, beer style is not a rigid scientific classification, and beers within one style often differ a lot, which further complicates the analysis of style as a model factor.

Our study is limited to beers from Belgian breweries. Although these beers cover a large portion of the beer styles available globally, some beer styles and consumer patterns may be missing, while other features might be overrepresented. For example, many Belgian ales exhibit yeast-driven flavor profiles, which is reflected in the chemical drivers of appreciation discovered by this study. In future work, expanding the scope to include diverse markets and beer styles could lead to the identification of even more drivers of appreciation and better models for special niche products that were not present in our beer set.

In addition to inherent limitations of GBR models, there are also some limitations associated with studying food aroma. Even if our chemical analyses measured most of the known aroma compounds, the total number of flavor compounds in complex foods like beer is still larger than the subset we were able to measure in this study. For example, hop-derived thiols, that influence flavor at very low concentrations, are notoriously difficult to measure in a high-throughput experiment. Moreover, consumer perception remains subjective and prone to biases that are difficult to avoid. It is also important to stress that the models are still immature and that more extensive datasets will be crucial for developing more complete models in the future. Besides more samples and parameters, our dataset does not include any demographic information about the tasters. Including such data could lead to better models that grasp external factors like age and culture. Another limitation is that our set of beers consists of high-quality end-products and lacks beers that are unfit for sale, which limits the current model in accurately predicting products that are appreciated very badly. Finally, while models could be readily applied in quality control, their use in sensory science and product development is restrained by their inability to discern causal relationships. Given that the models cannot distinguish compounds that genuinely drive consumer perception from those that merely correlate, validation experiments are essential to identify true causative compounds.

Despite the inherent limitations, dissection of our models enabled us to pinpoint specific molecules as potential drivers of beer aroma and consumer appreciation, including compounds that were unexpected and would not have been identified using standard approaches. Important drivers of beer appreciation uncovered by our models include protein levels, ethyl acetate, ethyl phenyl acetate and lactic acid. Currently, many brewers already use lactic acid to acidify their brewing water and ensure optimal pH for enzymatic activity during the mashing process. Our results suggest that adding lactic acid can also improve beer appreciation, although its individual effect remains to be tested. Interestingly, ethanol appears to be unnecessary to improve beer appreciation, both for blond beer and alcohol-free beer. Given the growing consumer interest in alcohol-free beer, with a predicted annual market growth of >7% 84 , it is relevant for brewers to know what compounds can further increase consumer appreciation of these beers. Hence, our model may readily provide avenues to further improve the flavor and consumer appreciation of both alcoholic and non-alcoholic beers, which is generally considered one of the key challenges for future beer production.

Whereas we see a direct implementation of our results for the development of superior alcohol-free beverages and other food products, our study can also serve as a stepping stone for the development of novel alcohol-containing beverages. We want to echo the growing body of scientific evidence for the negative effects of alcohol consumption, both on the individual level by the mutagenic, teratogenic and carcinogenic effects of ethanol 85 , 86 , as well as the burden on society caused by alcohol abuse and addiction. We encourage the use of our results for the production of healthier, tastier products, including novel and improved beverages with lower alcohol contents. Furthermore, we strongly discourage the use of these technologies to improve the appreciation or addictive properties of harmful substances.

The present work demonstrates that despite some important remaining hurdles, combining the latest developments in chemical analyses, sensory analysis and modern machine learning methods offers exciting avenues for food chemistry and engineering. Soon, these tools may provide solutions in quality control and recipe development, as well as new approaches to sensory science and flavor research.

Beer selection

250 commercial Belgian beers were selected to cover the broad diversity of beer styles and corresponding diversity in chemical composition and aroma. See Supplementary Fig.  S1 .

Chemical dataset

Sample preparation.

Beers within their expiration date were purchased from commercial retailers. Samples were prepared in biological duplicates at room temperature, unless explicitly stated otherwise. Bottle pressure was measured with a manual pressure device (Steinfurth Mess-Systeme GmbH) and used to calculate CO 2 concentration. The beer was poured through two filter papers (Macherey-Nagel, 500713032 MN 713 ¼) to remove carbon dioxide and prevent spontaneous foaming. Samples were then prepared for measurements by targeted Headspace-Gas Chromatography-Flame Ionization Detector/Flame Photometric Detector (HS-GC-FID/FPD), Headspace-Solid Phase Microextraction-Gas Chromatography-Mass Spectrometry (HS-SPME-GC-MS), colorimetric analysis, enzymatic analysis, Near-Infrared (NIR) analysis, as described in the sections below. The mean values of biological duplicates are reported for each compound.

HS-GC-FID/FPD

HS-GC-FID/FPD (Shimadzu GC 2010 Plus) was used to measure higher alcohols, acetaldehyde, esters, 4-vinyl guaicol, and sulfur compounds. Each measurement comprised 5 ml of sample pipetted into a 20 ml glass vial containing 1.75 g NaCl (VWR, 27810.295). 100 µl of 2-heptanol (Sigma-Aldrich, H3003) (internal standard) solution in ethanol (Fisher Chemical, E/0650DF/C17) was added for a final concentration of 2.44 mg/L. Samples were flushed with nitrogen for 10 s, sealed with a silicone septum, stored at −80 °C and analyzed in batches of 20.

The GC was equipped with a DB-WAXetr column (length, 30 m; internal diameter, 0.32 mm; layer thickness, 0.50 µm; Agilent Technologies, Santa Clara, CA, USA) to the FID and an HP-5 column (length, 30 m; internal diameter, 0.25 mm; layer thickness, 0.25 µm; Agilent Technologies, Santa Clara, CA, USA) to the FPD. N 2 was used as the carrier gas. Samples were incubated for 20 min at 70 °C in the headspace autosampler (Flow rate, 35 cm/s; Injection volume, 1000 µL; Injection mode, split; Combi PAL autosampler, CTC analytics, Switzerland). The injector, FID and FPD temperatures were kept at 250 °C. The GC oven temperature was first held at 50 °C for 5 min and then allowed to rise to 80 °C at a rate of 5 °C/min, followed by a second ramp of 4 °C/min until 200 °C kept for 3 min and a final ramp of (4 °C/min) until 230 °C for 1 min. Results were analyzed with the GCSolution software version 2.4 (Shimadzu, Kyoto, Japan). The GC was calibrated with a 5% EtOH solution (VWR International) containing the volatiles under study (Supplementary Table  S7 ).

HS-SPME-GC-MS

HS-SPME-GC-MS (Shimadzu GCMS-QP-2010 Ultra) was used to measure additional volatile compounds, mainly comprising terpenoids and esters. Samples were analyzed by HS-SPME using a triphase DVB/Carboxen/PDMS 50/30 μm SPME fiber (Supelco Co., Bellefonte, PA, USA) followed by gas chromatography (Thermo Fisher Scientific Trace 1300 series, USA) coupled to a mass spectrometer (Thermo Fisher Scientific ISQ series MS) equipped with a TriPlus RSH autosampler. 5 ml of degassed beer sample was placed in 20 ml vials containing 1.75 g NaCl (VWR, 27810.295). 5 µl internal standard mix was added, containing 2-heptanol (1 g/L) (Sigma-Aldrich, H3003), 4-fluorobenzaldehyde (1 g/L) (Sigma-Aldrich, 128376), 2,3-hexanedione (1 g/L) (Sigma-Aldrich, 144169) and guaiacol (1 g/L) (Sigma-Aldrich, W253200) in ethanol (Fisher Chemical, E/0650DF/C17). Each sample was incubated at 60 °C in the autosampler oven with constant agitation. After 5 min equilibration, the SPME fiber was exposed to the sample headspace for 30 min. The compounds trapped on the fiber were thermally desorbed in the injection port of the chromatograph by heating the fiber for 15 min at 270 °C.

The GC-MS was equipped with a low polarity RXi-5Sil MS column (length, 20 m; internal diameter, 0.18 mm; layer thickness, 0.18 µm; Restek, Bellefonte, PA, USA). Injection was performed in splitless mode at 320 °C, a split flow of 9 ml/min, a purge flow of 5 ml/min and an open valve time of 3 min. To obtain a pulsed injection, a programmed gas flow was used whereby the helium gas flow was set at 2.7 mL/min for 0.1 min, followed by a decrease in flow of 20 ml/min to the normal 0.9 mL/min. The temperature was first held at 30 °C for 3 min and then allowed to rise to 80 °C at a rate of 7 °C/min, followed by a second ramp of 2 °C/min till 125 °C and a final ramp of 8 °C/min with a final temperature of 270 °C.

Mass acquisition range was 33 to 550 amu at a scan rate of 5 scans/s. Electron impact ionization energy was 70 eV. The interface and ion source were kept at 275 °C and 250 °C, respectively. A mix of linear n-alkanes (from C7 to C40, Supelco Co.) was injected into the GC-MS under identical conditions to serve as external retention index markers. Identification and quantification of the compounds were performed using an in-house developed R script as described in Goelen et al. and Reher et al. 87 , 88 (for package information, see Supplementary Table  S8 ). Briefly, chromatograms were analyzed using AMDIS (v2.71) 89 to separate overlapping peaks and obtain pure compound spectra. The NIST MS Search software (v2.0 g) in combination with the NIST2017, FFNSC3 and Adams4 libraries were used to manually identify the empirical spectra, taking into account the expected retention time. After background subtraction and correcting for retention time shifts between samples run on different days based on alkane ladders, compound elution profiles were extracted and integrated using a file with 284 target compounds of interest, which were either recovered in our identified AMDIS list of spectra or were known to occur in beer. Compound elution profiles were estimated for every peak in every chromatogram over a time-restricted window using weighted non-negative least square analysis after which peak areas were integrated 87 , 88 . Batch effect correction was performed by normalizing against the most stable internal standard compound, 4-fluorobenzaldehyde. Out of all 284 target compounds that were analyzed, 167 were visually judged to have reliable elution profiles and were used for final analysis.

Discrete photometric and enzymatic analysis

Discrete photometric and enzymatic analysis (Thermo Scientific TM Gallery TM Plus Beermaster Discrete Analyzer) was used to measure acetic acid, ammonia, beta-glucan, iso-alpha acids, color, sugars, glycerol, iron, pH, protein, and sulfite. 2 ml of sample volume was used for the analyses. Information regarding the reagents and standard solutions used for analyses and calibrations is included in Supplementary Table  S7 and Supplementary Table  S9 .

NIR analyses

NIR analysis (Anton Paar Alcolyzer Beer ME System) was used to measure ethanol. Measurements comprised 50 ml of sample, and a 10% EtOH solution was used for calibration.

Correlation calculations

Pairwise Spearman Rank correlations were calculated between all chemical properties.

Sensory dataset

Trained panel.

Our trained tasting panel consisted of volunteers who gave prior verbal informed consent. All compounds used for the validation experiment were of food-grade quality. The tasting sessions were approved by the Social and Societal Ethics Committee of the KU Leuven (G-2022-5677-R2(MAR)). All online reviewers agreed to the Terms and Conditions of the RateBeer website.

Sensory analysis was performed according to the American Society of Brewing Chemists (ASBC) Sensory Analysis Methods 90 . 30 volunteers were screened through a series of triangle tests. The sixteen most sensitive and consistent tasters were retained as taste panel members. The resulting panel was diverse in age [22–42, mean: 29], sex [56% male] and nationality [7 different countries]. The panel developed a consensus vocabulary to describe beer aroma, taste and mouthfeel. Panelists were trained to identify and score 50 different attributes, using a 7-point scale to rate attributes’ intensity. The scoring sheet is included as Supplementary Data  3 . Sensory assessments took place between 10–12 a.m. The beers were served in black-colored glasses. Per session, between 5 and 12 beers of the same style were tasted at 12 °C to 16 °C. Two reference beers were added to each set and indicated as ‘Reference 1 & 2’, allowing panel members to calibrate their ratings. Not all panelists were present at every tasting. Scores were scaled by standard deviation and mean-centered per taster. Values are represented as z-scores and clustered by Euclidean distance. Pairwise Spearman correlations were calculated between taste and aroma sensory attributes. Panel consistency was evaluated by repeating samples on different sessions and performing ANOVA to identify differences, using the ‘stats’ package (v4.2.2) in R (for package information, see Supplementary Table  S8 ).

Online reviews from a public database

The ‘scrapy’ package in Python (v3.6) (for package information, see Supplementary Table  S8 ). was used to collect 232,288 online reviews (mean=922, min=6, max=5343) from RateBeer, an online beer review database. Each review entry comprised 5 numerical scores (appearance, aroma, taste, palate and overall quality) and an optional review text. The total number of reviews per reviewer was collected separately. Numerical scores were scaled and centered per rater, and mean scores were calculated per beer.

For the review texts, the language was estimated using the packages ‘langdetect’ and ‘langid’ in Python. Reviews that were classified as English by both packages were kept. Reviewers with fewer than 100 entries overall were discarded. 181,025 reviews from >6000 reviewers from >40 countries remained. Text processing was done using the ‘nltk’ package in Python. Texts were corrected for slang and misspellings; proper nouns and rare words that are relevant to the beer context were specified and kept as-is (‘Chimay’,’Lambic’, etc.). A dictionary of semantically similar sensorial terms, for example ‘floral’ and ‘flower’, was created and collapsed together into one term. Words were stemmed and lemmatized to avoid identifying words such as ‘acid’ and ‘acidity’ as separate terms. Numbers and punctuation were removed.

Sentences from up to 50 randomly chosen reviews per beer were manually categorized according to the aspect of beer they describe (appearance, aroma, taste, palate, overall quality—not to be confused with the 5 numerical scores described above) or flagged as irrelevant if they contained no useful information. If a beer contained fewer than 50 reviews, all reviews were manually classified. This labeled data set was used to train a model that classified the rest of the sentences for all beers 91 . Sentences describing taste and aroma were extracted, and term frequency–inverse document frequency (TFIDF) was implemented to calculate enrichment scores for sensorial words per beer.

The sex of the tasting subject was not considered when building our sensory database. Instead, results from different panelists were averaged, both for our trained panel (56% male, 44% female) and the RateBeer reviews (70% male, 30% female for RateBeer as a whole).

Beer price collection and processing

Beer prices were collected from the following stores: Colruyt, Delhaize, Total Wine, BeerHawk, The Belgian Beer Shop, The Belgian Shop, and Beer of Belgium. Where applicable, prices were converted to Euros and normalized per liter. Spearman correlations were calculated between these prices and mean overall appreciation scores from RateBeer and the taste panel, respectively.

Pairwise Spearman Rank correlations were calculated between all sensory properties.

Machine learning models

Predictive modeling of sensory profiles from chemical data.

Regression models were constructed to predict (a) trained panel scores for beer flavors and quality from beer chemical profiles and (b) public reviews’ appreciation scores from beer chemical profiles. Z-scores were used to represent sensory attributes in both data sets. Chemical properties with log-normal distributions (Shapiro-Wilk test, p  <  0.05 ) were log-transformed. Missing chemical measurements (0.1% of all data) were replaced with mean values per attribute. Observations from 250 beers were randomly separated into a training set (70%, 175 beers) and a test set (30%, 75 beers), stratified per beer style. Chemical measurements (p = 231) were normalized based on the training set average and standard deviation. In total, three linear regression-based models: linear regression with first-order interaction terms (LR), lasso regression with first-order interaction terms (Lasso) and partial least squares regression (PLSR); five decision tree models, Adaboost regressor (ABR), Extra Trees (ET), Gradient Boosting regressor (GBR), Random Forest (RF) and XGBoost regressor (XGBR); one support vector machine model (SVR) and one artificial neural network model (ANN) were trained. The models were implemented using the ‘scikit-learn’ package (v1.2.2) and ‘xgboost’ package (v1.7.3) in Python (v3.9.16). Models were trained, and hyperparameters optimized, using five-fold cross-validated grid search with the coefficient of determination (R 2 ) as the evaluation metric. The ANN (scikit-learn’s MLPRegressor) was optimized using Bayesian Tree-Structured Parzen Estimator optimization with the ‘Optuna’ Python package (v3.2.0). Individual models were trained per attribute, and a multi-output model was trained on all attributes simultaneously.

Model dissection

GBR was found to outperform other methods, resulting in models with the highest average R 2 values in both trained panel and public review data sets. Impurity-based rankings of the most important predictors for each predicted sensorial trait were obtained using the ‘scikit-learn’ package. To observe the relationships between these chemical properties and their predicted targets, partial dependence plots (PDP) were constructed for the six most important predictors of consumer appreciation 74 , 75 .

The ‘SHAP’ package in Python (v0.41.0) was implemented to provide an alternative ranking of predictor importance and to visualize the predictors’ effects as a function of their concentration 68 .

Validation of causal chemical properties

To validate the effects of the most important model features on predicted sensory attributes, beers were spiked with the chemical compounds identified by the models and descriptive sensory analyses were carried out according to the American Society of Brewing Chemists (ASBC) protocol 90 .

Compound spiking was done 30 min before tasting. Compounds were spiked into fresh beer bottles, that were immediately resealed and inverted three times. Fresh bottles of beer were opened for the same duration, resealed, and inverted thrice, to serve as controls. Pairs of spiked samples and controls were served simultaneously, chilled and in dark glasses as outlined in the Trained panel section above. Tasters were instructed to select the glass with the higher flavor intensity for each attribute (directional difference test 92 ) and to select the glass they prefer.

The final concentration after spiking was equal to the within-style average, after normalizing by ethanol concentration. This was done to ensure balanced flavor profiles in the final spiked beer. The same methods were applied to improve a non-alcoholic beer. Compounds were the following: ethyl acetate (Merck KGaA, W241415), ethyl hexanoate (Merck KGaA, W243906), isoamyl acetate (Merck KGaA, W205508), phenethyl acetate (Merck KGaA, W285706), ethanol (96%, Colruyt), glycerol (Merck KGaA, W252506), lactic acid (Merck KGaA, 261106).

Significant differences in preference or perceived intensity were determined by performing the two-sided binomial test on each attribute.

Reporting summary

Further information on research design is available in the  Nature Portfolio Reporting Summary linked to this article.

Data availability

The data that support the findings of this work are available in the Supplementary Data files and have been deposited to Zenodo under accession code 10653704 93 . The RateBeer scores data are under restricted access, they are not publicly available as they are property of RateBeer (ZX Ventures, USA). Access can be obtained from the authors upon reasonable request and with permission of RateBeer (ZX Ventures, USA).  Source data are provided with this paper.

Code availability

The code for training the machine learning models, analyzing the models, and generating the figures has been deposited to Zenodo under accession code 10653704 93 .

Tieman, D. et al. A chemical genetic roadmap to improved tomato flavor. Science 355 , 391–394 (2017).

Article   ADS   CAS   PubMed   Google Scholar  

Plutowska, B. & Wardencki, W. Application of gas chromatography–olfactometry (GC–O) in analysis and quality assessment of alcoholic beverages – A review. Food Chem. 107 , 449–463 (2008).

Article   CAS   Google Scholar  

Legin, A., Rudnitskaya, A., Seleznev, B. & Vlasov, Y. Electronic tongue for quality assessment of ethanol, vodka and eau-de-vie. Anal. Chim. Acta 534 , 129–135 (2005).

Loutfi, A., Coradeschi, S., Mani, G. K., Shankar, P. & Rayappan, J. B. B. Electronic noses for food quality: A review. J. Food Eng. 144 , 103–111 (2015).

Ahn, Y.-Y., Ahnert, S. E., Bagrow, J. P. & Barabási, A.-L. Flavor network and the principles of food pairing. Sci. Rep. 1 , 196 (2011).

Article   CAS   PubMed   PubMed Central   Google Scholar  

Bartoshuk, L. M. & Klee, H. J. Better fruits and vegetables through sensory analysis. Curr. Biol. 23 , R374–R378 (2013).

Article   CAS   PubMed   Google Scholar  

Piggott, J. R. Design questions in sensory and consumer science. Food Qual. Prefer. 3293 , 217–220 (1995).

Article   Google Scholar  

Kermit, M. & Lengard, V. Assessing the performance of a sensory panel-panellist monitoring and tracking. J. Chemom. 19 , 154–161 (2005).

Cook, D. J., Hollowood, T. A., Linforth, R. S. T. & Taylor, A. J. Correlating instrumental measurements of texture and flavour release with human perception. Int. J. Food Sci. Technol. 40 , 631–641 (2005).

Chinchanachokchai, S., Thontirawong, P. & Chinchanachokchai, P. A tale of two recommender systems: The moderating role of consumer expertise on artificial intelligence based product recommendations. J. Retail. Consum. Serv. 61 , 1–12 (2021).

Ross, C. F. Sensory science at the human-machine interface. Trends Food Sci. Technol. 20 , 63–72 (2009).

Chambers, E. IV & Koppel, K. Associations of volatile compounds with sensory aroma and flavor: The complex nature of flavor. Molecules 18 , 4887–4905 (2013).

Pinu, F. R. Metabolomics—The new frontier in food safety and quality research. Food Res. Int. 72 , 80–81 (2015).

Danezis, G. P., Tsagkaris, A. S., Brusic, V. & Georgiou, C. A. Food authentication: state of the art and prospects. Curr. Opin. Food Sci. 10 , 22–31 (2016).

Shepherd, G. M. Smell images and the flavour system in the human brain. Nature 444 , 316–321 (2006).

Meilgaard, M. C. Prediction of flavor differences between beers from their chemical composition. J. Agric. Food Chem. 30 , 1009–1017 (1982).

Xu, L. et al. Widespread receptor-driven modulation in peripheral olfactory coding. Science 368 , eaaz5390 (2020).

Kupferschmidt, K. Following the flavor. Science 340 , 808–809 (2013).

Billesbølle, C. B. et al. Structural basis of odorant recognition by a human odorant receptor. Nature 615 , 742–749 (2023).

Article   ADS   PubMed   PubMed Central   Google Scholar  

Smith, B. Perspective: Complexities of flavour. Nature 486 , S6–S6 (2012).

Pfister, P. et al. Odorant receptor inhibition is fundamental to odor encoding. Curr. Biol. 30 , 2574–2587 (2020).

Moskowitz, H. W., Kumaraiah, V., Sharma, K. N., Jacobs, H. L. & Sharma, S. D. Cross-cultural differences in simple taste preferences. Science 190 , 1217–1218 (1975).

Eriksson, N. et al. A genetic variant near olfactory receptor genes influences cilantro preference. Flavour 1 , 22 (2012).

Ferdenzi, C. et al. Variability of affective responses to odors: Culture, gender, and olfactory knowledge. Chem. Senses 38 , 175–186 (2013).

Article   PubMed   Google Scholar  

Lawless, H. T. & Heymann, H. Sensory evaluation of food: Principles and practices. (Springer, New York, NY). https://doi.org/10.1007/978-1-4419-6488-5 (2010).

Colantonio, V. et al. Metabolomic selection for enhanced fruit flavor. Proc. Natl. Acad. Sci. 119 , e2115865119 (2022).

Fritz, F., Preissner, R. & Banerjee, P. VirtualTaste: a web server for the prediction of organoleptic properties of chemical compounds. Nucleic Acids Res 49 , W679–W684 (2021).

Tuwani, R., Wadhwa, S. & Bagler, G. BitterSweet: Building machine learning models for predicting the bitter and sweet taste of small molecules. Sci. Rep. 9 , 1–13 (2019).

Dagan-Wiener, A. et al. Bitter or not? BitterPredict, a tool for predicting taste from chemical structure. Sci. Rep. 7 , 1–13 (2017).

Pallante, L. et al. Toward a general and interpretable umami taste predictor using a multi-objective machine learning approach. Sci. Rep. 12 , 1–11 (2022).

Malavolta, M. et al. A survey on computational taste predictors. Eur. Food Res. Technol. 248 , 2215–2235 (2022).

Lee, B. K. et al. A principal odor map unifies diverse tasks in olfactory perception. Science 381 , 999–1006 (2023).

Mayhew, E. J. et al. Transport features predict if a molecule is odorous. Proc. Natl. Acad. Sci. 119 , e2116576119 (2022).

Niu, Y. et al. Sensory evaluation of the synergism among ester odorants in light aroma-type liquor by odor threshold, aroma intensity and flash GC electronic nose. Food Res. Int. 113 , 102–114 (2018).

Yu, P., Low, M. Y. & Zhou, W. Design of experiments and regression modelling in food flavour and sensory analysis: A review. Trends Food Sci. Technol. 71 , 202–215 (2018).

Oladokun, O. et al. The impact of hop bitter acid and polyphenol profiles on the perceived bitterness of beer. Food Chem. 205 , 212–220 (2016).

Linforth, R., Cabannes, M., Hewson, L., Yang, N. & Taylor, A. Effect of fat content on flavor delivery during consumption: An in vivo model. J. Agric. Food Chem. 58 , 6905–6911 (2010).

Guo, S., Na Jom, K. & Ge, Y. Influence of roasting condition on flavor profile of sunflower seeds: A flavoromics approach. Sci. Rep. 9 , 11295 (2019).

Ren, Q. et al. The changes of microbial community and flavor compound in the fermentation process of Chinese rice wine using Fagopyrum tataricum grain as feedstock. Sci. Rep. 9 , 3365 (2019).

Hastie, T., Friedman, J. & Tibshirani, R. The Elements of Statistical Learning. (Springer, New York, NY). https://doi.org/10.1007/978-0-387-21606-5 (2001).

Dietz, C., Cook, D., Huismann, M., Wilson, C. & Ford, R. The multisensory perception of hop essential oil: a review. J. Inst. Brew. 126 , 320–342 (2020).

CAS   Google Scholar  

Roncoroni, Miguel & Verstrepen, Kevin Joan. Belgian Beer: Tested and Tasted. (Lannoo, 2018).

Meilgaard, M. Flavor chemistry of beer: Part II: Flavor and threshold of 239 aroma volatiles. in (1975).

Bokulich, N. A. & Bamforth, C. W. The microbiology of malting and brewing. Microbiol. Mol. Biol. Rev. MMBR 77 , 157–172 (2013).

Dzialo, M. C., Park, R., Steensels, J., Lievens, B. & Verstrepen, K. J. Physiology, ecology and industrial applications of aroma formation in yeast. FEMS Microbiol. Rev. 41 , S95–S128 (2017).

Article   PubMed   PubMed Central   Google Scholar  

Datta, A. et al. Computer-aided food engineering. Nat. Food 3 , 894–904 (2022).

American Society of Brewing Chemists. Beer Methods. (American Society of Brewing Chemists, St. Paul, MN, U.S.A.).

Olaniran, A. O., Hiralal, L., Mokoena, M. P. & Pillay, B. Flavour-active volatile compounds in beer: production, regulation and control. J. Inst. Brew. 123 , 13–23 (2017).

Verstrepen, K. J. et al. Flavor-active esters: Adding fruitiness to beer. J. Biosci. Bioeng. 96 , 110–118 (2003).

Meilgaard, M. C. Flavour chemistry of beer. part I: flavour interaction between principal volatiles. Master Brew. Assoc. Am. Tech. Q 12 , 107–117 (1975).

Briggs, D. E., Boulton, C. A., Brookes, P. A. & Stevens, R. Brewing 227–254. (Woodhead Publishing). https://doi.org/10.1533/9781855739062.227 (2004).

Bossaert, S., Crauwels, S., De Rouck, G. & Lievens, B. The power of sour - A review: Old traditions, new opportunities. BrewingScience 72 , 78–88 (2019).

Google Scholar  

Verstrepen, K. J. et al. Flavor active esters: Adding fruitiness to beer. J. Biosci. Bioeng. 96 , 110–118 (2003).

Snauwaert, I. et al. Microbial diversity and metabolite composition of Belgian red-brown acidic ales. Int. J. Food Microbiol. 221 , 1–11 (2016).

Spitaels, F. et al. The microbial diversity of traditional spontaneously fermented lambic beer. PLoS ONE 9 , e95384 (2014).

Blanco, C. A., Andrés-Iglesias, C. & Montero, O. Low-alcohol Beers: Flavor Compounds, Defects, and Improvement Strategies. Crit. Rev. Food Sci. Nutr. 56 , 1379–1388 (2016).

Jackowski, M. & Trusek, A. Non-Alcohol. beer Prod. – Overv. 20 , 32–38 (2018).

Takoi, K. et al. The contribution of geraniol metabolism to the citrus flavour of beer: Synergy of geraniol and β-citronellol under coexistence with excess linalool. J. Inst. Brew. 116 , 251–260 (2010).

Kroeze, J. H. & Bartoshuk, L. M. Bitterness suppression as revealed by split-tongue taste stimulation in humans. Physiol. Behav. 35 , 779–783 (1985).

Mennella, J. A. et al. A spoonful of sugar helps the medicine go down”: Bitter masking bysucrose among children and adults. Chem. Senses 40 , 17–25 (2015).

Wietstock, P., Kunz, T., Perreira, F. & Methner, F.-J. Metal chelation behavior of hop acids in buffered model systems. BrewingScience 69 , 56–63 (2016).

Sancho, D., Blanco, C. A., Caballero, I. & Pascual, A. Free iron in pale, dark and alcohol-free commercial lager beers. J. Sci. Food Agric. 91 , 1142–1147 (2011).

Rodrigues, H. & Parr, W. V. Contribution of cross-cultural studies to understanding wine appreciation: A review. Food Res. Int. 115 , 251–258 (2019).

Korneva, E. & Blockeel, H. Towards better evaluation of multi-target regression models. in ECML PKDD 2020 Workshops (eds. Koprinska, I. et al.) 353–362 (Springer International Publishing, Cham, 2020). https://doi.org/10.1007/978-3-030-65965-3_23 .

Gastón Ares. Mathematical and Statistical Methods in Food Science and Technology. (Wiley, 2013).

Grinsztajn, L., Oyallon, E. & Varoquaux, G. Why do tree-based models still outperform deep learning on tabular data? Preprint at http://arxiv.org/abs/2207.08815 (2022).

Gries, S. T. Statistics for Linguistics with R: A Practical Introduction. in Statistics for Linguistics with R (De Gruyter Mouton, 2021). https://doi.org/10.1515/9783110718256 .

Lundberg, S. M. et al. From local explanations to global understanding with explainable AI for trees. Nat. Mach. Intell. 2 , 56–67 (2020).

Ickes, C. M. & Cadwallader, K. R. Effects of ethanol on flavor perception in alcoholic beverages. Chemosens. Percept. 10 , 119–134 (2017).

Kato, M. et al. Influence of high molecular weight polypeptides on the mouthfeel of commercial beer. J. Inst. Brew. 127 , 27–40 (2021).

Wauters, R. et al. Novel Saccharomyces cerevisiae variants slow down the accumulation of staling aldehydes and improve beer shelf-life. Food Chem. 398 , 1–11 (2023).

Li, H., Jia, S. & Zhang, W. Rapid determination of low-level sulfur compounds in beer by headspace gas chromatography with a pulsed flame photometric detector. J. Am. Soc. Brew. Chem. 66 , 188–191 (2008).

Dercksen, A., Laurens, J., Torline, P., Axcell, B. C. & Rohwer, E. Quantitative analysis of volatile sulfur compounds in beer using a membrane extraction interface. J. Am. Soc. Brew. Chem. 54 , 228–233 (1996).

Molnar, C. Interpretable Machine Learning: A Guide for Making Black-Box Models Interpretable. (2020).

Zhao, Q. & Hastie, T. Causal interpretations of black-box models. J. Bus. Econ. Stat. Publ. Am. Stat. Assoc. 39 , 272–281 (2019).

Article   MathSciNet   Google Scholar  

Hastie, T., Tibshirani, R. & Friedman, J. The Elements of Statistical Learning. (Springer, 2019).

Labrado, D. et al. Identification by NMR of key compounds present in beer distillates and residual phases after dealcoholization by vacuum distillation. J. Sci. Food Agric. 100 , 3971–3978 (2020).

Lusk, L. T., Kay, S. B., Porubcan, A. & Ryder, D. S. Key olfactory cues for beer oxidation. J. Am. Soc. Brew. Chem. 70 , 257–261 (2012).

Gonzalez Viejo, C., Torrico, D. D., Dunshea, F. R. & Fuentes, S. Development of artificial neural network models to assess beer acceptability based on sensory properties using a robotic pourer: A comparative model approach to achieve an artificial intelligence system. Beverages 5 , 33 (2019).

Gonzalez Viejo, C., Fuentes, S., Torrico, D. D., Godbole, A. & Dunshea, F. R. Chemical characterization of aromas in beer and their effect on consumers liking. Food Chem. 293 , 479–485 (2019).

Gilbert, J. L. et al. Identifying breeding priorities for blueberry flavor using biochemical, sensory, and genotype by environment analyses. PLOS ONE 10 , 1–21 (2015).

Goulet, C. et al. Role of an esterase in flavor volatile variation within the tomato clade. Proc. Natl. Acad. Sci. 109 , 19009–19014 (2012).

Article   ADS   CAS   PubMed   PubMed Central   Google Scholar  

Borisov, V. et al. Deep Neural Networks and Tabular Data: A Survey. IEEE Trans. Neural Netw. Learn. Syst. 1–21 https://doi.org/10.1109/TNNLS.2022.3229161 (2022).

Statista. Statista Consumer Market Outlook: Beer - Worldwide.

Seitz, H. K. & Stickel, F. Molecular mechanisms of alcoholmediated carcinogenesis. Nat. Rev. Cancer 7 , 599–612 (2007).

Voordeckers, K. et al. Ethanol exposure increases mutation rate through error-prone polymerases. Nat. Commun. 11 , 3664 (2020).

Goelen, T. et al. Bacterial phylogeny predicts volatile organic compound composition and olfactory response of an aphid parasitoid. Oikos 129 , 1415–1428 (2020).

Article   ADS   Google Scholar  

Reher, T. et al. Evaluation of hop (Humulus lupulus) as a repellent for the management of Drosophila suzukii. Crop Prot. 124 , 104839 (2019).

Stein, S. E. An integrated method for spectrum extraction and compound identification from gas chromatography/mass spectrometry data. J. Am. Soc. Mass Spectrom. 10 , 770–781 (1999).

American Society of Brewing Chemists. Sensory Analysis Methods. (American Society of Brewing Chemists, St. Paul, MN, U.S.A., 1992).

McAuley, J., Leskovec, J. & Jurafsky, D. Learning Attitudes and Attributes from Multi-Aspect Reviews. Preprint at https://doi.org/10.48550/arXiv.1210.3926 (2012).

Meilgaard, M. C., Carr, B. T. & Carr, B. T. Sensory Evaluation Techniques. (CRC Press, Boca Raton). https://doi.org/10.1201/b16452 (2014).

Schreurs, M. et al. Data from: Predicting and improving complex beer flavor through machine learning. Zenodo https://doi.org/10.5281/zenodo.10653704 (2024).

Download references

Acknowledgements

We thank all lab members for their discussions and thank all tasting panel members for their contributions. Special thanks go out to Dr. Karin Voordeckers for her tremendous help in proofreading and improving the manuscript. M.S. was supported by a Baillet-Latour fellowship, L.C. acknowledges financial support from KU Leuven (C16/17/006), F.A.T. was supported by a PhD fellowship from FWO (1S08821N). Research in the lab of K.J.V. is supported by KU Leuven, FWO, VIB, VLAIO and the Brewing Science Serves Health Fund. Research in the lab of T.W. is supported by FWO (G.0A51.15) and KU Leuven (C16/17/006).

Author information

These authors contributed equally: Michiel Schreurs, Supinya Piampongsant, Miguel Roncoroni.

Authors and Affiliations

VIB—KU Leuven Center for Microbiology, Gaston Geenslaan 1, B-3001, Leuven, Belgium

Michiel Schreurs, Supinya Piampongsant, Miguel Roncoroni, Lloyd Cool, Beatriz Herrera-Malaver, Florian A. Theßeling & Kevin J. Verstrepen

CMPG Laboratory of Genetics and Genomics, KU Leuven, Gaston Geenslaan 1, B-3001, Leuven, Belgium

Leuven Institute for Beer Research (LIBR), Gaston Geenslaan 1, B-3001, Leuven, Belgium

Laboratory of Socioecology and Social Evolution, KU Leuven, Naamsestraat 59, B-3000, Leuven, Belgium

Lloyd Cool, Christophe Vanderaa & Tom Wenseleers

VIB Bioinformatics Core, VIB, Rijvisschestraat 120, B-9052, Ghent, Belgium

Łukasz Kreft & Alexander Botzki

AB InBev SA/NV, Brouwerijplein 1, B-3000, Leuven, Belgium

Philippe Malcorps & Luk Daenen

You can also search for this author in PubMed   Google Scholar

Contributions

S.P., M.S. and K.J.V. conceived the experiments. S.P., M.S. and K.J.V. designed the experiments. S.P., M.S., M.R., B.H. and F.A.T. performed the experiments. S.P., M.S., L.C., C.V., L.K., A.B., P.M., L.D., T.W. and K.J.V. contributed analysis ideas. S.P., M.S., L.C., C.V., T.W. and K.J.V. analyzed the data. All authors contributed to writing the manuscript.

Corresponding author

Correspondence to Kevin J. Verstrepen .

Ethics declarations

Competing interests.

K.J.V. is affiliated with bar.on. The other authors declare no competing interests.

Peer review

Peer review information.

Nature Communications thanks Florian Bauer, Andrew John Macintosh and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary information, peer review file, description of additional supplementary files, supplementary data 1, supplementary data 2, supplementary data 3, supplementary data 4, supplementary data 5, supplementary data 6, supplementary data 7, reporting summary, source data, source data, rights and permissions.

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ .

Reprints and permissions

About this article

Cite this article.

Schreurs, M., Piampongsant, S., Roncoroni, M. et al. Predicting and improving complex beer flavor through machine learning. Nat Commun 15 , 2368 (2024). https://doi.org/10.1038/s41467-024-46346-0

Download citation

Received : 30 October 2023

Accepted : 21 February 2024

Published : 26 March 2024

DOI : https://doi.org/10.1038/s41467-024-46346-0

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

By submitting a comment you agree to abide by our Terms and Community Guidelines . If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Quick links

• Explore articles by subject
• Guide to authors
• Editorial policies

Sign up for the Nature Briefing: Translational Research newsletter — top stories in biotechnology, drug discovery and pharma.

#SchoolsNotPrisons: Attend the Informatics Seminar on Participatory Action Research

Don’t miss the april 5 talk by dr. david turner iii, which doubles as the keynote for uci’s third datafication and community activism workshop..

• Share on Facebook
• Share on Twitter
• Share on LinkedIn
• Share through Email
• Copy permalink

You’re invited to attend an engaging talk at UC Irvine by Dr. David Turner III , senior advisor at the Alliance for Boys and Men of Color, on Friday, April 5, 2024 at 2 p.m. Turner will discuss how he has led efforts to develop internal infrastructure as well as campaigns that uplift community schools by transforming discipline and decriminalizing youth and communities of color.

#SchoolsNotPrisons “ #SchoolsNotPrisons: Participatory Action Research Notes from the abolitionist movement to reimagine public safety in The Carceral State ” will take place in DBH 6011 as part the Informatics Seminar Series of UCI’s Donald Bren School of Information and Computer Sciences ( ICS ). It also serves as the keynote for DCA 2024 , the third Datafication and Community Activism workshop at UCI.

Turner will detail how communities have worked collaboratively to support the #SchoolsNotPrisons movement, replacing policies that lead to mass incarceration with social welfare programs that benefit students. Drawing from his 12 years of organizing experience and field notes, he will share how he has helped local youth gather more than 30,000 participatory action research surveys and create successful campaigns at nearly every level of government. In particular, he will answer questions such as

• What strategies do these movements use?
• How do these movements incorporate direct action organizing, strategic policy demands, electoral strategies, and research to facilitate social change?

“It’s such an important talk for both undergrad and grad students as well as professors — for pretty much anybody in the university interested in looking at inspirational models of how to partner with our communities in a way that honors their questions, ideas and demands,” says Uriel Serrano , an ICS postdoc researcher in the Department of Informatics. “We as researchers tend to come up with the questions first, and then enter spaces and communities, and I think what Dr. Turner does is oftentimes even the questions and the projects themselves are co-created with community, including young people.”

Datafication and Community Activism

Serrano is leading the two-day DCA workshop in collaboration with Informatics Professor Roderic Crooks , who started DCA in 2019 to help scholars and the broader tech sector better understand the “datafication” of society — that is, the ways in which more and more parts of our public and private lives are mediated by data.

“In starting the workshop, I was looking for a way to describe the relationship between minoritized communities and datafication,” says Crooks. The workshop brings together scholars, journalists, graduate students, artists, adjunct lecturers, community activists and data/information professionals in support of activist responses to datafication in minoritized communities. “To date, we have been fortunate to work with many national and regional organizations who are generally interested in digital data but are specifically concerned with civic education, abolition of police, racial discrimination in computational systems, data-driven government services, reintegration of formerly incarcerated people, economic empowerment, and many other issues of interest to working-class communities of color.”

Crooks and Serrano are working to expose and address the harm caused by datafication and to support ethical data collection and analysis for community-based advocacy projects. “We’re interviewing about a hundred community organizers across the country and asking questions about their data practices,” says Serrano. “What we’ve learned is that many community organizers are looking for spaces where they can learn from other organizers, like Dr. Turner, to gain insight into tools and to learn how to write surveys, conduct focus groups, ask research questions and design participation research projects.”

The DCA workshop, including its keynote, thus facilitates conversations between community organizers and academics with access to resources and information that could inform advocacy work. “Community organizers are some of the best researchers and ethnographers I know. They’re already thinking very deeply about social issues,” says Serrano. “I can only imagine what it would mean for them to have access to data analysis software or other tools readily available in academia to support their work. The heart of DCA is bringing people together to facilitate these connections.”

Learn more about Dr. Turner and his #SchoolsNotPrisons talk and plan to attend! The 2-3 p.m. seminar will be followed by a one-hour networking session.

— Shani Murray

Related Posts

A multitasker’s guide to regaining focus (the new york times), haley marts highlights the benefits of experiential learning, statistics ph.d. students win distinguished paper awards, ics researchers awarded best technical poster at ndss, defeating deepfakes at the source, guiding light (hana hou).