Mapping supply chain collaboration research: a machine learning-based literature review

ABSTRACT

Supply chain collaboration has been widely discussed in the literature. With this maturity comes a plethora of heterogeneous research that is difficult to manage and navigate. This paper, therefore, applies a novel literature review approach based on text mining analyzing 10,556 articles to provide an overview of previous research themes and future directions of the field. The applied method enables researchers to systematically analyze and structure larger samples of research publications. It allocates articles to thematic clusters using a visual hierarchical clustering approach and subsequently aggregates them into nine overarching themes to determine potential research and insights for practice. Developments regarding research interest and attention within these themes are examined and journals publishing the most impactful articles are identified. The paper thus contributes to the field of Supply Chain Collaboration research by mapping its evolvement over the last five years and by deriving a research agenda for the next decade.

KEYWORDS:

• Supply chain collaboration
• supply chain management
• literature review
• text mining
• machine learning
• collaborative supply chain

1. Introduction

The broad and growing field of Supply Chain Management (SCM) and logistics research (Kohl and Pfretzschner 2018) also encompasses Supply Chain Collaboration (SCC) as an area of interest that has been researched extensively over the last decades and become a driver for organisational success and goal attainment (Singh, Garg, and Sachdeva 2018). The growing application and relevance of the SCC research field are supported by the numerous advantages of the extension of logistics towards a more comprehensive approach and network perspective (Christopher 2016; Eßig, Hofmann, and Stölzle 2013; Venegas Vallejos, Matopoulos, and Greasley 2020). Innovation and optimisation efforts such as collaborative SCM are considered to be of great relevance for the survival and thriving of organisations as they may constitute a competitive advantage (Christopher 2000; Seo, Dinwoodie, and Roe 2016; Van Lancker et al. 2016).

Despite the research field’s rich and long history as well as the large number of publications dedicated to SCC, there are only few attempts to systematically analyze this comprehensive topic from a bird’s eye perspective. Review articles often focus on specific aspects of SCC, such as readiness to collaboration (e.g. Singh, Garg, and Sachdeva 2018), sustainability (e.g. Chen et al. 2017), certain contexts or industries (e.g. Aktas, Bourlakis, and Zissis 2020; Badraoui, Van der Vorst, and Boulaksil 2020; Kim, Dinwoodie, and Roh 2020; Wandfluh, Hofmann, and Schoensleben 2016) or collaborative techniques (e.g. Nimmy, Chilkapure, and Pillai 2019). In addition, the SCC literature field is relatively broad, transdisciplinary and has grown substantially over the last few years. Thus, a consolidation of the fragmented literature base yields potential for providing relevant insights into the main research streams as well as prospective future research opportunities within the SCC field.

The purpose of this paper is to provide a comprehensive overview of the heterogenous research field of SCC based on a text mining and machine learning literature review (MLR) as proposed by Reuther (2019). Thus, this paper consolidates and extends previous research and contributes to SCC research not only by providing an exhaustive survey of the available literature but also by using a novel approach to enable a better understanding of this extensive topic. Moreover, the review approach empowers the authors to provide a clear map of current issues and presents future areas of SCC interest. The MLR method enables researchers to systematically analyze large samples of research publications in an accessible way, without the need for in-depth methodological knowledge in data-mining, by using the open-source software Orange (Demsar et al. 2013). As we find that the SCC research field has become rather complex, it would be impossible to conduct a thorough and rigorous review of all papers within the body of knowledge without the application of text mining and machine learning tools.

In the remainder of the paper, the approach and main results of the MLR are presented. This review discusses how the field has developed over time in terms of research interest and attention based on a bibliometric analysis, thus informing both research and practice and promoting the further application of SCC. The MLR identifies 46 clusters in the SCC literature, which have subsequently been merged into nine overarching themes. The findings are synthesised and conceptualised in a SCC research agenda and the paper thus contributes to a better understanding of supply chain types, network management, supply chain optimisation, green and sustainable SCM, relationship management, supply chain context and research, supply chain development and digital supply chains.

2. Literature review approach

Traditional literature review forms include narrative reviews (Baker 2016), systematic reviews (Denyer and Tranfield 2009) and bibliometric studies (Zupic and Čater 2014). Systematic reviews aim to provide and describe available knowledge for a specific practice by analyzing and summarising the existing literature (Briner and Denyer 2012; Fink 2019) and provide a valuable source for policy, practice and academic research (Okoli and Schabram 2010; Petticrew and Roberts 2006). The necessity of conducting systematic reviews is further amplified by the growth in number and length as well as the increasing complexity of academic and other publications (Fettke 2006; vom Brocke et al. 2015) due to the increasing pressure to publish (‘publish or perish’) and the pursuit of theoretical contributions (Bothello and Roulet 2018). Strongly increasing publication numbers and improved access have led to a larger, more transdisciplinary, and more diverse body of literature throughout the disciplines, making the application of traditional review forms more difficult and, particularly when an overarching perspective relying on a wider range of sources is needed, less effective. Due to these developments, the analysis, integration and critical appraisal of the existing body of literature has become all the more relevant (Fettke 2006) and literature reviews have proven useful as a key tool to manage the diversity of knowledge (Fink 2019; Schryen 2015; Tranfield, Denyer, and Smart 2003b). While systematic literature reviews are widespread in the management literature today (Kunisch et al. 2018), they tend to analyze less than one hundred up to a few hundred articles. Bibliometric studies appear to include more articles, but mostly not more than 1000.

The MLR approach used in this paper combines the rigorous approach and content analysis of systematic reviews with the quantitative assessment of bibliometric studies. Thus, the analysis of an even larger number of articles is enabled, despite using comparably fewer exclusion criteria. It consequently facilitates the identification, consolidation, and combination of the main research streams within the field of SCC research. Furthermore, new research avenues are potentially detected. This paper thus applies a MLR (Reuther 2019) using systematic search procedures based on Tranfield, Denyer, and Smart (2003a) to provide the literature basis for the review as well as the open-source machine learning and data visualisation software Orange (Demsar et al. 2013) for the definition and analysis of clusters and overarching research themes (OTs). Orange was selected due to several reasons. First, this tool does not require any coding competencies but creates a Python script at the backend, making it accessible to both researchers with a non-informatics or non-mathematics backgrounds as well as social science researchers. Thus, researchers of all fields are provided with the means to make use of text data mining and machine learning, whereby the use of a ready-made software solution represents a compromise between the best quality of the clustering and precisely this accessibility. Second, despite its user friendliness, Orange is transparent and editable when necessary, as the underlying coding is open source. Third, the software was developed in a university background and is supported through continuous development, updates and user recommendations. Thus, this review does not aim at a deeper understanding of individual research areas within the SCC field, but rather at providing a bird’s eye perspective of the thematic developments within the past few years at a higher level of abstraction.

Compared to traditional systematic literature reviews, the applied approach based on machine learning and text mining tools can be used to handle large quantities of publications to conduct comprehensive reviews. Due to the widespread availability of semi-structured data in electronic documents (Salloum et al. 2018) and the accelerating growth of publications, automated data extraction for literature analysis and synthesis is necessary. Text mining can be applied to all phases of a traditional systematic review, namely planning, conducting and reporting, post-review activities (Feng et al. 2017a). The MLR approach mainly applies text mining and machine learning for the conducting phase, which includes literature selection and assessment as well as data extraction. Due to the combination of machine learning and text mining tools with traditional review approaches, the MLR incorporates the benefits enabled by manual interventions, subjective experience, and knowledge-based decisions.

Following the definition of the review scope and purpose, the database search is executed. Therefore, relevant keywords need to be tested and constructed into search terms. The Scopus database is used to identify the relevant literature for this study due to its comprehensiveness and the data export capacity. Scopus offers access to more than 23,500 peer-reviewed journals covering social sciences, physical sciences, health sciences and life sciences (Elsevier 2019). Another advantage compared to other popular databases is that in addition to its coverage, researchers can easily export large quantities of citation information, thus facilitating the subsequent clustering and bibliometric analysis. The applied search phrase for this review is composed of two segments which are connected using the Boolean ‘AND’ operator:

(TITLE-ABS-KEY (organi*ational OR relationship OR coordinat* OR collaborat* OR cooperat* OR ‘logistic* service*’ OR network*) AND TITLE-ABS-KEY (‘supply-chain-management’ OR ‘supply-chain management’ OR ‘supply chain management’ OR ‘supply chain*’ OR ‘supply-chain*’ OR supply AND chain*))

The first segment summarises terms related to collaboration: organisational, relationship, coordination, collaboration, cooperation, logistics service, and network. The second segment covers SCM and includes supply chain management, supply chain, supply, and chain. Within the segments, individual terms are connected through Boolean ‘OR’ operators and wildcards are used where appropriate to take spelling and ending variations into account. Additionally, the search results are restricted to published journal articles from 2015 to 2020 (up to the search date July 9th, 2020) due to the high increase of publication numbers in recent years and existing restrictions regarding computing power. In addition, previous reviews have sufficiently covered developments further in the past, while novel research trends such as the integration of blockchain or Artificial Intelligence can be observed in this more recent time period. Only English language publications from an engineering, business, computer science, decision science, social sciences or economics context are included. The resulting list of 10,556 titles of the relevant Scopus literature search results are then exported as a CSV file and formatted as an Excel file. Then, a workflow containing various text-mining tools (widgets) is developed within the Orange software tool. The principal process flow including the Orange workflow used to identify the clusters in recent literature on SCC is illustrated in Figure 1.

Figure 1. General process flow including the orange workflow for the MLR (Reuther 2019).

Figure 1 visualises the main process steps of the literature review. Following the definition of the review scope and purpose, the database search is executed in the Scopus literature database using the keywords (see search string above). The relevant literature is identified based on inclusion criteria and subsequently imported into the text-mining software Orange. The first Orange workflow widget is corpus, through which the Excel file containing the article titles are imported. To pre-process the text, the pre-processor splits the text into smaller units (i.e. tokens). The pre-processor also uses the Porter stemming algorithm to remove morphological and inflectional endings from single words, thus normalising the text from all titles. During this step, it is also possible to remove words without meaning for the research context, such as pronouns or articles. As a last pre-processing step, the widget creates one-grams and two-grams based on the tokens (n-grams are contiguous sequences of n items from a text) which enables the software to start the text mining process. The next text mining widget is the so-called bag of words, adding word counts for each data instance created in the previous step. As a result of this widget, the data are rearranged to form rows and columns of entities and features. The rows of the database are formed by entities (i.e. individual titles) and the columns are formed by features (i.e. every word that appears in all titles). Following the bag of words widget, the distances widget is used to calculate Euclidian distances¹ between the rows of the dataset (i.e. the titles). Thereby, a distance matrix concerning all titles is created. Based on the distance matrix, the hierarchical clustering is carried out. The resulting dendrogram visualises the allocation of all titles to different clusters. It is also possible to create a silhouette plot which indicates the similarity of titles. Based on the dendrogram created through hierarchical clustering, word clouds and concordances of the most frequently appearing words and word combinations in the respective cluster and be used by the researcher to form thematic constructs. The clusters are analyzed separately by using the select rows widget. In case of unclarity of the clusters, the titles included in the construct can be used as a new corpus to derive sub-clusters. This process aims at determining more precise clusters by re-executing the steps described above. Following the cluster definition, they are merged into nine OTs for further analysis.

3. Literature review results

It was found that within the last five and a half years (2015–2020) 10,556 published journal articles were identified on SCC in the Scopus database. Following the methodological approach outlined in the previous section, this paper provides a descriptive overview of SCC research published over the last five and a half years by analyzing 10,556 articles from the Scopus database. These articles have been published in 1660 journals. Based on the entire sample, only 13 journals account for more than 100 articles each (see Figure 2), while 1473 journals contribute less than 10 publications. In addition, the quantitative development of the research field publications over time, including an estimate for 2020 and 2021, is presented in Figure 3.

Figure 2. Overview of Scopus export major journals.

Figure 3. Development of research field publications over time.

According to the MLR workflow explained in section 2, the Scopus export Excel file is imported through the corpus widget and subsequently prepared for the clustering process using the pre-processor, bag of words and distance matrix widgets. The hierarchical cluster analysis using Orange resulted in the identification of 46 clusters. Table 1 presents the thematic content of these clusters as well as their clarity and size. The cluster content is determined by analyzing the words most frequently occurring in the article titles contained within the respective cluster. The perceived clarity is judged based on the homogeneity of the contained article titles concerning the cluster content using the silhouette plot and the word cloud depicting the most frequently appearing words within each cluster. The silhouette plot is a graphical representation of the silhouette score as a measure of how similar an object is to its own cluster in comparison to other clusters. More homogenous clusters are described as ‘clear’ whereas clusters containing many titles of more varying focus are characterised as ‘fuzzy’. The themes derived directly from the cluster content have subsequently been merged into nine OTs:

• Supply chain types

• Network management

• Supply chain optimisation

• Green and sustainable supply chain management

• Relationship management

• Supply chain context

• Research

• Supply chain development

• Digital supply chain

Table 1. Supply chain collaboration thematic clusters, derived themes, and overarching themes.

Cluster no	Cluster content	Clarity (clear/fuzzy)	No of papers	Derived themes	Overarching themes (OT)	OT no
1	Closed-loop	c	267	Closed-loop supply chain	Supply chain types	1
2	Network	c	191	Network management	Network management	2
3	Design	c	266	Network design	Network management	2
4	Decision (making/support/model)	c	110	Decision making	Supply chain optimisation	3
5	Approach	c	216	Supply chain management approach	Network management	2
6	Model	c/f	227	Supply chain model	Network management	2
7	Problem optimisation	c	306	Optimisation	Supply chain optimisation	3
8	Green (management/performance/practices)	c	168	Green supply chain management	Green and sustainable supply chain management	4
9	(Performance) Management	c	349	Performance management	Supply chain optimisation	3
10	Food	c	177	Food supply chain	Supply chain types	1
11	Service	c	135	Logistics service providers	Relationship management	5
12	Logistics (reverse/network/third party)	c/f	172	Logistics	Network management	2
13	Demand and price	c/f	111	Demand and price management	Supply chain optimisation	3
14	Inventory	c	108	Inventory management	Supply chain optimisation	3
15	Channel	c	62	Channel type	Supply chain types	1
16	Coordination	c	207	Relationship management	Relationship management	5
17	Carbon emissions	c	87	Sustainability	Green and sustainable supply chain management	4
18	Sustainable management and development	c	141	Sustainability	Green and sustainable supply chain management	4
19	Risk	c	76	Risk management	Supply chain optimisation	3
20	Sustainability, environmental (assessment)	c/f	210	Sustainability	Green and sustainable supply chain management	4
21	Business	c/f	52	Business management	Supply chain context	6
22	Research, literature review	c	78	Literature review	Research	7
23	System management/approach	c	105	System management	Network management	2
24	Factors (structural, success)	c	122	Supply chain factors	Supply chain development	8
25	Game	c/f	114	Game approach	Relationship management	5
26	N.a.	f	3290	N.a.	N.a.
27	Networks	c	124	Network management	Network management	2
28	Production	c/f	138	Production management	Supply chain context	6
29	(Corporate) Social (responsibility)	c	125	Corporate Social Responsibility	Green and sustainable supply chain management	4
30	Value chain/creation	c	128	Value chain management	Supply chain development	8
31	Chains	c/f	239	Supply chains	Network management	2
32	CO2	c	38	Carbon emissions	Green and sustainable supply chain management	4
33	Data	c	97	Data management	Digital supply chain	9
34	Analysis	c	222	Data management	Digital supply chain	9
35	Empirical	c	96	Empirical research	Research	7
36	Case study	c	263	Empirical research	Research	7
37	Supplier selection	c	88	Supplier selection	Supply chain optimisation	3
38	Supplier	c/f	221	Relationship management	Relationship management	5
39	Manufacturing	c	128	Manufacturing management	Supply chain context	6
40	Industry	c	230	Industry management	Supply chain context	6
41	Information	c	194	Data management	Digital supply chain	9
42	Role	c	136	Relationship management	Relationship management	5
43	Small and medium sized enterprises	c	42	Relationship management	Supply chain context	6
44	Performance	c/f	267	Performance Management	Supply chain optimisation	3
45	Innovation	c	80	Innovation management	Supply chain development	8
46	Performance integration, integration effect	c	353	Supply chain integration	Supply chain optimisation	3

These OTs aggregate those thematic clusters that apparently refer to similar topics. For instance, the OT4 green and sustainable supply chain management comprises clusters green (management/ performance/practices) (C8), carbon emissions (C17), sustainable management and development (C18), sustainability, environmental (assessment) (C20), (corporate) social (responsibility) (C29), CO₂ (C32).

Table 1 reveals some basic characteristics of SCC research. Most of the clusters seem clear regarding their thematic content and contain between 100 and 400 articles. The largest thematic clusters among those with high clarity are C46 performance integration, integration effect, C9 (performance) management and C7 problem optimisation. Only some clusters are comparably small and comprise less than 100 publications: C33 data, C35 empirical, C37 supplier selection, C17 carbon emissions, C45 innovation, C22 research, literature review, C19 risk, C15 channel, C21 business, 43 small and medium-sized enterprises, 32 CO₂. General cluster size ranges from 38 articles (C32) to 353 articles (C46). Those clusters with a clear thematic focus make up 52,25% of the total amount of articles. Similarly, just a few clusters are less clear, the largest ones including C44 performance, C31 chains, C6 model, these account for 16,59% of all publications. One cluster, which is also the largest one with 3,290 articles (31,17% of the total number of publications), has been classified as fuzzy since the thematic content is quite mixed (C26). In the MLR, receiving one cluster of this size is typical and these fuzzy clusters tend to increase in size in relation to the research field heterogeneity. However, the purpose of the MLR is to identify the existing clusters and OTs within the literature pool, not to assign each individual article to a cluster. The authors furthermore assume that the distribution of content in the large fuzzy cluster likely reflects the content of the other clusters. Thus, an exclusion of the fuzzy cluster from the further in-depth analysis still yields valid findings for the research field.

4. Bibliometric analysis

4.1. Overview of the bibliometric analysis

The bibliometric analysis focuses on several key figures based on the number of published articles, citations, authors or journals and their relation (Donthu et al. 2021; Gingras 2016; Linnenluecke, Marrone, and Singh 2019). Thus, the topicality of the articles and clusters is investigated on the level of the nine OTs derived from the literature sample and assumptions about the attention, relevance or interest in an OT or journal are made. This bibliometric analysis uses the Scopus database export of basic article attributes and the clustering results from the analysis conducted in the Orange software to provide the database for further processing using Excel.

4.2. Influential journals

The analysis reveals that most articles relevant to the purpose of this paper have been published in the journals International Journal of Supply Chain Management, Journal of Cleaner Production, International Journal of Production Economics, Sustainability (Switzerland) and International Journal of Production Research. Table 2 provides an overview of the top 10 journals regarding the total number of citations as well as the number of publications and the average number of citations per article. The citations forming the basis of this table are directly related to SCC. Concerning the average citation per article within the different journals, the highest count is achieved by the Journal of Cleaner Production and International Journal of Production Economics. The table also appears to show that those journals that publish the most articles also have the highest number of average citations per articles and could thus be assumed to be where the essential debates take place. The overall similar number of average citations per publications also indicates that the top journals within the SCC research field are comparably equally influential. A notable exception is the journal Sustainability (Switzerland) which only reaches an average citation count of 6.01 per article despite a relatively high number of publications, which may indicate that researchers might be more likely to successfully publish SCC articles in this journal.

Table 2. Main metrics for the top 10 journals regarding the total number of citations derived from the literature review sample.^a

Journal name	No of citations	No of publications	Average citations per publication
Journal of Cleaner Production	11,173	488	22.90
International Journal of Production Economics	9884	429	23.04
International Journal of Production Research	4895	290	16.88
European Journal of Operational Research	3640	165	22.06
Computers and Industrial Engineering	2882	226	12.75
Sustainability (Switzerland)	2289	381	6.01
Transportation Research Part E: Logistics and Transportation Review	2234	112	19.95
Supply Chain Management	2185	144	15.17
Annals of Operations Research	1852	107	17.31
International Journal of Operations and Production Management	1732	101	17.15

^a Database search using Scopus: (TITLE-ABS-KEY (organi*ational OR relationship OR coordinat* OR collaborat* OR cooperat* OR ‘logistic* service*' OR network*) AND TITLE-ABS-KEY (‘supply-chain-management’ OR ‘supply-chain management' OR ‘supply chain management' OR ‘supply chain*' OR ‘supply-chain*’ OR supply AND chain*)).

4.3. Research interest

In addition to the journal characteristics, the research interest, which is expressed by the total number of published articles in an OT, is analyzed. The fuzzy cluster (C26) is for the most part not included in the bibliometric analysis as it is assumed that it contains a lot of articles that overlap with the other thematic clusters and would confound calculations and thus distort results (for example calculations regarding the number of articles). As the fuzzy cluster (C26) is excluded, the new total of nine OTs contains 45 thematic clusters and 7226 articles. The analysis of the research interest and its development over the last five years aims at providing some insights into the thematic focus of the SCC research community. To investigate the research interest, the total number of published articles per year and OT, the average number of articles per year, and the average number of articles per OT are calculated. Figure 4 shows the total number of published articles per year and OT. An overall increase in the number of publications is clearly visible, with a more intensive rise observable from 2017 onwards. Although a similar tendency can be observed, the relative increase in publications differs between themes. For instance, OT2 network management, OT3 supply chain optimisation, OT4 green and sustainable supply chain management, and OT5 relationship management exhibit especially steep increases. This leads to the assumption that the SCC research community’s interest is directed towards higher efficiency and sustainability within supply networks material flows as well as inter-organisational relations.

Figure 4. Number of publications per year and overarching theme (OT).

Table 3 illustrates the relative OT size based on the yearly average (left side) and OT average (right side). Here, only the full years (2015–2019) are included for a better comparability. Following the advice of Linnenluecke, Marrone, and Singh (2019), a colour-scaled heat map is used as a bibliometric tool to visualise the differences regarding the research interest.

Table 3. Number of articles per OT and year: relative research interest highlighted in nuances (blue = above average, grey = below average) based on the yearly average (left side) and OT average (right side).

For yearly average, the vertically highlighted colour gradient visualises the relative OT sizes for each year with darker blue colour nuances marking OT sizes above the respective year’s average and grey colour nuances identifying OTs containing less than the average number of articles. For instance, the yearly OT size average is 110.11 publications in 2015. Thus, the number of publications in OT5 (113) is coloured in light blue as this matches the year’s average, whereas OT3 (260 articles) is coloured a darker nuance of blue and OT8 (45 articles) in grey. The visualisation clearly shows that the constantly largest OTs are OT2 network management and OT3 supply chain optimisation, which could indicate a high relevance of collaborative SCM for performance- and optimisation-related aspects. From the observed heterogenous OTs, a variable research interest in diverse aspects of SCC research could be deduced. While topics such as performance management and supply chain optimisation appear to be of central interest, outliers which arouse limited interest include research approaches and methods, such as literature review, as well as supply chain development topics, such as structural and success factors and innovation management. This indicates that these topics constitute research niches in the field of SCC that could potentially grow over the next years. For instance, literature reviews are important tools for the consolidation and development of research fields (Rowe 2014) and thus likely to receive more research interest in the future. Similarly, innovation is often discussed as a crucial success factor in SCM, not only concerning sustainability practices (De Carvalho et al. 2020) but also regarding novel technology application (Hahn 2020).

For OT average, the horizontally highlighted colour gradient similarly visualises the relative OT sizes for each OT over time. This nuancing is used to gain an insight into the OT development from year to year and to detect year-specific anomalies. Overall, it is striking that the number of published articles has increased substantially in all OTs. While some OTs show distinct growth patterns (e.g. OT1 supply chain types), others appear to somewhat fluctuate (e.g. OT6 supply chain context). These short-term increases or decreases in the interest of the research community can have manifold reasons which would require further investigation. The general increase in the number of publications is in accordance with the growth in article quantity and an intensifying pressure to publish across academia mentioned in section 2.

4.4. Research attention

The research attention paid to specific topics is conveyed by the average number of citations for each OT. The fuzzy cluster (C26) is again excluded from this part of the bibliometric analysis. To analyze the research attention paid to specific topics, the number of citations for each OT is considered. Thus, the research fields receiving the most attention by SCC scholars during the last five years can be determined and managerial implications can be drawn. To investigate the research attention, the absolute number of citations, the average yearly number of citations and the average yearly number of citations per article are calculated.

The absolute number of citations in the SCC field over the past five and a half years (2015–2020) is 103,155. Without consideration of the time dimension, this equals an average of 9.77 cites per publication. When looking at the yearly average, it can be observed that articles in the SCC field have an average a yearly cite rate of 2.52. As this rate is influenced by the decisions made during the database search, for instance, the applied exclusion criteria, its meaning is limited. However, the yearly cite rate of 2.52 for SCC publications could be used for comparisons with other research fields for the same time period. Regarding the average number of citations per article for the individual full years 2015–2019, a steady decrease can be observed over the years. In 2015, publications achieved an average of 19.59 citations, however, in 2019, the average article has only 3.51 citations. This substantial drop in the number of citations per article might be the result of rising number of publications, better access to academic databases and the general practice to base new findings on a broad base of previous works in the field. To analyze the research attention within the field of SCC research in more detail, comparisons between OTs and the aggregation or relative performance over time are required.

When the development of citations over time is included in the calculations, the difference in the observed research attention is lower but a similar pattern is visible. Table 4 illustrates the relative average number of article citations based on the yearly average (left side) and OT average (right side). Again, only the full years (2015-2019) are included for better comparability and a colour-scaled heat map is used as a bibliometric tool (Linnenluecke, Marrone, and Singh 2019) to visualise the differences regarding the research attention. The visualisation confirms the findings from the previous paragraphs describing the research attention and mirrors the heterogeneity exhibited by Table 3.

Table 4. Relative number of article citations per OT and year: relative research attention highlighted in nuances (blue = above average, grey = below average) based on the yearly average (left side) and OT average (right side).

For yearly average, the vertically highlighted colour gradient visualises the relative OT research attention for each year with darker blue colour nuances marking OT article citations above the respective year’s average and grey colour nuances identifying OTs achieving less than the average article citations. For instance, the yearly OT article citation average is 3.61 in 2015. Thus, the article citation average of OT9 (3.65) is coloured in light blue as this matches the year’s average, whereas OT4 (5.97 average citations) is coloured a darker nuance of blue and OT6 (2.39 average citations) in grey. From the visualisation, it can be understood that the topics of supply chain types (OT1) and green and sustainable supply chain management (OT4) constantly receive the most attention within the SCC field. On the other hand, relatively little attention is given to themes such as network management (OT2), supply chain optimisation (OT 3), relationship management (OT 5) and supply chain development (OT8) over the last five years. These findings appear to again highlight the relevance of sustainability and novel supply chain types and approaches. Interestingly, sustainability is also a topical issue in public sector discussion and lawmaking across the globe. Managerial implications include the increasing relevance of supply chain modernisation, digitalisation, and accountability due to growing consumer and regulatory demands.

For OT average, the horizontally highlighted colour gradient similarly visualises the relative OT article citations for each OT over time. This nuancing is used to visualise the OT development from year to year and to detect year-specific anomalies. The colour-scaled heat map illustrates a relatively similar development among OTs, however, it is interesting to note that the years 2015 and 2017 exhibit the continuously highest OT averages regarding article citations while a sharp decline can be observed from 2018 onwards. It appears that the overall growth in publication numbers (research interest) is linked to a decrease in research attention. Thus, it seems likely that the recent increase in the publication quantity and overall interest in SCC has led to a blurring of the field and a subsequent dispersion of scholarly attention. Potentially, this development could continue over the next years and make it more difficult for scholars and managers alike to navigate this complex research field.

4.5. Summary of the bibliometric analysis

Figures 5 and 6 provide an overview of the main metrics that are considered relevant for the analysis of the research output regarding the nine OTs. The growth rates for the number of articles, the number of citations and the average cites per article are based on the literature sample derived from the Scopus database search and calculated by comparing the figures for the first and the last full year of the observed time period (2015 and 2019). Concerning the development in the number of articles that can be allocated to the OTs, OT4 green and sustainable supply chain management and OT7 research exhibit the largest growth from 2015 to 2019. This is hardly surprising considering the omnipresence of this theme both in the public domain, for example, the ‘Fridays for Future’ movement starting in 2018, but also in national and international politics, for instance, the international climate conference in Paris in 2015. In contrast to the development in the number of articles (research interest), both the number of citations and the average cites per article (research attention) have decreased over the five-year period within all OTs. For the number of citations, OT2 network management and OT8 supply chain development show the steepest decline. The same OTs account for the largest drop regarding the average cites per article. This appears to show a shift in research attention of the scientific community towards digital as well as green and sustainable SCM. The least decline in research attention can be observed for the OT6 supply chain context which at the same time shows an average growth in the number of articles. This suggests that this OT is more stable than the others.

Figure 5. Overview of main metrics per overarching theme (OT).

Figure 6. Overview of growth rates of main metrics per overarching theme (OT).

The bibliometric analysis focusing on influential journals, research interest and attention as well as OTs enables supply chain practitioners and researchers to navigate this complex topic more easily and supports the identification of a suitable platform for their research and needs. The analysis shows an overall increase in research interest with some exceptions. In general, the whole field of SCC research appears to experience an upward trend concerning the number of research publications, which could indicate a high level of scholarly and supply chain community interest in the various topics of the field. The clusters (performance) management (C9) and performance integration/ integration effect (C46) are large but still growing. Since the research interest in these topics does not yet seem to be saturated, potential for high growth rates in the next years can be presumed and researchers might find it rewarding to work on these topics and to serve the interest. Examples of thematic clusters with comparably stagnating or erratic growth rates include decision (making/ support/ model) (C4), service (C11) and networks (C27). An exception exhibiting a downward trend is inventory (C14). The reason for this development is difficult to determine. Potentially, the increasing interest in forecasting and data related issues incentivised researchers to adjust their focus.

Regarding the research attention, a more heterogeneous picture presents itself when comparing the clusters within the individual years. A comparison of the average number of citations per article per cluster over time reveals an overwhelmingly negative trend. This appears to be due to the link between the overall growth in publication numbers (research interest) and the decrease in research attention. Business (C21) and empirical (C35) are the only notable exceptions from this overall development, which is partly explained by their general smaller cluster size. Overall, it appears to have become easier (or more necessary) for researchers to publish more frequently. At the same time, it seems to be increasingly difficult to receive the attention of the respective research field. SCC research could thus potentially grow to be more competitive and challenging, especially for early career researchers.

The analysis of the OTs supports an understanding of general relations in bibliometric data. Digital, as well as, green and sustainable SCM are currently the most promising thematic areas within SCC research. Apparently, a correlation exists between the observed strongly increasing publication numbers and at the same time substantially decreasing average citation numbers per article. Researchers should be encouraged to explore niches, such as collaboration for small and medium-sized enterprises, in accordance with and related to their respective areas of expertise. The observed high diversity and size of SCC could potentially lead to a more fragmented literature base and greater competition.

5. Synthesis and discussion

5.1. Overview of the synthesis and discussion

To derive a sample literature pool representing the overall SCC research field, the five journal articles with the highest yearly citation rate since publication are selected from each cluster. Assuming that the most frequently cited articles of each cluster are highly influential, the sample publications are likely to adequately represent the different topics. Furthermore, the synthesis is structured according to the nine OTs identified from the thematic clusters that generalise the topics discussed in the literature. The synthesis thus aims at the exemplary demonstration of core topics within the clusters based and the identification of themes and insights across clusters. As the fuzzy cluster (C26) is excluded from the synthesis process, 45 clusters are considered and a representative selection of 225 articles is derived from a total quantity of more than 10,000 articles. The analysis and synthesis of the sample articles show that most of the publications fit well with the cluster they are assigned to. However, due to the functional nature of the algorithm, there are individual article allocations that are not optimally assigned due to the functional principle of the clustering algorithm which identifies words that are shared by several titles of articles within the literature pool.

Figure 7 illustrates the distribution of the methodological approaches of the 225 sample publications. Clearly, conceptual papers constitute the largest part, followed by empirical quantitative approaches. The remaining categories of empirical qualitative, mixed-methods and literature review only amount to a combined approximate quarter of the total sample. This indicates a preference of conceptual and numerical approaches, while qualitative methods are less popular. While most papers contain a literature review section, only a small proportion of the sample publications are standalone literature reviews. Potentially, qualitative approaches and systematic literature reviews could contribute to the research field.

Figure 7. Distribution of methodological approaches of the sample publications.

5.2. OT1 – supply chain types

The OT1 supply chain types comprises clusters that are related to closed-loop supply chains, food supply chains and supply channel types. Therefore, the articles within this topic area are derived from three clusters, namely closed-loop (C1), food (C10) and channel (C15). Within these clusters, the publications with the highest yearly citation rates cover aspects such as contract, cost and pricing related coordination (e.g. Heydari, Govindan, and Jafari 2017; Zhalechian et al. 2016) as well as differences concerning dual-channel and omni-channel approaches (e.g. Feng et al. 2017b; Murfield et al. 2017). Several publications discuss the potential of environmental collaboration (e.g. Centobelli, Cerchione, and Esposito 2017; Grekova et al. 2016) but also the effects of corporate social responsibility and channel coordination (Panda, Modak, and Cárdenas-Barrón 2017). A related topic is a sustainability regarding food waste. Here, authors consider causes, uses and effects of food waste in addition to methods to reduce it (Canali et al. 2017; Göbel et al. 2015) as well as specific methods such as cradle-to-farm gate approaches and cradle-to-grave assessment (Tasca, Nessi, and Rigamonti 2017) or regional and local food hubs (Berti and Mulligan 2016).

5.3. OT2 – network management

The second OT covers various aspects of network management such as network design, SCM approaches or models, logistics and system management. This theme is composed of eight thematic clusters: network (C2), design (C3), approach (C5), model (C6), logistics (reverse/network/third party) (C12), system management/approach (C23), networks (C27) and chains (C31). Again, sustainability and green SCM are central research issues within the broader theme of network management. This comprises sustainable design (Arampantzi and Minis 2017; Tsao et al. 2018) and influencing factors and barriers (e.g. Mangla, Govindan, and Luthra 2017; Rabbani et al. 2020). Hashemi, Karimi, and Tavana (2015) and Jakhar (2015) propose approaches towards green supplier selection. Risk management in supply network is also important as Kwak, Seo, and Mason (2018) investigate potential relationships between supply chain innovation and risk management capabilities and others discuss risk management and propagation (e.g. Garvey, Carnovale, and Yeniyurt 2015; Ojha et al. 2018; Qazi et al. 2018). Here, the authors review supply chain integration and network dominance (e.g. Stevens and Johnson 2016; Ziaee Bigdeli et al. 2018) and develop network design approaches (e.g. Fard and Hajaghaei-Keshteli 2018; Keyvanshokooh, Ryan, and Kabir 2016). Within network management, partners for logistics services and reverse logistics are central (e.g. Bing et al. 2015; Bouzon, Govindan, and Rodriguez 2018; Govindan and Chaudhuri 2016). On the one hand, the notion of systems in this OT concerns systems thinking as, for instance, systems theory and system dynamics are used to solve complex problems (e.g. Fera et al. 2017; Kochan et al. 2018). On the other hand, some authors refer to distribution or information systems (e.g. Dweekat, Hwang, and Park 2017; Mogale et al. 2017). Many authors suggest the use of Artificial Intelligence tools such as neural networks for network management purposes (e.g. Huang et al. 2015; Jang and Lee 2017).

5.4. OT3 – supply chain optimisation

Supply chain optimisation is the central topic of the third OT which consists of nine clusters (C4, C7, C13, C14, C19, C37, C9, C44, C46). This theme includes publications related to decision making, optimisation, performance management, demand and price management, inventory management, risk management, supplier selection, performance management and supply chain integration. Regarding decision making, sustainability aspects (e.g. Mirkouei et al. 2017; Osiro, Lima-Junior, and Carpinetti 2018), optimal decisions and agile hierarchical decision making structures (e.g. Wu et al. 2017; Xiao and Qi 2016) are relevant. As in the previous OTs, sustainability is itself an important subtopic concerning optimisation (e.g. Dubey et al. 2017b; Esfahbodi, Zhang, and Watson 2016; Hong, Zhang, and Ding 2018; Luthra, Garg, and Haleem 2015; Miranda-Ackerman, Azzaro-Pantel, and Aguilar-Lasserre 2017). Collaborative efforts are also made for optimised scheduling and inventory issues (e.g. Cannella et al. 2015; Ivanov et al. 2016; Mateen and Chatterjee 2015; Soysal et al. 2018) as well as uncertainty and disruption management (e.g. DuHadway, Carnovale, and Hazen 2019; Pasandideh, Niaki, and Asadi 2015). Some authors analyze the relationship between supply chain agility and performance (e.g. Gligor, Esmark, and Holcomb 2015; Jermsittiparsert et al. 2019). Overall, the OT discusses the benefits of coordination and collaboration, for instance concerning service level maximisation (Sarrafha et al. 2015) or the ripple or bullwhip effect (Dolgui, Ivanov, and Rozhkov 2020). Similarly, authors aim at the optimisation of supplier selection (e.g. Kannan 2018; Rajesh and Ravi 2015; Wan, Xu, and Dong 2017).

5.5. OT4 – green and sustainable supply chain management

The OT4 green and sustainable supply chain management comprises six clusters with the subtopics green SCM, sustainability, corporate social responsibility and carbon emissions (C8, C18, C20, C29, C17, C32). While holistic modelling of drivers for green innovation is discussed (e.g. El-Kassar and Singh 2019; Song, Fisher, and Kwoh 2019), most authors focus on specific aspects of sustainable development such as organisational learning or quantitative modelling (e.g. Brandenburg and Rebs 2015; Suryanto, Haseeb, and Hartani 2018). Supplier sustainability compliance as well as selection criteria are a central component of this OT and highlight the relevance of transparency throughout the supply chain (e.g. Grimm, Hofstetter, and Sarkis 2016; Luthra et al. 2017). Similarly, sustainability implementation in developing and emerging economies is analyzed (e.g. Mani et al. 2016; Silvestre 2015). To achieve these research goals, sustainability assessment and criteria are discussed and compared (e.g. Ahmadi, Kusi-Sarpong, and Rezaei 2017; Ivanova et al. 2016). Apart from environmental sustainability, the social dimension of sustainability and corporate social responsibility are emphasised (e.g. Ağan et al. 2016; Croom et al. 2018; Mani, Gunasekaran, and Delgado 2018). Lastly, carbon emissions and CO₂ are relevant issues within the OT (e.g. Kagawa et al. 2015; Wang, Wang, and Huang 2017; Yuyin and Jinxi 2018; Zhang et al. 2019). It is interesting to note that publications on this topic are sorted separately depending on the use of the full term or chemical formula of carbon dioxide.

5.6. OT5 – relationship management

Five clusters on logistics service providers, relationships roles and game approaches are summarised in the OT5 relationship management: C11, C16, C25, C38, C42. Again, a relevant proportion of articles covers sustainability-related aspects, for example regarding logistics service providers (e.g. Abbasi and Nilsson 2016; Centobelli, Cerchione, and Esposito 2017), theoretical or game-theoretic approaches (e.g. Madani and Rasti-Barzoki 2017; Sinayi and Rasti-Barzoki 2018), or the application of contracts (e.g. Bai, Chen, and Xu 2017; Song and Gao 2018; Zhang et al. 2015). In addition, relationship management is researched in different contexts, for instance, the impact of countries’ institutional pressure (e.g. Sancha, Longoni, and Giménez 2015), small and medium-sized enterprises (e.g. Centobelli, Cerchione, and Singh 2019), blockchain incorporation (e.g. Treiblmaier 2018), or supplier management (e.g. Rezaei, Wang, and Tavasszy 2015). Advantages of collaborative relationships reported in the literature include supply chain uncertainty and resilience (e.g. Flynn, Koufteros, and Lu 2016; Scholten and Schilder 2015) and performance (e.g. Narayanan, Narasimhan, and Schoenherr 2015; Ploenhad et al. 2019).

5.7. OT6 – supply chain context

The OT6 supply chain context deals with aspects such as business management, production or manufacturing management and industry management and contains the five clusters C21, C28, C39, C40 and C43. On the one hand, this concerns the impact of collaboration and digitalisation as well as Industry 4.0 (e.g. Ivanov, Dolgui, and Sokolov 2019; Jermsittiparsert and Wajeetongratana 2019; Pradabwong et al. 2017). On the other hand, specific strategies such as external knowledge sourcing or lean business are explored (e.g. Brunswicker and Vanhaverbeke 2015; Dey et al. 2019). The different contexts presented in this OT range from cooperation in the hospitality and travel industry (e.g. Ye, Zhang, and Li 2018), over the food industry (e.g. Govindan 2018; Sharma, Chandna, and Bhardwaj 2017) and construction projects (e.g. Li et al. 2016), to small and medium-sized enterprises (e.g. Alsaad, Mohamad, and Ismail 2017) and smart manufacturing (e.g. Fynes et al. 2015; Tuptuk and Hailes 2018).

5.8. OT7 – research

OT7 Research is a comparatively small OT comprising three clusters on literature reviews and empirical research: C22, C35, C36. Some publications present the findings of comprehensive literature reviews (e.g. Chen et al. 2017; Govindan, Fattahi, and Keyvanshokooh 2017; Linnenluecke 2017). Overall, the number of distinct literature review papers is relatively small, as already illustrated in Figure 7. Other publications emphasise empirical research methods such as case studies (e.g. Dubey et al. 2017a; Huq, Chowdhury, and Klassen 2016; Queiroz and Wamba 2019; Varsei and Polyakovskiy 2017).

5.9. OT8 – supply chain development

Supply chain development is again a smaller OT containing three clusters (C24, C30, C45) which cover topics such as supply chain factors, value chain management and innovation management. This OT aggregates articles discussing factors influencing supply chains (e.g. Aschemann-Witzel et al. 2017; Venkatesh, Rathi, and Patwa 2015; Yadav and Barve 2015). This also includes mega-trends such as shifting dynamics of global trade (e.g. Horner and Nadvi 2018), the use of social media information (e.g. Cui et al. 2018), innovation (e.g. Costantini et al. 2017; Zilberman, Lu, and Reardon 2019) and crowd logistics (e.g. Carbone, Rouquet, and Roussat 2017).

5.10. OT9 – digital supply chain

The last OT derived during this literature review is digital supply chain. The three clusters contained in this theme cover mainly data and information in a supply chain context: C33, C34, C41. Papers discussing the impact and application of big data constitute a relevant proportion (e.g. Chae 2015; Gunasekaran et al. 2017; Papadopoulos et al. 2017). Another novel technology used in this context is blockchain (e.g. Choi et al. 2019). Data gathered from supply networks is also employed for different modelling purposes (e.g. Chiu and Choi 2016; Hosseini, Ivanov, and Dolgui 2019). Finally, information processing and exchange (e.g. Srinivasan and Swink 2018; Vanpoucke, Vereecke, and Muylle 2017), as well as, information technology capabilities (e.g. Basheer et al. 2019; Liu et al. 2016) are discussed in this OT.

5.11. Summary of the synthesis and discussion

To conclude, this literature review provides a broad and holistic mapping of the SCC research field by consolidating and extending previous research. It contributes to SCC research not only by providing an exhaustive survey of the available literature but also by using a novel approach to enable a better understanding of this complex topic. The MLR method allows for a systematic analysis of a large sample of SCC research publications, thus identifying major research streams within the field and deriving a research agenda.

Overall, the main research streams and most dominant central issues across the OTs appear to relate to sustainability or green SCM and technology integration. There is also a large number of (mathematical) modelling approaches, which suggests that the ideal collaborative supply chain is both difficult to achieve and not yet practically implemented at a larger scale (i.e. Keyvanshokooh, Ryan, and Kabir (2016), Govindan, Jafarian, and Nourbakhsh (2015)). Extensive research into different kinds of supply chains, such as food supply chains, and an overall orientation towards practical applications are clearly visible from the frequency of case studies and real industry examples found during the review. Concerning relationship research, one can observe that in the SCC context this focuses more on suppliers and services providers and much less on customers. This appears to show an orientation towards the procurement side of SCM, which means the upstream supply chain partners. It is also interesting to note that coordination and collaboration, also more rarely cooperation, are used interchangeably in the literature.

This review informs research and practice, provides incentives for future research and promotes the further application of SCC by generating an understanding of the magnitude of and influential research streams within the related literature. Figure 8 thus depicts an Evolutionary Thematic Map (ETM) of the SCC field from 2015 to the next decade to inform a future research agenda. For this purpose, based on the total of 10,556 articles included in the MLR, individual Excel files containing the titles of the articles published within each year are loaded into the Orange workflow and the 10 most frequent key terms (except supply chain) are extracted using the word cloud. The results were then aggregated and an outlook for the next decade of SCC research is provided. When reading the key terms for individual years, the frequency of occurrence increases from the bottom to the top of the lists.

Figure 8. Evolutionary thematic map of SCC research.

The ETM illustrates a gradually shifting focus of SCC research towards sustainable and green practices as well as relationship management while the term logistics appears to become less influential. Management, performance, modelling, and analysis are central in all years under consideration. While previous and current SCC research focused on logistics performance and the design and analysis of dedicated models, the authors expect a growth in the number and relevance of publications covering green and sustainable SCM, risk or disruption management, relationship focused research and case study methods for the next decade of SCC research. Thus, the future research agenda visualised through the ETM points towards an increasingly transdisciplinary and inter-organisational perspective on SCC. As expressed by Jayaraman, Ross, and Agarwal (2008), closed-loop supply chains contribute to the sustainability efforts. Driven by Industry 4.0 and smart environments, industry and manufacturing, as well as information technology content and the application of innovative technologies are likely to expand. Furthermore, systemic approaches such as system dynamics could provide a holistic and long-term development of supply networks and relating sectors. A systems consideration might be of advantage for the analysis and advancement of more aspects regarding SCC, as interconnectedness, collaborative interdependencies and complexity are central characteristics of supply chains (Orozco-Romero, Arias-Portela, and Marmolejo- Saucedo 2019). Comprehensive approaches towards collaboration taking into account multiple categories, supply chain agents and levels could be relevant avenues as publications often focus on specific areas such as maritime (e.g. Seo, Dinwoodie, and Roe 2016) or last mile collaboration (e.g. Yao, Cheng, and Song 2019), or certain aspects such as financial (e.g. Wandfluh, Hofmann, and Schoensleben 2016) or transport (e.g. Venegas Vallejos, Matopoulos, and Greasley 2020) processes. As illustrated by Aktas, Bourlakis, and Zissis (2020) and Yao, Cheng, and Song (2019), the application of SCC to consumer-oriented aspects of supply networks, such as last mile delivery, instead of supplier-focused advances are promising research areas.

5.12. Theoretical and practical contribution

This paper provides valuable contributions to both theory and practice. Three main theoretical contributions include (1) providing a comprehensive overview of the heterogenous research field of SCC, (2) identifying future research avenues, and (3) applying and extending the MLR method to SCC research.

First, this paper consolidates and extends previous SCC research and visualises the gradually shifting focus of SCC research in an ETM. The ETM illustrates a tendency towards sustainable and green practices as well as relationship management, while management, performance, modelling, and analysis are found to be continuously relevant topics. The findings presented in this paper enable researchers to gain an overview of the comprehensive field of SCC research and to navigate this complex topic more easily.

Second, several research gaps and consequently potential areas of interest for future research could be identified. Despite the apparent importance of performance management, relatively few comprehensive studies covering performance gains through collaboration were found, as most focus on specific industries or types of organisations (e.g. Centobelli, Cerchione, and Singh 2019) or kinds of supply chains (e.g. Esfahbodi, Zhang, and Watson 2016). Surprisingly, the barriers hindering collaborative behaviour also do not seem to be of predominant importance within SCC research. Despite some articles discussing the application of new technologies for collaborative supply chain networks (e.g. Queiroz and Wamba 2019), this area is researched much less extensively than the authors assumed. It could be expected, that innovative technologies such as AI will enhance collaborative supply chain processes as mathematical approaches are widespread (e.g. Zhalechian et al. 2016). Lastly, there are only few research papers presenting a systems perspective on supply chain networks among the most highly cited publications and those mostly apply systems dynamics modelling (e.g. Kochan et al. 2018). In addition to systemic approaches, literature reviews only constitute 7% of the 225 sample publications and could provide relevant consolidation of and insights into this fragmented field.

Third, this paper applies and extends the MLR method to SCC research. The MLR method enables researchers to systematically analyze large samples of research publications in an accessible way, without the need for in-depth methodological knowledge in data-mining. Thus, researchers from SCC and related fields are enabled to apply the MLR method as a useful tool to consolidate and advance their own focus areas.

In addition, this literature review provides three practical contributions by (1) enabling practitioners from SCC and related disciplines to navigate this complex topic more easily, (2) providing incentives to apply theoretical findings into practice, and (3) suggesting managerial implications.

First, the SCC literature field is relatively broad, transdisciplinary and has grown substantially over the last few years. While SCC has been researched extensively over the last decades and become a driver for organisational success and goal attainment, the large number of publications dedicated to SCC has made it difficult for practitioners to navigate the topic to identify relevant inspirations and advice.

Second, this paper highlights an expected growth in the number and relevance of publications covering green and sustainable SCM, risk or disruption management, relationship focused research, case study methods, the application of innovative technologies, and systemic approaches. Thus, the identified OTs and research clusters and their in-depth analysis provide incentives for practitioners to apply recent theoretical findings into practice. The paper consequently serves as an accelerator for the realisation of green and sustainable SCM.

Third, managerial implications of this paper include the following. We recommend cooperating with researchers by providing access to organisational data or by becoming involved in studies. This enables a better understanding of SCC influence factors, barriers hindering collaborative behaviour, etc. Supply chain performance appears to be a continuously important aspect of SCC research and should thus remain a central organisational goal, albeit extended by sustainability performance. The application of new technologies and systemic approaches for collaborative supply chain networks have been identified as relevant future topics. We advise managers to dedicate resources to integrate innovative technologies into day-to-day supply chain operations as well as to consider their supply network from a systems perspective. We recommend taking up a strategic approach to collaboration to enhance long-term performance throughout the organisation’s supply chain.

6. Conclusion and limitations

The purpose of this paper is to gain an overview of the heterogenous research field of SCC. For this purpose, a MLR of the literature is performed to consolidate the last five of SCC research. The resulting thematic constructs are aggregated into nine OTs which are subsequently analyzed to determine potential research and practical gaps. The paper thus discusses how the field of SCC research has evolved over the last five years and derives OTs from the SCC literature. SCC presents itself as a heterogenous field that can be clustered into 46 thematic segments, most of which contain between 100 and 400 articles. On the level of the OTs, sustainability SCM and technology integration appear to be the most dominant central issues while an overall orientation towards practical applications and procurement can be observed.

In addition, developments regarding research interest and research attention within the clusters and OTs are examined and the journals publishing the most impactful articles concerning the different themes are identified. The research interest is expressed by the total number of published articles in a cluster per year, while the research attention paid to specific topics is conveyed by the average number of citations for each thematic cluster. The visualisation of the research interest shows that the number of published articles has increased in nearly all clusters and indicates a high relevance of collaborative SCM for performance-related aspects while an overall variable research interest in diverse aspects of SCC research could be deduced. While topics such as performance management and problem optimisation appear to be of central importance, outliers which arouse limited interest include SCC literature reviews and collaborative efforts within small and medium-sized enterprises. Concerning the OTs, green and sustainable SCM research exhibits the largest growth. The general increase in the number of publications is in accordance with the overall growth in article quantity and an intensifying pressure to publish across academia. However, a steady decrease in research attention can be observed across the clusters and OTs. While in 2015, publications achieved an average of 19.59 citations, in 2019 the average article has dropped to only 3.51 citation, potentially due to better access to academic databases and the general practice to base new findings on a broad base of previous works in the field. The development of citations over time again highlights the relevance of sustainability, novel technologies and data analysis, as well as supply chain relationships and transparency. Regarding the OTs, the steep decline in average cites per article within OT2 network management and OT8 supply chain development appears to show a shift in research attention of the scientific community towards digital as well as green and sustainable SCM. It seems likely that the recent increase in the publication quantity and overall interest in SCC has led to a blurring of the field and a subsequent dispersion of scholarly attention. Potentially, this development could continue over the next years and make it more difficult for scholars and managers alike to navigate this complex research field. The analysis reveals that the Journal of Cleaner Production and International Journal of Production Economics can be regarded as the most impactful journals in the field. The heterogeneity of the field is again highlighted as only 13 journals account for more than 100 articles each while 1,473 journals contribute less than 10 publications.

This paper has several practical and theoretical implications. First, the consolidation of the streams of SCC theory in an ETM illustrates a gradually shifting focus of SCC research towards sustainable and green practices as well as relationship management, while management, performance, modelling, and analysis are found to be central topics in all years under consideration. Regarding the next decade of SCC research, the authors expect a growth in the number and relevance of publications covering green and sustainable SCM, risk or disruption management, relationship focused research, case study methods, the application of innovative technologies and systemic approaches. Second, the paper enables researchers to gain an overview of the comprehensive field of SCC research and furthers our understanding of logistics and SCM as researchers and practitioners from SCC and related disciplines are enabled to navigate this complex topic more easily.

Furthermore, the paper proposes different avenues for future research. The review indicates that some topics, such as literature reviews and collaborative efforts within small and medium-sized enterprises constitute research niches in the field of SCC that could potentially over the next years. Further identified research gaps include studies covering performance gains through collaboration and barriers to collaborative behaviour. As literature reviews account for only 7% of the publications, further SCC reviews could provide relevant consolidation of and insights into this fragmented field. Third, practitioners can use the ETM to identify appropriate research streams and the related influential journals for their business context, needs and objectives. Supply chain managers might, therefore, find it easier to access topical information on relevant issues. Managerial implications of the observed research trends include the increasing significance of supply chain modernisation, digitalisation, and accountability due to growing consumer and regulatory demands.

This paper has its recognised limitations. First, the choice of a broad instead of an in-depth approach clearly restricts the thoroughness of the resulting analysis and discussion. Despite the large sample of journal articles, this review cannot claim to provide a full overview of SCC due to restrictions concerning the choice of keywords, publication year etc. while some extent of fuzziness needs to be acknowledged and, therefore, thematic clarity cannot be guaranteed. The analysis of the five articles with the highest yearly citation rate is not representative of the cluster but provides a glimpse of the most influential articles within the thematic area. Further focused and critical analysis of the OTs identified through the MLR could provide interesting insights.

Acknowledgements

This work was supported by the tax revenues on the basis of the budget adopted by the Saxon State Parliament under Grant SAB/100379142 and by a full scholarship towards Kevin Reuther by the Research and Development Management Association (RADMA).

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Data sharing is not applicable to this article as no new data were created or analyzed in this study.

Additional information

Funding

This work was supported by the tax revenues on the basis of the budget adopted by the Saxon State Parliament through the European Social Fund [grant number SAB/100379142]; Research and Development Management Association (RADMA).

Notes

1 $=\sqrt{{\displaystyle \sum _{i=1}^n(q_i-p_i{)}^2}}$