Supply Chain Integration: A Review and Bibliometric Analysis

Supply chain integration has been widely identified as a key research topic by both practitioners and academicians. In such environment, it is essential to vividly illustrate the publications contribution during the period of time and recognize research area and interests as well as the direction of research trend for future studies. With the availability of bibliometric data and variety of analytical tools for evaluation purposes, disregarding bibliometric analysis would be a missed opportunity for this area. Therefore, the current research attempts to deliver a comprehensive comparison thorough using rigorous bibliometric tools that provides a better understanding not previously fully grasped or evaluated by prior studies in the area of supply chain integration. The objective of this research is to recognize the global scientific production; the most productive authors, journals, articles and countries as well as to extract the most influential articles. The analysis begins by identifying over 500 published studies during the period of 1980 to 24th February 2016, which are then purified to works of proven influence and those authored by influential investigators. Web of Science Core Collection (formerly known as ISI), category of management was utilized to identify the relevant articles. Gaps are also discussed in knowledge about literature and bibliometrics analysis. The findings provide wisdom and a vigorous roadmap for further investigation in this field.


Introduction
Over the last decade, the supply chain integration (SCI) has become a vigorous topic of research in the context of operations management (OM). SCI is widely considered by both practitioners and researchers on investigating collaborative relationships between a manufacturing firm and its supply chain partners (Prajogo and Olhager 2012). As supply chains grew up, a variety of relevant issues have occurred. While customer expectation is continuously increasing, companies have forced to invest and focus their attention to coordination across and beyond the organizational boundaries. Customers expect wider choice, better service, fast delivery and higher quality as all these are critical factors which affect company's strategies and the bottom line of its operations (Li et al. 2012). Accordingly, the increasing global competition has driven firms to not only develop their own performance, but also concentrate on integrating their supply chain partners into the overall supply chain competitive advantage (Alfalla-Luque et al. 2013;Zhang et al. 2010).
The initial debate about supply chain integration began with horizontal alignment of operations across processes which discussed in the 1990s (Frohlich and Westbrook 2001). Stevens (1989) divides SCI with two distinct flows of material and information. Most concepts in such relationships are supplier integration, customer integration and internal integration (Zhao et al. 2014;Wong et al. 2011;Flynn et al. 2010;Koufteros et al. 2007;Das et al. 2006;Power 2005;Pagell 2004;Huang et al. 2002) and the intention was to cover the basic concepts usually used to measure SCI, in both operational and strategic purposes.
In developing viable organizational performance, Kim (2006) conceptualized the linkages between firm's supply chain integration strategy and its competitive strategy. Effective SCI links a manufacturer firm with upstream suppliers and downstream customers, as well as other channel members by integrating their relationships, activities, functions, processes and locations (Kim and Narasimhan 2002). Based on a review of prior studies, SCI was defined from three perspectives of material flow, information flow and important, companies are failing in their attempts at internal and external integration (Jayaram and Tan 2010). Many suppliers have invested in modern technologies, but cannot use their own technologies to build matchless supply chain capabilities (Fawcett et al. 2011). The resource-based view (RBV) theory stresses how strategically it is important for a firm to have the resources and competencies that are invaluable, rare, and which are difficult to imitate or replace (Barney 1991;Carr and Pearson 2002;Feng et al. 2010). In addition, the RBV theory framework points out not all resources confer enduring and longlasting value and it is important to be able to differentiate those that offer sustainable valuable resources (Das et al. 2006). Hence, there must be tangible and intangible integration of both internal and external resources to develop organizational capabilities.
Whereas studies focusing on more narrow conceptualizations of information flow, material flow and financial flow integration, there is a dearth of research investigating of other aspect of integration like infrastructure technology integration and knowledge integration in which lead organization to a sustainable performance. Practice of how to exchange knowledge may support firms change from a traditional relationship with supply chain partners to dyadic integration; therefore, suppliers and customers become as a tactical partners and more integration appeared into their business (Asgari et al. 2015).
The current research attempts to provide a comprehensive evaluation of SCI through using precise bibliometric tools; thus, we started with a pool of over 500 published articles and considered more related studies. The aim of this research is to provide a unique structure to better understanding of SCI related publications as well as to provide a systematic and objective mapping of different themes and concepts in the development of SCI field. In addition, this research attempts to recognize the contribution to the development of SCI by identifying research gap. Formerly, studies have provided insight into the field, but additional analysis of previous studies using rigorous bibliometric analysis which can provide further insights, not previously fully grasped or evaluated to demonstrate the value of scientific research.

Document type
Document type distribution of supply chain integration research literatures during the time span was illustrated in Fig. 1. There were five document types in WoS, namely article, proceedings paper, review, editorial material, and book review. These document types were found among the total 386 publications in which English was the dominant language. From these documents, 242 publications belong to publication type of article, and it was the most frequently used document type, comprising 62.70 percent of the total document, followed by proceedings paper (118; 30.06 percent), review (15; 3.90 percent), editorial material (7; 1.80 percent), and book review (4; 1.00 percent) as illustrated in Fig. 1. Journal articles produce the most useful information because of their timeline as well as the evaluation of the scholars.

Bibliometric analysis
Structural methodology and associated analytic types, which allow organization of the reviewed literature, can be obtained deductively or inductively. Methodology, in a deductive approach, is employed before the analyzing of material. Whereas inductive method methodology are developed from the material by means of generalization (Seuring and Müller 2008). With a given model of a bibliometric analysis research, an inductive approach is employed for the purpose of data analysis (Fahimnia et al. 2015). The bibliometric analysis and various visualization technologies instinctively presented the analysis results.
Bibliometric analysis using SPSS, Excel and Kutools provide additional data statistics including author, affiliation, and country distribution. Kutools can be used to analyze the frequency of occurrence of a text in different fields of the bibliographic data. In addition, Tableau was employed to demonstrate geographic distribution of publications and VOSviewer text mining and visualization functions was used to create a co-occurrence map of title field keyword.

Country/territory and institution
We investigated the geographical distribution of SCI research by analyzing the production of articles on per country.  Globalisation has increased the scope of opportunities and the geographical spreading shows that SCI research and practice has involved institutes and research centers from throughout the World. The distribution by country/territory helps future studies for further research collaboration and expanding relevant issues. The USA as a developed country and China as a developing country held the majority of total world production. The issue of SCI and its solution is complex and varies in different categories, environment and culture (Flynn et al. 2016). As it has been shown in Fig. 3, there is a lack of empirical research in African and South American countries. It seems quite clear that research on SCI should be expanded in different environment and culture. Therefore, future research should encompass research with extending to a broader population of firms, including other countries, for generalizability of the results and to detect potential country effects.
According to the top 10 subject categories, four research perspectives on supply chain integration were as follows: management science, business, and engineering.  (Garfield 2006), has become perhaps the most popular bibliometric product used in bibliometrics itself but also outside the scientific community (Glänzel and Moed 2002).
In addition, the value of the impact factor is affected by different factors such as subject area, type of documents or length of the citation measurement window (Bordons and Barrigón 1992). The journal impact factor, originally developed as a tool to select journals for indexing in the Science Citation Index published by the Institute for Scientific Information (ISI), has nowadays become the bibliometric construct most widely used for evaluation in the scholarly and publishing community (Moed et al. 2012).
The Impact Factor introduced by Eugene Garfield and regularly published in the annual updates of the Journal Citation Reports (JCR) ( Glänzel 2002) is the general bibliometric indicator for assessing the citation pattern and article quality. The journal impact factor once employed to show the quality of an article will upgrade the bad ones and downgrade the good ones (Gisvold 1999). In other words, impact factor may have improperly used to evaluate the quality and value of articles. Impact factor depends on the research field, since citation habits and citation dynamics can be altered in every research fields (Seglen 1997 Table 6 outlines the top ten contributing authors and the quantity of papers they authored or co-authored and ranked by the total number of publications. As can be seen in this data, Baofeng Huo and Xiande  Note: TP, the total publications; TP R, the rank; TP (%), the share in publications; TC, the total citation (times cited); ACR, the average citation rate; HC, the highly cited paper; FA, the first author

The most influential articles
Follow-up future studies are needed to analyses citation patterns in details and identify core articles in the research domain. While most studies, in OM discipline, picked up the highly cited/total citation (TC) articles as the most important articles (e.g. Asgari et al. 2016;Cobo et al. 2015) a few researcher in other discipline used average citation rate (ACR) as an extra indicator to evaluate the most influential articles (e.g. Maghami et al. 2015). Previous studies, traditionally, extracted and ranked articles based on the number of times citation or ranked them based on journal's impact factor. Although total citation may allow us to distinguish productive articles, it is not enough to enjoy a substantial evaluation. However, there is not a clear cut-off criterion to recognize influential articles among a set of objects. Therefore, we are going to conduct a systematic way to achieve the goal of identifying the most influential research.
Whether a study is authoritative, it can be judged from two aspects; the first aspect is the number of times citation (TC) and the next one is the average of citation rate (ACR). The former aspect refers to the studies that appeared most frequently in the references of SCI. Building upon these factors, a multidimensional scaling analysis was employed to provide a visual depiction based on importance-performance analysis (IPA) matrix to identify most influential articles and overcome the weaknesses of previous studies. TC is essential to distinguish articles with the high performance in citation. In addition to TC, ACR allows us to recognize the most important articles. On the other word, this technique enables us to figure out the most effective articles which published so far and helps to examine the origin and direction of future research trends (See Table 5). Martilla and James (1977) introduce the IPA matrix into the marketing domain to help target audiences identify and rate certain product or service attributes, based on their importance to the rater and their impact on the organisation's overall performance. This analysis has become popular among researchers to enlarge the priority of finding analyses for managerial usage and academic purpose (see Matzler et al. performance is depicted along the x-axis and importance along the y-axes. IPA extends the outcome by taking the importance and performance of each study into account. In this two-dimensional matrix, sources of data fall into one of four quadrants; Quadrant I (high TC, high ACR), Quadrant II (low TC, high ACR), Quadrant III (high TC, low ACR) and Quadrant IV (low TC, low ACR). Quadrant I, reflecting high level of both total citation and average citation rate and represent areas which contribute to recognize the most influential articles.
To construct the importance-performance matrix, the mean of ACR and TC were calculated. To calculate the accurate cut-off criterion, we removed documents with no citation and they do not accounted for calculating the cut-off rate. The number of 159 documents was ignored due to zero citation. We calculate the value cut-off criterion of ACR and TC by considering 229 documents from 386. The total of times citation is 6061and the total of average citation rate is 799.88. We propose cut-off values of 3.52 (799.88 divided by 229) for ACR and 26.7 (6061divided by 229) for TC. Therefore, any cases greater than these cut-off criterions is located in Quadrant I and considered as influential articles. Fig. 4 shows the scatter plot of data to visualize highly cited articles with highest citation rate. This Fig. represents the total citations (performance) in vertical axis and average citation rate (importance) in horizontal axis corresponding yearly average. Spots are displayed by case numbers. By a close examination of the two rankings, we found out 46 articles out of the total number of 386, which have the highest total citation and average citation rate and representing roughly 11.92% of total publications. The sequence of most importance with the highest performance articles is provided in   The most influential articles in "SCI" by taking quadrant II into account

Authorship pattern
Collaboration is found as an essential effort to publish in the area of SCI. Individual efforts account to only 15.8 percent of the total, whereas 84.2 percent of publications are produced in collaboration. Fig. 6 shows that collaboration pattern of co-authorship has categorized into 5 groups. However, maximum output is observed in the team strength of two authors (34.46%), followed by group of three (28.24%), four (17.10%), single author (15.80%) and five authors (4.40%) respectively. Inzelt et al. (2009) believe that scientific collaboration and co-authorship has significant influence on citation rate of publications.
The most evident form of scientific collaboration is co-authorship, which is a frequent and reliable target of scientometric studies on collaboration (Glänzel and Thijs 2004). According to Raan (1998), international collaboration among researchers would increase the citation rate through spurious causes, such as self-citations. However, it is essential to use necessary references in scientific publications which are made for professional reasons principally (Vinkler 2010).

Cooperation network
The cooperation network of authors is shown in Fig. 8, the size of the rectangles and labels represent the authorship weight. The cooperation network of authors was analyzed by VOSviewer, authors that have published ≥2 papers were extracted from the database. The co-authorship relations are relations representing whether an author have written a paper with another author. Typically, a paper is written by two or more authors. Investigating co-authorship information on a larger database of scientific publications will support in recognizing a set of researchers who work closely together (Van Eck and Waltman 2011). In the co-authorship network, each author corresponds to a rectangle and each link between two rectangles of different authors indicates that there is a co-authorship. The authors, Xiande Zhao and Baofeng Huo, have most collaboration with each other and other authors. This means, not only Xiande Zhao and Baofeng Huo produced the greatest number of papers, but also has the most collaboration with other researchers. Thus, these authors can be considered as the most influential author in the area of SCI.

Co-authorship countries
Authorship collaboration is the main pattern in scientific collaboration, one thing worth noting is that visualizing the scientific collaboration at the country level exposes that country collaboration is significant in order to provide detailed information in international collaboration levels. The indicators of collaboration are based on authors' address. Fig. 9 represents the network of international co-authorship among countries according to the information about research competency and co-authorship relations.
Bibliometric map of co-authorship of countries was analyzed by VOSviewer, countries that have published ≥2 papers were extracted from the database; hence, 18 countries were taken into account. In this view, countries are indicated by a label and by a rectangle. The top productive countries have larger label and rectangle. This figure shows the major players in the network in terms of international coauthorship relations. The size of each rectangle shows the number of papers written by authors from the country. Moreover, co-authorship relation among countries corresponds to a link and the width of links clarifies the number of co-authorships between countries. In the field of SCI, the US and China have most collaboration with each other as well as in the World. It can be concluded that the top productive countries carried out most international collaboration with others in the SCI research.

Conclusion
This article provided a quantitative perspective of bibliometric and citation analysis of journals citing and research trends of Supply chain integration (SCI) area in which brings contributions for both academics and practitioners. SCI is the main building block of every supply chain due to requiring coordinating and cooperation and managing of physical flow, information flow and financial flow between parties. Indeed, SCI is an interesting research topic, is of managerial relevance as well as academic importance, to achieve competitive drivers such as; "increased cost competitiveness, shorter product life cycles, faster product development cycles, globalisation and customisation of product offerings, and higher overall quality" (Hale et al. 2008). Likewise, SCI is a set of procedures that effectively integrates various business parties, including suppliers, manufacturers, warehouses, and stores, so that the company's products enjoy proper distribution in the required quantities, to the right places at the right time and done at minimum cost and at satisfactory service levels.
We argue that the principal aim of the development of bibliometric method and indicators in the field under investigation on the nature of scholarly quality; number of publications and citation trends, the most productive articles, authors, journals and countries as well as identifying the most influential articles. However, a few review articles on supply chain management area and related fields have been published, thorough bibliometric analysis. Therefore, the current study shows an evolution of the influential articles and contributes to the SCI field by further mapping the relationships amongst the higher impact works.
The 386  Furthermore, the SCI in USA has attracted an increasing level of attention in recent years and was second to China in the number of total publications. The findings also show that industrialized countries have had more publications contribution to the development of SCI. Likewise, little attention has been paid to the research in African and South American countries; thus, there is a lack of empirical research in these scopes of study in the area of SCI.
The current research will be very helpful to identify the latest research trends in the SCI field to offer guidance to future studies in this field. The novel aspect of this research is that the indicators have not been studied in this context before. While previous studies evaluated the quality of articles based on the number of citations, the current research utilized a complementary factor to identify the most influential articles in the research field. To the best of our knowledge, this is the first study to include total citation and average citation rate simultaneously and utilized importance-performance analysis (IPA) to prioritize the most influential articles. The 63 articles were categorized in two quadrants for developing an effective literature review, identifying the derivation of research gaps, helping future researchers and practitioners to detect the broad range of business activities. It is important to understand how the IPA contributes to the formulation of bibliometrics analysis, whereby researchers and practitioners are required to choose priorities from a number of studies.
The analysis which is conducted in this research is not without the limitation, and these pose challenges for the future development of this study. Firstly, the citation count is limited to the Web of Science Core Collection documents and citations from other documents are not included. However, one of the most popular data source for bibliometrics research is the ISI (Web of Science) Citation Databases (i.e., the Science Citation Index Expanded, the Social Science Citation Index and the Arts and Humanities Index) (Bar-Ilan 2008). Nonetheless, future research may need to concentrate on both database; Web of Science and Scopus. Secondly, the concentration of this research was extremely on journals of an international scope which mainly published in English; thus, we ignored local journals and publications. Thirdly, we have dealt entirely with scientific articles, ignoring other channels of scholarly revelation, such as working documents. Nevertheless, in a given the nature of the tendencies identified in our analysis, we do not believe that these are seriously conditioned by the biases these limitations involve. Lastly, for future studies comparing judgements of experts and bibliometric indicators, datasets are necessary which cover a broad range of different disciplines (Bornmann and Marx 2015).
Furthermore, supply chain members should recognize different types of relevant issues covered within the supply chain before employing the strategy of SCI. For instance, in a recent study, Flynn et al. (2016) show that different types of uncertainty require different levels of SCI. They discuss that without an appropriate consideration managers may invest in SCI that is not aligned with the type of uncertainty their firm faces. Despite the awareness that integration is important, companies are failing in their attempts at internal and external integration (Jayaram et al. 2010). Companies have bought advanced technologies, but not used them to build non-imitable supply chain competencies that deliver unique customer value (Fawcett et al. 2011). Managers should not assume that all investments are equally effective and the same level of investment will assure the same result. Therefore, a better understanding of the concept of integrating supply chain, its dimensions and its implications, is of managerial relevance as well as academic importance may enhance the overall supply chain competitive advantage. In order to do this, firms may need to develop a rating system or rank strategic areas by their perceived level of importance based on the management of the environmental supply chain.