2.1 Big Data: A Revolution Arrived

In the current century, ‘Big data’ has emerged as one of the most used buzzwords in managerial literature. A surge has occurred in the number of published researches about big data within the management and marketing domain (Wamba & Mishra, 2017). As it is possible to imagine, most of these researches target to improve the existing knowledge about how big data can be turned into money and improve an organizations’ functioning. As a result, the present chapter explores and further comments on the findings presented by the extensive literature research proposed by Rialti, Marzi, Ciappei, and Busso (2019).

From this perspective, it has commonly been observed how big data and Big Data Analytics (BDA) has dramatically affected the ways of conducting business (McAfee & Brynjolfsson, 2012). Managers of today have large amounts of information at their disposal and, therefore, are more informed about the status of internal operations, processes involving supply chain, the performance of the workforce, and the behavioral patterns of the consumer (Hofacker, Malthouse, & Sultan, 2016). In fact, in the digitalized economy, it has been observed how any organization could potentially collect large amounts of data that was unthinkable just a few years ago. Along with the ability to collect data, decision making processes by the management has evolved, and the managers are now capable of making the decision about implementing the best-suited strategy based on the newly acquired information (Mardi, Arief, Furinto, & Kumaradjaja, 2018).

Due to the emerging potential of big data, the need for systems to decodify these extremely large datasets became evident (Labrinidis & Jagadish, 2012). Indeed, it has become apparent that these systems have the capability to transform unstructured datasets into information in the form of insights that can impact management decisions and organizational performances. Big data and BDA capable systems are also related to improved organization performances in terms of agility, flexibility, dynamism, and ambidexterity. This is particularly important when considering that such systems may be adaptable to different kinds of data and, hence, may provide a continuous flow of information to ensure the continuation of a process, even amidst difficulties (Xuemie, 2017). It has also become apparent that these systems could assist organizations in identifying and exploiting new opportunities. Here, the importance of dynamism emerges, and the organization becomes capable of scanning the environment continuously to obtain a competitive edge. As a result of correctly utilizing big data, they can generate value (Rialti, Marzi, Silic, & Ciappei, 2018).

Since the availability of BDA capable Information Systems (IS) and procedures have been deemed to be linked with organizational agility, flexibility, and dynamism, therefore, scholars have recently begun to use literature pertaining to dynamic capabilities as a theoretical lens through which they interpret the effects of big data on organizations. BDA capable IS can be used in different situations; and during any environmental turbulent situation, it can even give organizations an advantage over the competition (Lee, Kao, & Yang, 2014). The use of BDA capable information system is normally associated with usual processes and routines that may be useful in solving different problems that are data-related.

Several other theoretical approaches have also been used by scholars to underpin the topic, for example, existing research has explored big data focusing on the information value theory, resource-based view, technology acceptance model, or the dynamic capabilities. The literature on the topic is still fragmentary. Particularly, the literature on the topic of big data and the impact of big data on management and dynamic capabilities need further systematization (Xuemie, 2017). It is, therefore, clear that there is a need to address this specific literature gap; which is demonstrative of the requirement of recording and organizing the existing literature (Ardito, Scuotto, Del Giudice, & Petruzzelli, 2018). For this, we need to perform a bibliometric analysis and literature review.

This chapter is organized as follows. The following section focuses on the importance of big data and BDA capable systems for the organizations, and how the dynamic capabilities contribute to an organization in this stream of literature. The next section describes the methodological procedure. In the fifth section, bibliometric results are discussed which are followed by the literature review in the next section. Finally, the suggestions for future research are presented by the author to provide a guide to the reader.

2.2 What Is Big Data? Towards a Consensual Definition

According to McAfee and Brynjolfsson’s seminal research (2012, p. 5), “smart leaders across industries will see using big data for what it is: a management revolution”. Today, the enormity of the impact that the big data has in the world of management is clear to all researchers who are associated with big data. For any organization, information is recognized as one of the essential value-creating factors. Because of the characteristics of big data, the extent to which value can be created by using the information has simply reached to an unmatched level.

Big data is different from conventional large datasets in terms of seven parameters acknowledged as the ‘Seven Vs of Big Data’: Volume, Velocity, Variety, Veracity, Value, Variability, and Visualization (Mishra, Luo, Jiang, Papadopoulos, & Dubey, 2017). The seven parameters of big data are defined below.

  1. 1.

    Volume: This characteristic refers to the sheer dimensions of this typology of datasets. Indeed, big data’s dimensions frequently exceed terabyte (Mishra et al., 2017). Such a phenomenon is a direct consequence of the fact that worldwide computers, mobile devices, and machines generate data that exceeds 2.5 exabytes per day. Coherently, companies such as Walmart or Amazon have more data from consumer transactions than ever-imagined just 10 years ago.

  2. 2.

    Velocity: Velocity parameter in terms of big data is the “rate at which data is generated and the speed at which it should be analyzed and acted upon” (Gandomi & Haider, 2015, p. 138). In the current digital era, businesses such as social networks are collecting data at every instant of their operations. Facebook, for example, collects about 136,000 photos, 510,000 comments, and 290,000 updates every minute; which is a large amount of data to be managed in real-time.

  3. 3.

    Variety: This is a trait associated with the “heterogeneous sources of big data” (i.e. sensors embedded in machines, consumers’ activities on social media, B2C or B2B digital interactions, etc.) and the consequent assorted formats that the files composing big data may assume” (Rialti et al., 2018, p. 1098). Considering the Facebook example again, all this collected data may be in different formats like .jpg format, .doc format .mp3, and .mp4 format, etc. The problem is then related to merging all the formats in order to create files that may be analyzed using a single analytical tool.

  4. 4.

    Veracity: Veracity is a characteristic associated with the required degree of trustworthiness that must be possessed by the sources of big data (Mishra et al., 2017). As it is easy to imagine, unreliable data generate unreliable results and unreliable decisions. Thus, it is necessary to consider only data whose origin could be traced and assessed periodically.

  5. 5.

    Value: It is a characteristic related to the ultimate economic value of big data. This is the value that might be generated after any organization uses processes and technologies to analyze the collected big data (Xuemie, 2017). Obviously, in the massive amount of data, there is always the need to identify the ones that could be used to generate strategic decision.

  6. 6.

    Variability: This trait suggests that there might be variations in flow rate of data, processing, and data sources (Wamba & Mishra, 2017). As the data depends on the tool used to generate it, it may be possible to comprehend how the quantity may vary with time.

  7. 7.

    Visualization: It concerns with the possibility for data analysts to get visual insights as an output of big data analysis (Liao et al., 2018). This characteristic, albeit often neglected by scholars, is fundamental as there is no use of a large dataset if it cannot be transformed into sharable information.

Moving ahead from these premises, big data could then be defined as any huge dataset of unstructured data, that could not be analyzed using traditional database tools and methods, and that originate from reliable sources and intrinsically contain information that could be transformed to give strategic insights that help in making decisions.

Considering the characteristics of big data, organizations face a difficult challenge while trying to extract any insight from these datasets. Big data is identified as large and complex datasets because of which traditional database management software cannot be used for processing big data (Manyika et al., 2011). For collecting, storing, and analyzing big data, tools are required that are based on artificial intelligence paradigms such as data lakes, NoSQL data models, and schema-less data storage-retrieval. It is, therefore, necessary for organizations to define proper BDA processes for the defined stages. These stages are data acquisition, cleansing, integration, modeling, and interpretation. According to Prescott (2014), to gain any competitive advantage and economic value form collected information, it is important that the organizations develop systems and procedures that can not only collect data but also remove any unworthy components (messages without any useful content or spam messages). The implementation of the identified processes has its own challenges. Frequently, to accept BDA capable systems or processes, the whole organization’s culture needs to be changed (Rialti et al., 2018). The resistance to change may come from managers and employees. Because of the fear of change, they may oppose the implementation of automatic processes that are able to complement human intervention in the process of decision making. It is observed that when managers’ and employees’ mindsets become familiarized with big data and BDA processes, the whole organization could be characterized by a culture assisting the use of big data in the business.

Once these systems are in place, overcoming all potential difficulties, the changes are likely to give positive outcomes (Akter, Wamba, Gunasekaran, Dubey, & Childe, 2016). More accurate decisions may be made by managers based on the insights they gather from BDA capable IS. Specifically, the implementation of BDA provides the following benefits.

  1. a.

    Information from big data offers managers the possibility of knowing their consumers better than ever. It is possible to predict individual consumer’s behavior and propose tailored offerings in terms of prices (Erevelles, Fukawa, & Swayne, 2016).

  2. b.

    Big data can dramatically improve organizations’ internal operational efficiency. On the one hand, BDA may prove extremely useful for controlling business processes (Del Giudice, 2016). Indeed, BDA capable IS for business process management (or BDA capable BPMS) may allow managers to identify bottlenecks in the production processes, inefficiencies in the usage of machinery and causes for wastage of resources. On the other hand, BDA capable systems have also been linked to better workforce utilization as they permit managers to monitor the performance of each employee.

  3. c.

    BDA capable systems may also positively impact an organization’s ability to pursue collaborations with partners. Particularly, BDA capable systems may improve knowledge flow and facilitate sharing between partners as BDA systems are frequently built around jointly-developed database architectures (Vera-Baquero, Colomo-Palacios, & Molloy, 2016).

  4. d.

    BDA capable IS may play a role in fostering the organizational capability of identifying and seizing new opportunities. Because of the newly extractable information, BDA capable systems can improve organizational exploitation and exploration capabilities and, consequently, ambidexterity (Rialti et al., 2018).

In short, big data is progressively influencing organizations’ competitiveness.

Better performance of organizations is found to be positively linked with the big data impact and the effect of BDA capable systems on organizations, according to the relevant literature (Gunasekaran, Yusuf, Adeleye, & Papadopoulos, 2018). Particularly, organizations may see an improvement over the time in the performance metrics which include supply chain efficiency, workforce utilization, and production processes efficiency and economic performance metrics with the utilization of big data and BDA (McAfee & Brynjolfsson, 2012). Nevertheless, all the positive impacts of big data are dependent on the acceptance of big data by organizations’ managers (Teece, 2009). Thus, organizational processes, which include allocation of resources, organization, and utilization, play a vital role in the ability of an organization to obtain the benefits that can be derived from big data (Côrte-Real, Oliveira, & Ruivo, 2017).

In this perception, pertinent literature pointed out that for a modern organization it is important not only to implement BDA systems and processes but also to develop the so-called organizational BDA capabilities (Wamba et al., 2017). Organizational BDA skills are an overall ensemble of capabilities that comprise BDA infrastructure flexibility (the BDA capable IS and processes that were previously explained), BDA management skills and BDA personnel skills. Indeed, the three of them are fundamental for any organization trying to face big data advent. If the BDA infrastructures, which include all the technical IS that are capable of collecting, storing, processing, and analyzing big data, are too unyielding, and are unable to self-adapt to different kinds of data, they cannot make sure that there is any data flow in any situation. BDA management skills are fundamental to the selection and the implementation of the correct BDA infrastructure and to extract the right information from the datasets. On the one hand, managers need to have enough skills to decide which technical solution is the best for their organization, and on the other hand, they need to be skilled enough in BDA to understand the information extracted from the datasets and to make the right decisions according to the new insights. The personnel should be skilled in BDA as personal skills are related to the lower rejection of BDA within the organization—less resistance to new IS implementation—and better functioning of the BDA infrastructure. Additionally, as employees are often the ones doing the analysis of the data, they need adequate skills, both, to maintain the infrastructure working and to identify the right data to be analyzed.

From this perspective, the studies investigating the importance of big data for organizations often focus on dynamic capabilities (Wamba et al., 2017). In fact, the development of such an intertwined system of capabilities is necessarily related to the organizational capability to use existing elements in a different way to adapt to new situations.

2.3 BDA, Dynamic Capabilities, Ambidexterity and Performance

Teece, Pisano, and Shuen in 1997 initially coined the concept of dynamic capabilities. As per their seminal manuscript, the organization’s ability to adapt a resource base to make it work in different situations gives the crux of the concept of dynamic capabilities (Eisenhardt & Martin, 2000). In short, the expression ‘dynamic capabilities’ may be reassumed by the organization’s degree of ability to adapt sufficiently and in time to any change in the environment by modifying external or internal resources and processes, depends on existing competencies (Teece, Pisano, & Shuen, 1997).

While dynamic capabilities definitions link changes to organizational improvisation, dynamic capabilities may “actually consist of identifiable and specific routines” (Eisenhardt & Martin, 2000, p. 1107). Indeed, the best practices within an organization may get diffused by some organizational routines and processes. Complete organizational routines can be split into smaller routines or small processes, that can be considered as the ‘bricks’ for forming the complete routine or process (Eisenhardt & Martin, 2000). In the case of a mutated environment, these bricks may be assembled differently to create a new routine or process so that the organization can survive and succeed. This phenomenon signifies the importance of knowledge and expertise accumulation within an organization and highlights the fact that this knowledge base is useful in multiple situations. Dynamic capabilities are, therefore, related to reusable routines or processes.

Theoretical perspectives that unpack the effects of big data and the BDA system’s availability in an organization have frequently been adopted by scholars as dynamic capabilities, in the big data era (Erevelles et al., 2016). Big data is indeed a resource related to information that requires routines and processes to be turned into meaningful insights. For increasing the efficiency of big data analysis, previous expertise of analysts and managers can play a key role. The knowledge spanning the organization about BDA processes and BDA capable IS is indeed fundamental to the extraction of insights from big data. Coherently, a certain degree of routinization may also be an advantage to process big data. The smaller routines that represent the bricks may be useful in solving different big data-related problems (Braganza, Brooks, Nepelski, Ali, & Moro, 2017). Building on these findings, it has been observed that the presence of existing adaptable routines and processes to analyze big data may increase organizational flexibility and agility (Chen, Chiang, & Storey, 2012), a phenomenon related to the huge amount of information that BDA systems can process and provide to managers in real-time. The consequence of this is that the organization may better identify and exploit the opportunities arising in the market, hence, imbibing the characteristics of an ambidextrous organization (Rialti et al., 2018). Thus, while traditional information systems have emerged almost as a factor fostering organizational rigidity, BDA capable IS have the potential to enhance organizational dynamism by increasing organizational agility, flexibility, and opportunity exploitation potential. The characteristics of big data and BDA systems and processes coupled with the possibility to adapt different routines (bricks) to different kinds of data. Hence, the adaptation capability of BDA systems and processes to modify themselves for different kinds of data may allow the organization to thrive in different situations and to survive eventual environmental shocks. In this way, in line with the basic assumptions of dynamic capabilities research (Teece, 2009), it is evident that this adaptation capability may positively influence organization performance (Wamba et al., 2017).

Moving on from these premises, researchers have explored the effect of big data and BDA capable processes on systems using dynamic capabilities as the main theoretical lens. For example, researchers have utilized dynamic capabilities for interpreting the means in which BDA capable processes can create competitive advantages. To overcome rivals, the BDA processes are based on re-applicable routines, which can generate new information (Braganza et al., 2017). Hence, BDA processes are equivalent to continuous knowledge generation and diffusion processes. This way, these processes give analysts and managers the ability to recognize good opportunities and they can reject opportunities that are not profitable. Secondly, to study the effect of big data on marketing strategies, a theoretical approach to analyze the dynamic and adaptive capabilities of BDA was used (Khan & Vorley, 2017). This phenomenon has been associated with the BDA processes’ and systems’ potential to unpack the consumers’ behavioral patterns and, therefore, to promote the creation of marketing strategies that are customized (Erevelles et al., 2016). Thirdly, the organizational and financial performance may be influenced by both, process-oriented dynamic capabilities and BDA (Wamba et al., 2017). Within an organization, the diffusion of BDA processes and systems is influenced by the ability of processes to adapt to changing situations. Finally, another study reported that a competitive advantage can be generated based on the ability of BDA processes and systems to adapt to different kinds of data and environments that are continually evolving (Sivarajah, Kamal, Irani, & Weerakkody, 2017).

With this theoretical perspective, it is possible to imagine and to understand in-depth how the overall three specific BDA capabilities are related to organizational performances. Indeed, these three capabilities are on their own related to routines that are necessary to transform big data into value. In respect of the first capability, BDA infrastructures (which are the ensemble of information systems capable to collect, store, process and analyze big data) should be self-adaptable to different kinds of data. This capability related to systems is indeed fundamental to ensure that technologies will be able to process different data flows and formats in any situation (Labrinidis & Jagadish, 2012; Rialti et al., 2018). Thus, BDA infrastructure flexibility (or capability) may generate routines concerning the management of information flows and to deal with different problems. Next, BDA managerial capabilities are fundamental in the selection and the implementation of the BDA infrastructure that is right, and in the identification of the right information extracted from the datasets. Managers should, in fact, have enough competence to decide which technical solution is the best for their organization. Similarly, they need to have enough data analytic skills to take the right decisions according to the newly available data. Such kind of capabilities could influence significantly how managers behave in respect of data related concerns and affect the strategy they select to challenge new information. Yet, they could improve managerial skills concerning the identification of the most valuable information among data at disposition. Insights are frequently hidden among data that may seem worthless, and individual BDA capabilities of managers may influence looks of information and the suggestion these figures may provide to employees. The personnel, finally, should be skilled in BDA too. Personnel’s BDA capabilities are related to lower rejection of BDA within the organization, less resistance to new IS implementation and better functioning of the BDA infrastructure. Additionally, since employees are often the ones conducting the analysis of the data, they need adequate skills to identify the right data to be analyzed (Wamba et al., 2017). Only if employees are skilled enough, they may report proper insights to managers.

For these reasons, the dynamic capabilities are also useful guidance to unpack how BDA is related to organizational ambidexterity. In particular, it is observed that ambidexterity is considered as one of the dynamic capabilities of an organization. Ambidexterity, indeed, is the organizational capability to use existing capacities to explore the environment and exploit opportunities available.

The pertinent literature shows that the structure of published research regarding big data needs to be properly studied (Braganza et al., 2017; Bresciani, Ferraris, & Del Giudice, 2017). Specifically, a better understanding is required about how literature about big data and dynamic capabilities is organized. It is observed that the bibliometric method is used by a maximum number of manuscripts to investigate other streams of literature that are related to big data (Ardito et al., 2018) and very few of them have focussed on this particular topic.

2.4 Exploring Existing Literature on Big Data, Dynamic Capabilities and Ambidexterity: Defining the Research Methods

Before the theoretical and empirical explorations of the relationship between big data, BDA, dynamic capabilities and ambidexterity, it is necessary to analyze the existing literature on the topic. In this sense, two methodologies have been selected: bibliometric method and the systematic literature review.

Comprehensive maps have been generated by using a bibliometric method. These maps give the knowledge structure of a given stream of literature. However, while studying a field of research that is still developing, it is important to implement a literature analysis accurately. For this purpose, Caputo, Marzi, Pellegrini, and Rialti (2018) used a combination of techniques, bibliometric analysis, and review of the literature. The first step included a bibliometric analysis. The results of this analysis were collected and the literature was reviewed. The basis of the bibliometric analysis is the technique of ‘visualization of similarities’ (VOS) (Van Eck & Waltman, 2010). The authors followed the method recommended by Tranfield, Denyer, and Smart (2013) for the review of the literature. The complete process of two methodologies consists of six steps. These six steps are described in detail below.

  1. a.

    Step 1: The Query. The first step is associated with searching the database, Clarivate Analytics Web of Science Core Collection, for the specified research query. The most valuable and high-impact data has been collected in this database and it is very useful for bibliometric studies because of its high reliability. In detail, the search for the query process is fundamental, as the query is an ensemble of words used to identify a set of papers in international online repositories. The procedure of selection of a research query begins with a review of literature of the manuscripts focussed on using BDA for management and considering the primary underlying theory as dynamic capabilities for grasping all of the terms that are utilized for explaining the phenomena that needs to be analyzed (i.e. Akter et al., 2016; Wamba et al., 2017). After multiple iterations targeting to outline an expansive research query, the following query was selected:

    • TS = ((“big data*” OR “big data analytics*”) AND (“dynamic capabilities*” OR “ambidexterity” OR performance*) AND (organization* OR firm* OR business* OR enterprise*))

    This query will identify most of the existing research on the topic. Indeed, specific words such as “big data” or “big data analytics” allow to collect the papers on such topics in relation to the theoretical streams of the research because of use of words such as “dynamic capabilities”, “ambidexterity” and “performance”. Yet, the inclusion of words like, “organization”, “firm” and “business” allows the inclusions of papers coming from the business and management field only as business organizations are the main object of the research.

    In the titles, abstracts, and keywords, a full search of the chosen terms is performed by the ‘TS’ operator. The search can be limited by document type, such as “articles, books, book chapters, book reviews, early access articles, and editorial material”. The timespan of search is also defined such as a ten-year cross-section—namely 2008 to 2018—can be considered as timespan.

    In order to be sure about the inclusion of the most significant papers, the query is iterated on the three most relevant databases, EBSCO, Scopus and Web of Science (WOS). The searches on all the relevant databases may offer similar results in terms of the number of papers. The analysis using WOS database is considered the most up-to-date on extremely recent literature in management. For the example query and the above conditions, 423 entries are obtained.

  2. b.

    Step 2: The Inclusion Criteria. The second step is focused on defining the criteria of inclusion of the documents to be utilized in the study, followed by the analysis and selection of each document manually. In detail, the inclusion criteria are used by a researcher to clean the dataset identified in Step 1. Hence, the author decides which paper is included for review based on the presence of the inclusion criteria in the paper itself.

    Using two inclusion criteria, the first of which was the most accepted definition of big data as “datasets whose size is beyond the ability of a typical database software tools to capture, store, manage and analyze” (Manyika et al., 2011, p. 1). The authors then excluded manuscripts without a research perspective focused on dynamic capabilities and manuscripts that did not belong to management-related literature. This led to the dataset being reduced from 423 to 217 entries.

  3. c.

    Step 3: Collaborative Analysis of the Papers. The third step consists of critically analyzing the manuscripts that were selected in the previous step and working to derive an effective knowledge of BDA and how it is associated with dynamic capabilities and organizations’ performances (Wamba et al., 2017). In this perspective, during this phase, it becomes possible for the users to be safe about the content of an included research and to critically explore the perspectives considered by original authors.

  4. d.

    Step 4: Bibliometric Indicators Analysis. The fourth step is regarded as the preliminary segment of the bibliometric analysis. Specifically, the research volume is analyzed using the activity indicators which reflect how the literature evolved quantitatively over time. Particularly, the number of papers for a year and the most prolific authors are identified. In addition, an analysis concerning the most institutions and the countries their authors (and related manuscripts) come from has also been performed with the help of the analytics functions of VOSviewer 1.6.5.

  5. e.

    Step 5: Bibliometric Analysis. The fifth step involves proper bibliometric analysis. For this purpose, as suggested by Van Eck and Waltman (2010), to aggregate the manuscripts, an aggregation mechanism is utilized by using the VOSviewer 1.6.5 algorithm with bibliographic coupling. When two papers cite a common third paper in their references, it is indicative of bibliographic coupling. These two papers can then be referred to as bibliographically coupled. The output generated by the VOSviewer is in the form of a map in which the relatedness of the terms can be inferred by the distance between the terms. A strong association between the terms is reflected by a smaller distance (Van Eck & Waltman, 2010). In addition to this, the knowledge base diversity is highlighted by the cluster analysis in an aggregate way: if the manuscripts are in the same cluster, then based on their shared references they are a group that is strongly linked together. Thereby, based on similarity, a stream of research is represented in a cluster. The manuscripts are presented in a useful way on the map generated by VOSviewer which optimizes their visualization; thus, there is no meaning of the axes of the map (Van Eck & Waltman, 2010).

  6. f.

    Step 6: Systematic Literature Review. This last step includes the process of literature review (Tranfield, Denyer, & Smart, 2013) and is based on the results generated from the VOS aggregation. The clustering results generated by VOSviewer were used and the most influential manuscripts inside the displayed clusters are analyzed to highlight their main areas of interest. In detail, the focus during this phase is on the exploration of the linkages existing between papers moving from their contents and the perspectives used by previous researches. Similarly, the evolution of literature is observed.

2.5 Results of Bibliometric Analysis

In this section, the results of the bibliometric analysis are presented and analyzed. The distribution of manuscripts over the years is presented in Fig. 2.1.

Fig. 2.1
figure 1

Manuscripts’ temporal distribution

Full size image

As seen in Fig. 2.1, most of the manuscripts that are selected are published within the last 5 years. Only 2 manuscripts were published in 2012 and before this, there were no publications. Although since the previous decade, the significance of big data and BDA for management has been researched extensively, there have been explorations into this field with a focus on the theoretical principle of dynamic capabilities only during the last five years. Based on the pattern observed by the number of manuscripts that are published each year, this area has still not reached maturity. Indeed, every year, the number of manuscripts published on this topic is increasing.

In regards to the principal outlets selected by authors on the topic, as shown in Table 2.1, Journal of Knowledge Management and Journal of Business Research present the most elevated number of manuscripts (8 and 7, respectively). Such a phenomenon is obviously related to the interest of these two particular journals toward the multifaceted emerging importance of Big Data. The other journals dealing with the topic mostly deal with the operation management area. Indeed, one of the principal effects of the implementation of BDA is the evolution of traditional operation management procedures. Finally, as it is possible to observe, counterintuitively, several journals concerning environmental concerns (i.e. Sustainability and Journal of Cleaner Production) are present on the list. This fact is related to scholars’ attention at the impact of BDA in reducing organization impact on the environment.

Table 2.1 Principal outlets for publications on big data and ambidexterity
Full size table

The most prolific authors, instead, are the pioneers of this field of research. As an example, Stephen Childe, Angappa Gunasekaran, Rameshwar Dubey, Samuel Fosso Wamba, and Shahriar Akter are the most prolific authors on these topics. As a matter of fact, they were among the first authors to deal with the topics related to the implementation of BDA within modern organizations. Moreover, they were among the first authors to use dynamic capabilities as a theoretical lens to observe how BDA could change the way organizations compete. The results are shown in Table 2.2.

Table 2.2 Most prolific authors
Full size table

To what concerns the most prolific institutions and the most prolific countries, results are shown in Tables 2.3 and 2.4 respectively.

Table 2.3 Most relevant institutions
Full size table
Table 2.4 Country of origin of the manuscripts
Full size table

In Fig. 2.2, the VOS analysis results are presented for the manuscripts that are most influential and only the first author’s surname is included.

Fig. 2.2
figure 2

VOSviewer result

Full size image

From the above analysis of the 217 manuscripts, five clusters can be identified. As set in the query, the selected manuscripts in the clusters have used dynamic capabilities as a theoretical principle.

In detail, the analysis of 5 clusters shows: a cluster (the blue one) containing papers about Big Data, Dynamic Capabilities, Supply Chain Management, and Performance; a cluster (the red one) including papers about Big Data, Dynamic Capabilities, Production Processes, and Manufacturing; another cluster (the green one) is about Big Data, Dynamic Capabilities, and Strategy Making; and the last cluster (the yellow one) dealing with Knowledge Exploitation Through BDA Systems, and, finally, a cluster (the purple one) about BDA and performance.

The contents of the clusters are explored in-depth in the next section.

2.6 Systematic Review of Literature

Comprehensibly with another research which contains both, a bibliometric analysis and a literature review (Wamba & Mishra, 2017), the authors followed a similar approach, resulting in the analysis of the ten most influential manuscripts belonging to each identified cluster (Caputo et al., 2018; Rialti et al., 2019). In this perspective, the selected papers were then characterized by the highest scores in terms of normalized citations. These manuscripts, in addition, also get selected as they were, in most of the cases, the ones most closely connected to others, according to VOSviewer 1.6.5.

2.6.1 Blue Cluster: Big Data, Dynamic Capabilities, Supply Chain Management and Performance

This cluster includes manuscripts dealing with how big data affects supply chain management, related dynamic capabilities, and performance. BDA capable systems and processes may significantly affect supply chain management (Gunasekaran et al., 2018). The routinization of BDA processes may have an important role in making sure that these processes are able to adapt to different situations. Hence, there may be an increase in organizational agility, dynamism, and flexibility because of BDA processes specifically identifying problems and opportunities in supply chain management. According to a similar perspective, increased sustainability has also been associated with the improvement of the supply chain management ability of BDA processes. This phenomenon is associated with the fact that as efficiency increases in the management of supply chains, the quantity of waste may be reduced (Papadopoulos et al., 2017). Similarly, BDA processes may influence supply chain management, more precisely, how the application of BDA enhances and speeds up managerial decision processes about procurement (Lamba & Singh, 2017). These findings have also been supported by the empirical exploration by Ghasemaghaei, Ebrahimi, and Hassanein (2017).

It emerged that, to a certain degree, the impact of BDA on supply chain management, may positively affect stock replenishment and product availability. For example, the ways in which BDA systems based on networks of sensors may increase the efficiency of retail chains have been examined by Li and Wang (2017) and by Wu and Lin (2018). The organization can, thus, dynamically re-organize processes according to emerging needs (Stefanovic, 2015).

Increased supply chain and stock replenishment efficiency may, as a by-product increase organizational performances. Specifically, Martin, Borah, and Palmatier (2017) and Money and Cohen (2018) have concluded that the ability to forecast resource usage may improve performance. In fact, when a business is capable to increase the accuracy of predicted resources’ consumption, it may be capable to achieve significant savings in costs.

In respect of the relationship of this phenomenon with the notion of dynamic capabilities, it is possible to observe how the more the predictions become a routine in operative practices the more they will become fast and accurate.

2.6.2 Red Cluster: Big Data, Dynamic Capabilities, Production Processes, and Manufacturing

The second cluster contains manuscripts on big data and production processes related to dynamic capabilities. BDA capable IS or processes may offer managers an opportunity to monitor production processes (Chen, Mao, & Liu, 2014). From this perspective, it emerges that with BDA methodologies, managers may control any part of the production process associated with less resource wastage (Sivarajah et al., 2017). Similarly, Tan assessed how BDA may foster innovation in all material processes, from supply chain to production. According to Wu, managers may get the ability to make informed choices on strategies based on the large quantities of data that can be provided by BDA. BDA adoption was also studied by Zudor for finding a way to make strong collaborations amongst partners belonging to the same production chain. The architecture of BDA capable IS may encourage data sharing which is related to the efficiency of the production process between partners. Finally, BDA capable IS may ensure continuous control over the quality of the production Kibira, Morris, and Kumaraguru (2016).

Additionally, some manuscripts included in this cluster outline how BDA capable systems may improve the efficiency of processes involving products or services (Erickson & Rothberg 2017; Moeuf, Pellerin, Lamouri, Tamayo-Giraldo, & Barbaray, 2018) and show that big data competencies are fundamental in making this possible (Wang & Cotton, 2018).

In detail, it has been observed how BDA in most of the modern businesses could decide in an easier way to prioritize the production of a batch of products with respect to machinery’ capacity. Algorithms for BDA tools reasoning are, sometimes, more accurate than the human mind, in the identification of production priorities, production slots and in the daily division of work. In addition, compared to the human mind, artificial intelligence is unbiased, as decisions taken by computers are based only on objective data. All these findings are consistent with existing and emerging literature on Industry 4.0. In conclusion, automation could help managers to increase the efficiency of the business and generate greater income and, BDA capable IS could integrate with automatic machines by embedding their AI with the right information to take autonomous decisions.

2.6.3 Green Cluster: Big Data, Dynamic Capabilities, and Strategy Making

Scholars have substantially focused on the impact of big data on strategy making processes because of the information that big data can contain.

BDA can provide managers with opportunities to know more about their consumers and insights arising from BDA capable IS can offer real-time actionable knowledge (Prescott, 2014). BDA capable IS can also enhance strategic decision-making processes related to marketing (Erevelles et al., 2016). It has been shown that information about consumers can allow managers to react dynamically to evolving consumer’s preferences. The integration of BDA related routines or processes in decision making has been also seen to influence organizational adaptive capabilities. Such a phenomenon, however, isn’t typical only of organizations producing goods, but can also be seen in services providers, to better deal with the requirements of their consumers and to effectively monitor the utilization of resources. (Wang, Kung, & Byrd, 2018). The manuscripts in the cluster show how managers can better allocate resources to respond to critical situations. BDA may allow managers to respond to changes concerning day-to-day operations (Melnyk, Flynn, & Awaysheh, 2018). This phenomenon is related to the knowledge flow generated by BDA capable IS, which may change the managers’ way of thinking and acting. Additionally, such systems may allow knowledge to be distributed across the whole organization (Wang, Zhu, Song, Hou, & Zhang, 2018).

Manuscripts belonging to this cluster deal also with the importance of the alignment between BDA capabilities of organizations and their expected outcomes. As stressed by Akter et al. (2016), if big data systems and processes are aligned with managerial objectives, these can offer proper information required to managers for formulating satisfactory strategies. In these instances, organizations may create value from big data (Bedeley, Ghoshal, Iyer, & Bhadury, 2018; Mamonov & Triantoro, 2018).

BDA could, thus, influence how managers decide to behave and plan the strategic paths to be followed by an organization. This is because BDA managers may potentially decide using accurate information regarding patterns concerning both consumers and competitors.

2.6.4 Yellow Cluster: Knowledge Exploitation Through BDA Systems

The fourth cluster includes manuscripts about the importance of BDA capable IS and processes for knowledge generation and its exploitation. In this perspective, He stressed how organizations that aim to utilize the informative potential of big data should adopt developed processes capable to extract the best contents from big data. Yet, it emerges that the systems to collect and analyze the data should be integrated with the systems capable to diffuse the knowledge within the organization. This is necessary to make knowledge flows arrive at the right decision-maker. These findings are also confirmed by Xie, Kwok, and Wang (2017) for big data in the hospitality sector. Indeed, it emerged that when the right information arrived at the right hotels, it may gain competitive advantages. Therefore, managers may give better responses to consumers’ requests (Pournarakis, Sotiropoulos, & Giaglis, 2017). Similarly, Arnaboldi, Azzone, and Sidorova (2017) evaluated how big data may be advantageous to firms only if the insights may be used by the right manager to deal with the equilibrium between the organization and the surrounding environment.

Other most influential manuscripts deal with very closely related topics. Lee (2017), for example, has emphasized that the major challenge of this era of big data is how the required information can be extracted from big data and how to exploit this new knowledge. Maklan, Peppard, and Klaus (2015) explored how knowledge generated from big data may improve marketing strategies. Then, Tirunillai and Tellis (2014) investigated the techniques to extract knowledge from big data.

2.6.5 Purple Cluster: BDA and Performance Related Outcomes

Superior performance is the target of most managers and managers are usually more than eager to follow any possible path to improve the performance of their organization. In this perspective, BDA capable IS and processes may represent a fundamental tool to improve the performance of organizations (Chen et al., 2013). As emphasized by Gani, Siddiqa, Shamshirband, and Hanum (2016) and Kowalczyk and Buxmann (2015), organizational capabilities may be influenced by BDA which can be used for identifying and exploiting opportunities that exist in the external environment. On the one hand, BDA capable IS may help the organization in scanning the surrounding environment for additional information, while on the other hand, managers have at their disposition the insights to extract new information extracted from BDA, and for formulating the strategies required for capturing and exploiting the new emerging opportunity. This is linked to the real-time insights that may be mined from big data and their accuracy. At present, managers have access to the perceptions and the ideas of consumers as they are freely and easily available on the internet and they may be analyzed with the help of proper tools (Christensen, Nørskov, Frederiksen, & Scholderer, 2017). The alignment between BDA capable IS and processes by using knowledge management tools can result in the diffusion of these insights within the organization. This may help in making strategic managerial decisions that are supported by accurate information. An organization may improve its performance by reducing costs or increasing revenues (Centobelli, Cerchione, & Esposito, 2018).

The manuscript by Vera-Baquero et al. (2016) confirms that BDA is associated with better performance of an organization as it facilitates better monitoring of internal processes by the manager. BDA managers may get familiarized with the performance of the processes, identify bottlenecks or problems, and may find solutions to the problems. Thus, managers may accomplish an improvement in the performance of the organization by increasing the individual processes’ performance (Ramanathan, Philpott, Duan, & Cao, 2017).

As it is possible to observe from Fig. 2.2, such a cluster is also intertwined with the majority of other ones. In fact, several papers from the purple cluster are positioned amidst the ones of the yellow, blue and red clusters. In this perspective, it is then possible to assess that literature exploring the impact of BDA on organizational performances is also related to other streams like the ones exploring the role of BDA in improving supply chain efficiency and in process management. These observations suggest that the effect of BDA on performances exists where these capabilities, processes, and technologies are capable to influence organizational aspects. BDA alone then may not have any positive effect if the organization is not willing to change accordingly.

2.7 What’s Next?

The existing knowledge is systematized after the review of important manuscripts on big data and dynamic capabilities. Firstly, it is established that there are five clusters in all the research on dynamic capabilities (Akter et al., 2016; Wamba & Mishra, 2017). Then, assessment is done to find how BDA can increase dynamic capabilities related to the adaptation of operation processes and supply chain strategies to mutated environments (Gunasekaran et al., 2018). It was also observed that BDA is frequently linked to superior performances (Prescott, 2014). Indeed, as BDA capable IS and processes may adapt to changing data and situations, they can affect organizational dynamic capabilities by providing insights on future strategies that need to be implemented. Consequently, organizations may re-organize processes and re-allocate resources, which may affect performance positively (Wamba et al., 2017).

From the example studied, according to Akter et al. (2016), the alignment between the capabilities of big data and manager’s expectations regarding the implementation of BDA should be monitored by the manager. It is also found that managers may not find the insights needed for developing new strategies if the big data capabilities and objectives of the organization are not aligned.

What clearly emerged from the example is that big data could influence performance with the help of newly available information and become instrumental in the development of new routines to exploit it. Yet, extant literature still must clarify how BDA could become a dynamic capability, and how it could influence ambidexterity.

Moving from these premises, the next chapters will first try to outline the mechanisms concerning how BDA could influence organizational routines from a theoretical standpoint.