Big Data Analytics and Value Creation: A
Meta-Analysis
Master Thesis Business Administration –
Change Management
Sietze Hamming S3272613 Obrechtstraat 27
8031 AH, Zwolle, The Netherlands i.i.hamming@student.rug.nl
University of Groningen Faculty of Economics and Business
MSc Business Administration – Change Management
January 2019
Supervisor: dr. John Q. Dong Co-assessor: dr. I. Maris-de Bresser
ABSTRACT
Big Data Analytics (BDA) is a concept that is researched a lot in recent years, because the amount of data is growing and increasingly important for creating business value. However, BDA has been conceptualized and operationalized in different ways and various forms of business value associated with BDA are examined in the literature. In this study, a meta-analysis is performed to understand the relationship between BDA and value creation. In addition, other research choices (e.g. Kohli & Deveraj, 2003) have been considered in the meta-analysis. From the results, when BDA is conceptualized as a capability and value is conceptualized as organizational performance, a stronger relationship between BDA and value can be found. Furthermore, the duration of data and national contexts also matter to explain the magnitude of the relationship between BDA and value. As a general guidance for future research, BDA defined and captured as a capability and linked to organizational performance is a better research choice. Moreover, data should be collected over a longer period and firms in European countries seem to benefit more from value creation of BDA.
TABLE OF CONTENTS
ABSTRACT _______________________________________________________ 2 1. INTRODUCTION ________________________________________________ 4 2. METHODS ____________________________________________________ 7 2.1 Meta-analysis ________________________________________________ 7 2.2 Sample _____________________________________________________ 8 2.3 Coding _____________________________________________________ 9 2.4 Analysis; Weighted Least Square regression (WLS) __________________ 123. RESULTS ____________________________________________________ 16 4. DISCUSSION _________________________________________________ 19
4.1 Main findings________________________________________________ 19 4.2 Concept of BDA ______________________________________________ 20 4.3 Concept of value ___________________________________________ 23 4.4 Combining BDA and value; an overall analysis ______________________ 27
5. IMPLICATIONSANDLIMITATIONS _________________________________ 29
5.1 Theoretical implications and future research _______________________ 29 5.2 Practical implications _________________________________________ 30 5.3 Limitations _________________________________________________ 31
1. INTRODUCTION
Nowadays, big data analytics is becoming an increasingly more important topic for managers, practitioners and researchers (Chen, Chiang & Storey, 2012). According to Fichman, Dos Santos and Zheng (2014), we entered a golden age of digital innovation. A new digital infrastructure (e.g. computers, mobile devices) is created, and thereby new technologies (e.g. analytics and big data), that transform the way we live and work (Fichman, et al. 2014). Baesens et al. (2016), identified five major sources where big data arises from; (1) Large enterprise systems, (2) online social graphs, (3) mobile devices, (4) Internet-of-things, (5) Open data /public data. Data is growing and will be growing at an exponential rate for the foreseeable future (Manyika et al., 2011). This is one of the reasons why interest in Big Data Analytics did increase (Bumblauskas, Nold, Bumblauskas & Igou, 2017).
Before we dig deeper in the concept of BDA and value creation, we first need to elaborate more on the concept of big data. Because of the youngness of the concept a consistent and wide used definition is not available (De Mauro, 2016). Many researchers (e.g. Lycett, 2013; Erevelles, Fukawa & Swayne, 2016; Laney, 2001; McAfee & Brynjolfsson, 2012), did research about the concept of big data. They came up with the classic characteristics of big data, namely the ‘3V’s’; volume, velocity and variety. In recent years, researchers (Wamba et al., 2015; Ahmed & Ameen, 2017) added two V’s in addition to the classic V’s, to wit, value and veracity. Based on the 5V’s definition of big data, De Moura (2016, p.131) comes up with overarching definition for big data; “Big Data is the information asset characterized by such a High
Volume, Velocity and Variety to require specific technology and analytical methods for its transformation into Value.”
et al. (2012), to capture the value of BDA, organizations need to make data-driven decisions. It enables managers to make decisions based on evidence instead of intuition. The way organizations use technologies has an important influence on the value it will capture from it. Organizations need to transform their decision making processes in order to generate value through the use of analytics (Sharma, et al., 2014). In addition, if an organization wants to capture the value of BDA, the organization needs to move their analytics from solely IT to their core business and daily operations (Davenport, Barth & Bean, 2012). These are examples how to generate or capture the value of BDA for the organization. Some researchers, did define organizational value in more measureable terms like, Return on Equity (ROE) (Chi, Ravichandran & Andreveski, 2010), Ideated Innovation (Joshi, et al., 2010), or as market and operational performance (Gupta & George, 2016). In general, the definition of value is not always consistent among the different papers. One major difference is, the value capturing of BDA is divided in a qualitative and quantitative way, leading towards different conceptualizations and substantiation.
Fosso Wamba et al. (2017, p.356), came up with overall definition of BDA based on the 5V’s; “BDA is a holistic approach to managing, processing, analyzing the 5 V
related dimensions. In order to create ideas for delivering sustained value, measuring performance and establish competitive advantage”. BDA is a broad term with different
Like mentioned above, how to threat and define the concepts of BDA and value differ among researchers. Some argue that value is achieved through improved decision making (Sharma, et al., 2014), where others (Chi, et al., 2010; Joshi, et al., 2010) define value more as an outcome that can be measured like, ROE or ideated innovation. Besides, researchers did perform individual research relating to BDA and value creation. A general overview of all the conceptualizations together is missing. To make sense of the literature about the relationship of BDA and value creation, this paper will make an overall view of the current conceptualization, and tries to bundle them together to make a more comprehensive overview. In order to come up with a general overview, this paper consists of a meta-analysis. This method is important because it will map the current quantitative literature in one analysis, trying to find which concepts of BDA are affecting value creation, and which value construct is an important outcome(s). With this paper more general conclusions can be drawn on this topic, leading to a more solid argumentation, if the relationship between BDA and value creation, absolutely exists.
2. METHODS
A meta-analysis tries to identify a general trend among the results of different studies. In order to combine evidence from different studies with similar theoretical predictions, or about a relationship around the same phenomenon (Bangert-Drowns, 1986; Combs, Crook & Rauch, 2019). Because of this, a meta-analysis is suitable for this paper. Often BDA and value creation are researched by individual researchers, without a linkage among the different articles. To make sense of all the individual papers, a meta-analysis can help to make a general overview and identify, if available, different trends on the topic. The remainder of this section has the following build-up; first the general steps for performing a meta-analysis will be described. Then will explained the way towards the final sample size. The third part is about the variable coding, ending with explaining the analysis performed.
2.1 Meta-analysis
Kohli and Deveraj (2003) performed a meta-analysis and adopted the method of Glass et al. (1981). Glass et al. (1981) introduced how to perform a meta-analysis by propose the following steps (these steps are redefined to this study);
(1) Development of framework listing factors that contribute to explaining BDA and value creation. This step is outlined in more detail in the discussion
section.
(2) Selection of studies to be included in the analysis. How the final sample size
is generated and completed is mention in the section sample (§2.2);
(3) Documentation and coding of the characteristics of studies included in the analysis. See the codebook of the final sample size in the appendix;
(4) Statistical meta-analyses through regression. For this study the weighted
least square (WLS) regression method is used. Why this method is suitable for this study is described in the section analysis (§2.4);
(5) Documentation of the findings from the WLS-regression, and explanation of the conceptualization of BDA and Value from the sample size. This step is
2.2 Sample
In this meta-analysis the focus is on the relationship between BDA and Value. Focusing on empirical research, with a quantitative focus. To estimate the sample size a broad search is performed. For both searches, performed on November 1st, the following
keywords were used; ‘“Big Data” OR “Analytics”’. The first search gave a result of 110.351 articles. For the second search the “basket of eight” Information Systems (IS) journals (see table 1) (Günther, Mehrizi, Huysman & Feldberg, 2017) were used as source, these journals are listed as the most important journals in the IS research field. Because they are the best journals in the field of IS, they publish the most important work, meaning that the quality of the papers is high and reliable. In addition, the journals; ‘decision support systems’ and ‘Information and Management’ were added to the research. The journal of decision support systems was added, because this journal is, besides its high quality of papers, important journal that publish papers on different kind of topics (e.g. data base management, user interface management), BDA could also be a topic that is discussed in various published papers. The Information and Management journal is added, because it is a high qualitative journal that publishes papers on different topics of information systems. These two journal are not added to the “basket of eight” IS journals list, however, the quality of the papers is high and the papers have important contributions towards the IS research field. Thus, including them will improve the sample size for the meta-analysis.
The second search (i.e. “basket of eight” journals plus ‘decision support systems’ and ‘information and management’) gave a result of 428 articles. The first search gave too many results for performing this meta-analysis, therefore the sample of the second search is the starting point.
presentation of correlations. This criterion is applied to the sample size of 401 articles, after the executing the second selection the sample size was 109 articles. In the third selection, the 109 articles were studied in more detail. These 109 articles did have some kind of statistical analysis about ‘analytics’ or ‘big data’, meaning they could be relevant for this meta-analysis. The third criterion was; the article should discuss or conceptualize the concept of analytics, the way it is discussed or conceptualized did not matter as long as it was related to analytics. And it should present a correlation between analytics and value, here is the conceptualization also not important. Applying this criterion on the 109 articles, each articles was studied in more detail (i.e. reading the methodology and in some cases the appendix where the conceptualization is defined), this lead to a final sample size of 14 articles that met all the criteria. For the overview of the final sample size see the codebook in the appendix. In the following paragraph, the rest of the codebook will be explained.
Table 1 List of Journals
Journal Number of
Articles European Journal of Information Systems 37
Information Systems Journal 14
Information System Research 50
Journal of Information Technology 61 Journal of Management Information Systems 25 Journal of Strategic Information Systems 18 Journal of the Association for Information Systems 12 Management Information Systems Quarterly 58
Decision Support Systems* 110
Information and Management* 58
Total 428
* = Not included in the “basket of eight” IS journals.
2.3 Coding
Deveraj; (a) sample size, (b) method design, (c) duration, (d) country location, (e) industry sector. In addition to these variables two other variables in the codebook are the constructs of BDA and value. From each article from the final sample size these variables are reported in the codebook, with the correlation, reliability and significance of each article (see appendix for the codebook). Further in this section each variable will explained in more detail, with the explanation of how it is coded for the analysis, see also table 2 and 3, here the coding variables and coding scheme are presented.
2.3.1 Sample size
The sample size is often a given number in a study. It is the number of observations which the study uses for their analysis. Which could be, for example, the number of firms or survey respondents that are included (Kohli & Deveraj, 2003). In this meta-analysis the sample size is the weight variable in weighted least square (WLS) regression. And coded with the number of observation that is presented in each study.
2.3.2 Method design
This variable is divided in two parts namely, cross sectional or longitudinal. Cross sectional data is easier to obtain because it is often about a one-time event that is researched. Where longitudinal is more resource intensive because time-series are needed to perform this kind of analysis, which costs more time to obtain (Kohli and Deveraj, 2003). Each article is studied, defined as cross sectional or longitudinal. Sometimes, in the text is explained what the method design of the study is. If it is not mentioned, the article is labeled as cross-sectional if it is one-time event, and if it is about multiple events it is labeled as longitudinal. The labels are coded with dummy variables, with cross sectional coded as “1” and longitudinal as “0”.
2.3.3 Duration
This variable is coded as the number of years that the data is collected (Kohli and Deveraj, 2003). In this study the duration is coded as the number that is presented in the article. In each article the number of years is presented, if time was a few months this was coded as 1 year, instead of working with numbers with decimals below zero.
2.3.4 Country location
with North America, Asia, Europe or not available. The last category is added, because many articles did not mention the origin of their data in which country/region they performed the study. For coding this variable Europe is taken as baseline, thus all the results are compared to Europe. Again, for this variable a dummy variable is created. Country 1 (i.e. North America) is coded as a “1” for North America and a “0” for otherwise. Country 2 (i.e. not available) is coded as a “1” for not available and a “0” for otherwise. Country 3 (i.e. Asia) is coded as a “1” for Asia and a “0” for otherwise.
2.3.5 Industry sector
Each articles can differ in industry sector where the research is performed. It matters to take this variable into account, because it could be possible that in some industries stronger effects are identified. In this paper the industry sector is divided in two parts namely, cross industry or in specific industry. For the coding a dummy variable is created with specific industry as baseline. So, Industry 1 (i.e. cross industry) is coded as “1” for cross industry and a “0” for specific industry.
2.3.6 Big Data Analytics
This is one of the main variable of this research. Each articles in the final sample size did conceptualize BDA in a certain way, the conceptualization of each paper differs from each other. The conceptualizations are mentioned in the codebook, based on these conceptualizations three categories were identified to create a more general overview and to bundle them together. The categories are, capability, usage and resource. Based on these categories dummy variables are created. IV 1 (i.e. capability) is coded as “1” for capability and a “0” for otherwise. IV 2 (i.e. usage) is coded as “1” for usage and “0” for otherwise. Where the concept of BDA as resource is taken as baseline.
2.3.7. Value
DV 2 (i.e. innovation performance) is coded as a “1” for innovation performance and a “0” for otherwise.
2.4 Analysis; Weighted Least Square regression (WLS)
Weighted Least Square (WLS) regression is comparable with standard regression, however there are some advantages of WLS regression. WLS benefits from meta-analysis by relying on all the available data to examine a particular phenomenon (Gonzalez-Mulé & Aguinis, 2017). Another advantage of WLS regression is, unlike linear/nonlinear least square regression, that WLS adds weights into the mode-fitting criteria. With adding a weight to each observation, it gives the right amount of impact on the final parameter estimates (Wu & Lederer, 2009). A Weighted Least Square (WLS) regression permits testing of moderating effects or mediating factors (Wu & Lederer, 2009; Combs, et al., 2019).
Table 2 Coding variables
Category Variable Definition Example
Big Data Analytics
1. Capability
Capability refers to the technical as the nontechnical side the three types of resources (i.e. tangible, human IT, intangible).
“This study defines BDA capability as a firm’s ability to assemble, integrate, and deploy its big data-based resources” (Gupta and George, 2016. p.1024)
2. Usage
Usage refers to the actual use of tool and advanced technologies in order to analyze the gathered data.
“As noted previously, BDA usage is defined in the current study as the extent to which organizations use BDA to process information generated across key supply chain processes.” (Chen, et al., 2015. p.20) 3. Resource Resource refers to applications or a platform that an organization can possess in order to
store or gather data.
focuses on strategically exploring investments in future resources and assets.” (Nwankpa and Datta, 2017. p.472) Value 1. Organizational Performance Organizational performance refers to the numerical /measurable tools (e.g. ROE, ROI, sales, profitability) to illustrate value.
“We used return on equity (ROE) as an approximation for firm performance.” (Chi, et al., 2010. p.556) 2. Innovation Performance Innovation performance refers to creating or implementing news ideas for products or services.
“We define ideated innovation as knowledge that is created through firm’s innovation efforts and embodied in forms such as inventions, discoveries, developed ideas, and/or solutions of technical problems.” (Joshi, et al., 2010. p.476) 3. Process Performance Process performance refers to the way the decision making process is designed, with different indicators (e.g. decision quality/efficiency).
“It refers to the firm's ability to sense and quickly respond to changes in the environment which often involves reconfiguring firm resources.” (Ghasemaghaei, et al., 2017. p.96) Industry Sector 1. Cross Industry Cross industry is a combination of two or more industry sectors (e.g. service, manufacturing, non profit).
and communication, scientific services, and business services.” (Dong and Yang, 2018. p.3)
2. Specific Industry
Specific industry contains of one industry sector.
“The survey instrument was intended for a single respondent, who represented the supply chain management organization as the unit of analysis.” (Chen, et al., 2015. p.19)
Country location 1. North America - -
2. Not Available - - 3. Asia - - 4. Europe - - Method design 1. Cross Sectional Cross sectional refers to a study where the data is gathered in a one-time event
“Our OLS results should be interpreted as association rather than causation because of the cross-sectional nature of our data.” (Dong and Yang, 2018. p.5) 2. Longitudinal Longitudinal refers to a study where the data is gathered during a series of events over a longer period of time
“In this regard, we collect longitudinal secondary data about firms’ use of IT for supporting knowledge management initiatives to capture IT- enabled knowledge
Table 3 Coding scheme
Category Coding variables
Big Data Analytics
IV 1 = Capability IV 2 = Usage Baseline = Resource Value DV 1 = Organizational Performance DV 2 = Innovation Performance Baseline = Process Performance
Industry Sector Industry 1 = Cross Industry
Baseline = Specific Industry
Country location
Country 1 = North America Country 2 = Not Available Country 3 = Asia
Baseline = Europe
Method design Method 1 = Cross Sectional
3. RESULTS
In this section the results of the WLS regression will be presented. A stepwise analysis is performed, with three different models. All the models are presented in table 4, the explanation of the different variables are presented in table 2 and 3. The following build up is used for the analysis. In all the models the correlations from the final sample size are used as the dependent variable (correlations can be found in the codebook). The weight variable is the sample size of all the final sample size. The first model consists of method and duration. In the second model country and industry sector are added. The full model consists of the above mentioned variables with the addition of concepts of BDA and Value.
In the first model (methodology) the duration and method are added as independent variables. In this model duration is significant (sig. of 0.008) and has a beta of -0.523. For method no significant relationship is found.
For the second model (research context) the industry sector and country variables are added. In this model none of the variables have a significant relationship, or no significant relationship is found. Despite duration, as in the first model, has a significant relationship.
Table 4 WLS regression results
Model 1 Model 2 Model 3
Coefficients Coefficients Coefficients
Beta Std. Error Beta Std. Error Beta Std. Error
Constant .319 .352 .473 Duration -.523** .005 -.985* .007 -1.511** 0.009 Method1 -.034 .323 -.039 .330 -.024 .314 Industry1 .284 .104 -.142 .132 Country1 -.287 .126 -.352*** .096 Country2 -.310 .063 -.391*** .044 Country3 -.192 .151 -.646*** .145 IV1 .694** .115 IV2 .254 .108 DV1 .802*** .110 DV2 -.081 .193 Sample Size 26 26 26 Adjusted R2 0.217 0.251 0.638 F Change 4.460 1.260 6.092
Dependent variable: Correlation
Table 5 Correlations (1) (2) (3) (4) (5) (6) (7) 1) Correlation . 2) Duration .003 . 3) Method1 .299 .242 . 4) Industry1 .394 .003 .466 . 5) Country1 .295 .033 .061 .394 . 6) Country2 .340 .019 .435 .159 .299 . 7) Country3 .311 .017 .476 .002 .424 .316 . IV1 .000 .009 .099 .288 .215 .056 .358 IV2 .003 .000 .190 .019 .092 .013 .267 DV1 .126 .000 .176 .012 .095 .011 .444 DV2 .498 .006 .072 .000 .397 .060 .323
Table 6 Descriptive statistics
Mean Std. Deviation N Minimum Maximum
4. DISCUSSION
In this section the results of the previous section will be discussed. First the main findings of the WLS regression will discussed and interpreted. After the main findings, the different concepts (i.e. BDA and Value) will be described in more detail. Per paper the different concepts will be described in a more qualitative way, to elaborate more on the different concepts. In the last paragraph of the discussion a general overall view of the conceptualizations will be discussed, bringing all the information together.
4.1 Main findings
As explained in the results section the analysis is divided in three models. The reason that the DV and IV variables (i.e. BDA and value) are added in the last model is because, the variables are the most important in this research. And if the variables are added earlier the results may change over time, instead of one clear result. To gain the best result, a stable result is necessary.
In the first model, duration is the only methodology variable that has a significant relationship. The relationship is negative and this could mean that in the beginning BDA can add value. However, over time the link between BDA and Value decreases for some reason. A reason could be that competitors start to use BDA, or use it more effective or efficient. For the second model no significant relationships are found. So, with adding the research context no relationship can be found between BDA and Value.
significant and has a positive relationship with BDA and value. Meaning that if value is measured, the conceptualization of value in terms of organizational performance, has a stronger relationship than innovation performance and process performance. With measuring innovation and process performance no significant relationship is found.
Concluding on the main findings, BDA conceptualized as capability has a stronger relationship than the conceptualization as usage or resource. For value, if it is conceptualized as organizational performance it has a stronger impact than innovation or process performance. In the following section the variables of BDA and value will be explained in more detail.
4.2 Concept of BDA
In this paragraph the concept of BDA (i.e. capability, resource, usage) will described. The structure of the paragraph is as follows; first the conceptualization of BDA as capability will be described, then as usage and finally as resource. Within each subgroup the individual papers will be described in more detail.
4.2.1 Conceptualization of BDA as Capability
Chen et al. (2015), define in this paper IT integration as the concept of BDA, as part of the overall IT capability of the organization. IT integration is a sub part of IT capability; an organization’s ability to mobilize and deploy IT-based resources in combination with other resources/capabilities. IT integration means combining information from different suppliers, with the goal of helping the partners to exchange information, communicate and establish a collaborative relationship (Chen, et al., 2015).
Chi et al. (2010), describe in their paper that the IT-enabled capability, consists of two indicators, knowledge oriented IT applications and partner scope. BDA is knowledge oriented IT applications in this paper. Knowledge oriented IT applications are systems that are used to connect with external partners, to manage knowledge, mine and interpret data and based on business intelligence (Chi, et al. 2010).
learning. Human recourses are managerial and technical skills. At last, tangible resources, these are data, technology and basic resources. They argue that a BDA capability is achieved through not only the use of digital resources but also other resources (e.g. intangible and human resources) (Gupta & George, 2016).
In this paper Ghasemaghaei et al. (2018) draw on the framework of Gupta and George (2015), that the data analytic competency consists of the three resources (i.e. intangible, tangible and human). Their main BDA construct is the data analytics competency. It is defined as the firm’s ability to deploy and combine data analytics resources for analyses of data. To expand the data analytics competency, they use also use Data quality as construct for BDA. Data quality is the quality of the raw data facts; these facts reflect characteristics of an event or entity. The quality of the data used in analytics could be defined or measured as, for example, reliable, secure, timely (Ghasemaghaei, Ebrahimi & Hassanein, 2018).
In this paper of Joshi et al. (2010) IT-RACAP is the concept of BDA, and specifically knowledge transformation capability. IT-enabled realized absorptive capacity (IT-RACAP), is realized through the use of information technologies that supports knowledge transformation and exploitation. Knowledge transformation capability is enabled through IT; this capability enhances innovations within an industry or helps to create new knowledge by merging, reclassifying or categorizing existing knowledge. Different information technologies can help or improve the firm’s knowledge transformation capability, like data mining and analytics tools (Joshi, et al. 2010).
4.2.2 Conceptualization of BDA as Usage
Dong and Yang (2018) conceptualize BDA as the use of BDA, that makes it possible to analyze data effectively and efficiently, for deriving insights on customers. They specify the concept of BDA in combination with the use of social media diversity (Dong & Yang, 2018).
and prescriptive. Data analytics itself are the technologies and processes that support data mining (Ghasemaghaei, Hassanein & Turel, 2017).
They, Chen et al. (2015), define big data analytics as the concept that uses advanced technologies in order to process or analyze big data. In this paper the overall conceptualization of BDA is usage, and defined as follows; the extent to which organizations use BDA to process information that is collected from key supply chain processes. In their study they used different drivers (i.e. expected benefits, technology compatibility, organizational readiness, competitive pressure and Top Management Support) as the explanatory drivers for BDA usage (Chen, Preston & Swink, 2015).
Business Intelligence (BI) success is the relationship between BI investments and the value the organization obtains from the investment. Isik et al. (2013), use data quality as the conceptualization of BDA usage. Data quality is the consistency and comprehensiveness of data; this can be qualitative data or quantitative data. If the information is not accurate or consistent, organizations usage is or can be poor. Nowadays, organizations have to deal higher data volumes, velocity and variety (i.e. big data). The big data the organization collets can be internal or external (Isik, Jones & Sidorova, 2013).
In this study of Song et al. (2018), the usage of data analytics is divided in the demand-side and the supply-side data analytics. The demand-side data analytics is defined as; to see buyers’ purchases and interactions, and listen to the needs and want of customers, and spends a lot of time to use data analytics to map customer’s needs. Supply-side data analytics is to see the internal operational activities and improve efficiencies, use data analytics to examine the quality of logistic services, and spend a lot of time using data analytics to generate knowledge for improving supply-side operation (Song, et al., 2018).
Besides the conceptualization of BDA as capability, Ghasemaghaei et al. (2018) also define BDA as the construct of usage. Analytical skills, like data quality, is a part to expand the data analytics competency. They define analytical skills as necessary for generating new business insights, that lead to improved decision making performance. The skills refer to the competency of using big data analytics, that the users have sufficient knowledge, skills and expertise to use analytics (Ghasemaghaei, Ebrahimi & Hassanein, 2018).
4.2.2 Conceptualization of BDA as Resource
Nwankpa and Datta (2017) define BDA as a resource, namely as Digital Business Intensity (DBI), where DBI focuses on exploring new investments for future resources or assets. These opportunities can be integrated or adopted in the organization. BDI consists of different technologies (e.g. analytics, big data, cloud) that are used in business transactions, firm operations or business operations. Besides, DBI is also a measure how much an organization has invested in digital innovations (Nwankpa & Datta, 2017).
In this paper of Quaadgras et al. (2014), BDA conceptualized as resources is defined as ‘working smarter with information’. Creating a digital platform can reduce costs, however, deriving a competitive advantage from the platform can be a struggle. Working smarter with technology will help decision makers to gather information and create clear business rules. Likewise, with using analytics revising these business rules or creating new ones. Leading to the digital platform, or the business rules as resource (Quaadgras, Weill & Ross, 2014).
Tarafdar and Tanriverdi (2018) define IT Unit’s standardization support as the concept of BDA as resource. IT Unit’s Standardization Support is the extent to which the IT organization identifies and maintains appropriate technology standards for the emerging information technologies in the organization. Even with different systems from different vendors, the systems should be linked with each other (Tarafdar & Tanriverdi, 2018).
4.3 Concept of value
4.3.1 Conceptualization of value as organizational performance
Chi et al. (2010) describe the variable of value as organizational performance. Because the concept of value is defined as firm performance, and conceptualized as return on equity (ROE). IT interacts with the network that shapes competitive action of the firm. Competitive action is linked to firm performance, however, the strategic use of IT is a better indicator to explain differences in firm performance (Chi, et al., 2010).
Gupta and George (2016) divide the concept of value in two parts namely, operational performance and market performance (market performance is classified as process performance. And operational performance is classified as organizational performance). Operational performance is about productivity, profit, return on investment, sales revenue. Market performance is related to entering new markets, introducing new products/services, market share. The study shows a clear relationship between big data analytics competency and firm performance (i.e. operational and market performance). Furthermore, this study expands the theoretical framework of BDA capability with technical and nontechnical resources across the three categories (Gupta & George, 2016).
The variable value is in the paper of Dong and Yang (2018) classified as organizational performance. Because in this paper value is conceptualized as firm performance, which is defined as the sales in the marketplace. They argue that the combination of social media diversity and the use of big data analytics has a positive effect on firm performance (Dong & Yang, 2018).
In this study Isik et al. (2013) define value as BI success (organizational performance). BI success leads, for example, to increased profitability, reduced costs, increased efficiency. The relationship between data quality and BI success is negatively related. An explanation is that the quality of the data is ‘good enough’ to work with, and improving the quality will come at the expense of other BI capabilities. Data quality is necessary but not sufficient to achieve BI success (Isik, Jones & Sidorova, 2013).
information of consumers to make better decisions (e.g. new products, services, innovations in customer experience). This all leads towards a high organizational performance (Song, et al., 2018).
Luftman et al. (2017) define value as company performance, this is divided in two concepts to measure performance, namely Return On Assets (ROA) and Return On Equity (ROE). Based on this definition, value is classified as organizational performance. This study confirms that different elements (e.g. value analytics, communications) contributes to IT-business alignment, which has a moderate positive effect on company performance (Luftman, Lyytinen & Zvi, 2017).
Organizational performance is a measure of how the organization reach its objectives and goals, this is characterized with profitability, sales growth and Return on Investment (ROI). Therefore, in the paper of Nwankpa and Datta (2017), value is classified as organizational performance. This study shows that organizational performance is not only related towards exploiting existing IT resources and assets. Instead, it should extend the existing IT resources and assets with an exploration through DBI. DBI is a mediator and a moderator of the effect of IT capabilities on company performance. It should be aligned with IT capabilities to gain the greatest advantage (Nwankpa & Datta, 2017).
In this study Quaadgras et al. (2014), financial performance is the conceptualization of value, thus it can be classified as organizational performance. The financial performance is measured through industry adjusted Return on Equity (ROE). Working smarter with information contributes to the concept of business impact of IT, which has a positive influence on the financial performance of the organization (Quaadgras, Weill & Ross, 2014).
4.3.2 Conceptualization of value as innovation performance
In this paper, Chen et al. (2015) define value as product innovation performance. Meaning that the value construct is classified as innovation performance. Creating and implementing new ideas is the basis for innovation. They argue that IT-capabilities should not be linked directly to firm performance (i.e. financial outcomes, profit), instead it should be linked to innovation performance. Their results show that IT-capabilities enable product innovation, through the use of business intelligence or analytics technologies (Chen, et al., 2015).
Joshi et al. (2010) discusses in his paper that ideated innovation is knowledge that an organization gains through innovation efforts, such as inventions, discoveries, solutions to technical problems. Therefore, value can be classified as innovation performance. These innovations can lead to new products or services, which lead to organizational profit. The study finds a relationship between IT-RACAP and ideated innovation, the results suggest that firms who use knowledge transformation through IT, are better capable of creating new knowledge that can be used for new products or services (Joshi, et al., 2010).
In this paper Chen et al. (2015) use business growth for defining value (i.e. innovation performance). Business growth is about sales growth, market expansion (entering new markets) and market share growth. It is defined as the capability of creating new temporary advantages. The authors find that BDA use, as an information processing capability, has strong influence on business growth (Chen, Preston & Swink, 2015).
4.3.3 Conceptualization of value as process performance
In the paper of Ghasemaghaei et al. (2017), they use agility to define the variable value (i.e. process performance). Agility is the ability of the organization to respond to threats and opportunities in the business environment. The concept agility is divided in operational adjustment agility and market capitalizing agility. The mere use of big data analytics will not generate organizational agility. The moderating effects of analytical tools, data, people and tasks, should be taken into account. These effects increase big data analytics use on agility (Ghasemaghaei, Hassanein & Turel, 2017).
4.4 Combining BDA and value; an overall analysis
In this paragraph an overall analysis will be discussed about the findings of the WLS regression and the input from the articles from the final sample size will also be used. From the findings the conceptualization of BDA as capability and value as organizational performance were significant. Meaning that when an organization uses BDA as capability it has a positive effect on value creation (i.e. organizational performance).
Chen et al (2015) used IT integration as part of the IT capability, and Chi et al (2010) did use knowledge oriented IT applications as part of IT-enabled capability. Joshi et al (2010) argue that knowledge transformation capability is part of BDA, and that IT enables this capability. Gupta & George (2016) argue that the BDA capability consists of three types of resources (i.e. tangible, intangible and human), arguing that BDA capability is more than the technological side. Ghasemaghaei et al (2018), used the framework of Gupta and George. However, they added data quality meaning that data quality matters for explaining the characteristics of an event or an entity.
Taking these articles and their conceptualization in mind, many papers are closely related to the work of Bharadwaj (2000). She argues that possessing IT capabilities leads to a higher organizational performance. Bharadwaj (2000) categorize IT resources based on the work of Grant (1991). Leading to the following resources; (1) tangible resource (e.g. IT infrastructure), (2) human IT resources (e.g. technical and managerial skills), (3) intangible IT-enabled resource (e.g. knowledge assets, customer orientation). Looking at how the articles conceptualize BDA capability in their papers, it has similarities with the notion of IT capability in Bharadwaj (2000).
sustainable competitive advantage. By possessing IT capabilities firm’s revenue will increase and/or cost will decrease (Bharadwaj, 2000; Kristandl and Bontis, 2007). According to these authors, possessing a capability will lead to a sustainable competitive advantage. When BDA is conceptualized as an IT capability, therefore, it is valuable, rare, and leads to sustainable competitive advantage with larger value. When BDA is conceptualized as usage only or resource, it is still valuable and probably rare, but not imperfectly imitable and non-substitutable. Therefore, studies found a weaker relationship between BDA and value.
From the WLS regression two moderators are significant, namely duration and country location. Duration makes the relationship between BDA and value weaker, meaning that the value of BDA decreases over time. This could be, for example, that competitors also invest in and use BDA and that the competitive advantage is not sustainable. Taking this finding together with conceptualization of BDA, it is important for research on BDA and value look into BDA capability and examine its impact on sustainable competitive advantage.
5. IMPLICATIONS AND LIMITATIONS
In this section of the study the following subjects will be discussed. First, the theoretical implications and directions for future research will be discussed. Second the practical implications of this paper will presented, and ending this section with the limitations of this paper.
5.1 Theoretical implications and future research
In this paper, an overall overview is presented of the current literature about BDA and value creation. In the field of BDA and value creation, until now, a meta-analysis was never performed to make sense of the current status of the literature. This study makes several contributions.
First, the main finding of this study is that when BDA is conceptualized as a capability it has a positive effect on value creation. Meaning that the other two conceptualizations were not significant (i.e. usage and resource). This finding is in line with the work of Bharadwaj (2000), he argues that capabilities are hard to imitate, and lead to a competitive advantage for the firm, based on the VRIN characteristics. Value creation has a significant effect when it is conceptualized as organization performance, this is also in line with the work of Bharadwaj (2000). In his study firm performance is defined in terms of revenue and costs, hard measurable terms. Like in this study, organizational performance is often conceptualized as ROE, ROI, profitability etc. Hard measureable items, that can be expressed in a quantitative way.
Based on this implications, future research could perform research about the concepts of capability and organizational performance. In this paper the conceptualization of BDA as capability is based on the final sample size of 14 papers, with a few papers that use BDA as capability. Based on the findings of this study BDA as capability is important for value creation. Thus, future research could focus on how to conceptualize BDA as capability, how to define a BDA capability and also a practical viewpoint to creating or acquire a capability.
This study confirms some current finding on capabilities and organizational performance. However, in the field of BDA it is a new contribution.
As with the first implication future research could focus more on the conceptualization of BDA as capability, and of value creation as organizational performance.
Third, the relationship between BDA and value creation has one steady moderator and one moderator that is only presented in the full model. The first moderator is duration, which could implicate that having a capability last for a certain amount of time. The second moderator is the country location, benefiting full from BDA as capability the region where the firm is located could matters.
In this study country location and duration are moderators on the relationship between BDA and value creation. Future research could take these two moderators into account when performing their research. This study does not give a clear explanation of what these moderators contain and in which circumstances they become more relevant. Also for duration, gathering data over a longer period of time could also help explaining why the value of BDA decreases over time. For country location, future research could look deeper into BDA as capability in different regions, and focusing why in Europe they benefit more from BDA than other regions.
5.2 Practical implications
This study has also some practical implications for managers and practitioners. In order to create a capability managers need to focus on the three resource categories namely, tangible resources, human IT resources and intangible IT-enabled resource. These three resources are key in creating a capability for the firm. First the organization needs to have an effective and efficient IT-infrastructure that is reliable for using different kind of systems, and linking different departments with each other. Secondly, managers need to have the right people in their organization, meaning that the organization is in possession of the right technical and managerial skills to use the IT infrastructure, and to improve it for the future. At last, managers need to have a customer oriented focus, taking this perspective the organization tries to relate closely to the customer needs of the organization. Managers need to make use the needs of customers in their internal processes.
weakness, and then identify organizations that are leaders on the weaker activities or function. Based on the finding from the benchmark research, the organization could improve the weaker activities and functions of the organization.
5.3 Limitations
6. CONCLUSION
Appendix A
Codebook
Paper (Name, Journal, Year)
Big Data Analytics Value Correlation Sig. Reliability Industry sector Country location
Duration Sample Size Method design
Nwankpa-Datta, European Journal of Information Systems, 2017 Digital Business Intensity Organizational performance
0.47 0.01 N/A Cross Industry USA 1 315 Cross sectional
Chi, Information Systems Research,
2010
IT-enabled
capability performance - Firm ROE
0.156 0.05 N/A Automobile
Industry USA 16 153 Cross sectional
Joshi, Information Systems Research,
2010
IT-RACAP Ideated
Innovation 0.427 0.01 N/A Cross Industry USA 1 110 Longitudinal
Quaadgras, Journal of Information Technology, 2014
Working Smarter
with information performance Financial 0.29 0.05 N/A Cross Industry N/A 1 210 Cross sectional
Luftman, Journal of Information Technology, 2017
Value Analytics Company
performance 0.112 0.205 N/A Cross Industry N/A 13 3029 Cross sectional
Chen, Journal of Management
Information Systems, 2015
Technological;
Appendix A Chen, Journal of Management Information Systems, 2015 Technological; Technology Compability
Business Growth 0.321 0.01 0.857 Suply Chain N/A 1 161 Cross Sectional
Chen, Journal of Management
Information Systems, 2015
Top Management
Support Business Growth 0.264 0.01 0.964 Suply Chain N/A 1 161 Cross Sectional
Chen, Journal of Management Information Systems, 2015 Organizational; organizational readiness
Business Growth 0.232 0.01 0.847 Suply Chain N/A 1 161 Cross Sectional
Chen, Journal of Management Information Systems, 2015 Environmental; competitive pressure
Business Growth 0.077 0.01 0.862 Suply Chain N/A 1 161 Cross Sectional
Chen, Journal of Management
Information Systems, 2015
Big data analytics
use Business Growth 0.301 0.01 0.945 Suply Chain N/A 1 161 Cross Sectional
Ghasemaghaei, Journal of Strategic Information Systems, 2018 Data analytics competency Decision making performance
0.883 0.001 N/A Cross Industry N/A 1 151 Cross Sectional
Ghasemaghaei, Journal of
Strategic Information Systems, 2018
Appendix A Ghasemaghaei, Journal of Strategic Information Systems, 2018
Analyitcal skills Decision Quality 0.166 0.05 N/A Cross Industry N/A 1 151 Cross Sectional
Ghasemaghaei, Journal of
Strategic Information Systems, 2018
Data Quality Decision
Efficiency 0.202 0.05 N/A Cross Industry N/A 1 151 Cross Sectional
Ghasemaghaei, Journal of
Strategic Information Systems, 2018
Analyitcal skills Decision
Efficiency 0.262 0.01 N/A Cross Industry N/A 1 151 Cross Sectional
Ghasemaghaei , Decision Support
Systems, 2017
Data Analytics Use Adjustment
Agility 0.17 0.01 N/A Cross Industry America North 1 215 Cross Sectional
Ghasemaghaei , Decision Support
Systems, 2017
Data Analytics Use Capitalizing
Agility 0.16 0.01 N/A Cross Industry America North 1 215 Cross Sectional
Chen, Information & management, 2015 IT integration Product innovation performance
0.11 0.88 Manufacturing North China 2 138 Cross Sectional
Gupta, Information & management,
2016
Big Data Analytic capability
Market performance
Appendix A Gupta,
Information & management,
2016
Big Data Analytic
capability performance Operational 0.67 0.001 N/A Cross Industry N/A 1 340 Cross Sectional
Isik, Information & management,
2013
Data Quality BI Success 0.356 0.1 N/A Cross Industry USA 1 92 Cross Sectional
Song, Information & management,
2018
Demand side data
analytics Performance 0.389 0.01 N/A B2C platform China 1 309 Cross Sectional
Song, Information & management,
2018
Supply side data
analytics Performance 0.324 0.01 N/A B2C platform China 1 309 Cross Sectional
Dong, Information & Management,
2018
Big data analytics Market
Performance 0.219 0.001 N/A Cross Industry Italy 13 18816 Cross Sectional
Tarafdar, Journal of the Association for Information Systems, 2018 IT Unit's technology standardization support Organizational Performance
REFERENCES
Agarwal, R., & Dhar, V. (2014). Big data, data science, and analytics: The opportunity and challenge for IS research.
Ahmed, W., & Ameen, K. (2017). Defining big data and measuring its associated trends in the field of information and library management. Library Hi Tech News, 34(9), 21-24.
Baesens, B., Bapna, R., Marsden, J. R., Vanthienen, J., & Zhao, J. L. (2016). TRANSFORMATIONAL ISSUES OF BIG DATA AND ANALYTICS IN NETWORKED BUSINESS. MIS quarterly, 40(4).
Bangert-Drowns, R. L. (1986). Review of developments in meta-analytic method. Psychological Bulletin, 99(3), 388.
Bharadwaj, A. S. (2000). A resource-based perspective on information technology capability and firm performance: an empirical investigation. MIS Quarterly, 169-196.
Bumblauskas, D., Nold, H., Bumblauskas, P., & Igou, A. (2017). Big data analytics: transforming data to action. Business Process Management Journal, 23(3), 703-720.
Chen, D. Q., Preston, D. S., & Swink, M. (2015). How the use of big data analytics affects value creation in supply chain management. Journal of Management Information Systems, 32(4), 4-39.
Chen, H., Chiang, R. H., & Storey, V. C. (2012). Business intelligence and analytics: from big data to big impact. MIS quarterly, 1165-1188.
Chen, Y., Wang, Y., Nevo, S., Benitez-Amado, J., & Kou, G. (2015). IT capabilities and product innovation performance: The roles of corporate entrepreneurship and competitive intensity. Information & Management, 52(6), 643-657.
Clarke, R. (2016). Big data, big risks. Information Systems Journal, 26(1), 77-90.
Combs, J. G., Crook, T. R., & Rauch, A. (2019). Meta‐analytic research in management: contemporary approaches, unresolved controversies, and rising standards. Journal of Management Studies, 56(1), 1-18.
Davenport, T. H., Barth, P., & Bean, R. (2012). How big data is different. MIT Sloan Management Review, 54(1), 43.
De Mauro, A., Greco, M., & Grimaldi, M. (2016). A formal definition of Big Data based on its essential features. Library Review, 65(3), 122-135.
Dong, J. Q., & Yang, C. H. (2018). Business value of big data analytics: A systems-theoretic approach and empirical test. Information & Management.
Erevelles, S., Fukawa, N., & Swayne, L. (2016). Big Data consumer analytics and the transformation of marketing. Journal of Business Research, 69(2), 897-904.
Fichman, R. G., Dos Santos, B. L., & Zheng, Z. E. (2014). Digital innovation as a fundamental and powerful concept in the information systems curriculum. MIS quarterly, 38(2).
Ghasemaghaei, M., Ebrahimi, S., & Hassanein, K. (2018). Data analytics competency for improving firm decision making performance. The Journal of Strategic Information Systems, 27(1), 101-113.
Ghasemaghaei, M., Hassanein, K., & Turel, O. (2017). Increasing firm agility through the use of data analytics: The role of fit. Decision Support Systems, 101, 95-105.
Glass, G. V., B. McGaw, M. L. Smith. 1981. Meta-analysis in Social Research. Sage, Beverly Hills, CA.
Grant, R. M. (1991). “The Resource-based Theory of Competitive Advantage: Implications for Strategy Formulation.” California Management Review 33 (3): 114– 135.
Günther, W. A., Mehrizi, M. H. R., Huysman, M., & Feldberg, F. (2017). Debating big data: A literature review on realizing value from big data. The Journal of Strategic Information Systems.
Günther, W. A., Mehrizi, M. H. R., Huysman, M., & Feldberg, F. (2017). Debating big data: A literature review on realizing value from big data. The Journal of Strategic Information Systems, 26(3), 191-209.
Gupta, M., & George, J. F. (2016). Toward the development of a big data analytics capability. Information & Management, 53(8), 1049-1064.
IşıK, Ö., Jones, M. C., & Sidorova, A. (2013). Business intelligence success: The roles of BI capabilities and decision environments. Information & Management, 50(1), 13-23.
Joshi, K. D., Chi, L., Datta, A., & Han, S. (2010). Changing the competitive landscape: Continuous innovation through IT-enabled knowledge capabilities. Information Systems Research, 21(3), 472-495.
Kristandl, G., & Bontis, N. (2007). Constructing a definition for intangibles using the resource based view of the firm. Management decision, 45(9), 1510-1524.
Laney, D. (2001). 3D data management: Controlling data volume, velocity and variety. META Group Research Note, 6(70).
Luftman, J., Lyytinen, K., & ben Zvi, T. (2017). Enhancing the measurement of information technology (IT) business alignment and its influence on company performance. Journal of Information Technology, 32(1), 26-46.
Manyika, J., Chui, M., Brown, B., Bughin, J., Dobbs, R., Roxburgh, C. and Byers, A.H. (2011), Big Data: The Next Frontier for Innovation, Competition, and Productivity, McKinsey Global Institute, New York, NY
McAfee, A., Brynjolfsson, E., Davenport, T. H., Patil, D. J., & Barton, D. (2012). Big data: the management revolution. Harvard business review, 90(10), 60-68.
Nwankpa, J. K., & Datta, P. (2017). Balancing exploration and exploitation of IT resources: the influence of Digital Business Intensity on perceived organizational performance. European Journal of Information Systems, 26(5), 469-488.
Quaadgras, A., Weill, P., & Ross, J. W. (2014). Management commitments that maximize business impact from IT. Journal of Information Technology, 29(2), 114-127.
Sharma, R., Mithas, S., & Kankanhalli, A. (2014). Transforming decision-making processes: a research agenda for understanding the impact of business analytics on organizations. European Journal of Information Systems, 23(4), 433-441.
Song, P., Zheng, C., Zhang, C., & Yu, X. (2018). Data analytics and firm performance: An empirical study in an online B2C platform. Information & Management.
Tarafdar, M., & Tanriverdi, H. (2018). Impact of the Information Technology Unit on Information Technology-Embedded Product Innovation. Journal of the Association for Information Systems, 19(8)
Wamba, S. F., Akter, S., Edwards, A., Chopin, G., & Gnanzou, D. (2015). How ‘big data’can make big impact: Findings from a systematic review and a longitudinal case study. International Journal of Production Economics, 165, 234-246.
Wamba, S. F., Gunasekaran, A., Akter, S., Ren, S. J. F., Dubey, R., & Childe, S. J. (2017). Big data analytics and firm performance: Effects of dynamic capabilities. Journal of Business Research, 70, 356-365.
