Association of Big Data with Sustainable Competitive Advantage in Online Retail Segment: A Serial Mediation Model for Relating Big Data with Strategic Management Tools

ARTICLE DETAILS ABSTRACT


Introduction
In recent times organizations are thrilled to use digital platforms for enhancement in business and sales.These optimizations in these business activities and operations might become more effective with the incorporation of Big-Data analytics.The use of Big-Data might also result in business intelligence innovation which can provide an edge to the company through strategic innovativeness of decision-making (Jayakrishnan et al., 2018).Businesses are used to extract required information and knowledge from larger data sets.Similar is the purpose of Big-Data analytics i.e., to process larger data sets through enhanced analytical techniques for extracting useful knowledge.Hence it is treated as a potential source to achieve a competitive edge (Dahiya et al., 2021).
Although some of the studies argued that the benefits of applying Big Data analytics are conditional and there is some evidence that firms have failed to attain a competitive advantage through the use of Big Data.The failure might happen due to technical issues rather than strategic roles and requirements associated with the implementation of Big-Data technology.Hence, it is optimal to reflect that managerial and organizational competency must support practitioners to attain the desired level of competitive edge (Ramadan et al., 2020).A study by Dahiya et al (2021) indicated that the Resource-Based (RB) Theory is the base by which it is legitimate to believe that firm's specific knowledge might result in the attainment of competitive advantage.
Therefore, in the light of RB Theory, it is optimal to reflect Big-Data models in association with advanced level firm's knowledge are essential to attain a sustainable competitive edge.Hence research must consider elements used by Ramadan et al (2020), to reflect the flow required for knowledge creation specifically for different types of firms.Inclusions of these elements are also aligned with the Knowledge-Based View (KBV) which is the extension of RB-Theory and is reflected by Dahiya et al. (2021).

Literature Review
There is a severe lack of research work that is associated with the research work in the area of Big-Data Analytics and its association with business value (Mikalef et al., 2017).Dahiya et al (2021) posited that several firms tried to opt for Big-Data analytics to have a competitive edge.However, there is still a requirement for further research work to explore the relationship between knowledge attained through Big-Data Analytics & competitive advantage.The major reason behind these sorts of requirements is the success ratio i.e., there is no guarantee of success even after using Big-Data analytics.These conditions are surrounded by a severe level of ambiguity due to dealing with enormous amounts of data (Jayakrishnan et al., 2018).The complexity enhances significantly when most of the research on big data is not associated with the capacity of the building of a firm.Hence, it is optimal to quote Dahiya et al (2021) and Ramadan et al (2020) to relate the use of Big-Data analytics with the increase of a firm's specific and differentiated knowledge.Therefore, this study will be focusing significantly on the relationship of Big-Data with the Big-Data knowledge that may affect the innovative capabilities of the firm and business improvement.Sivarajah et al. (2017) indicated that Big-Data and Big-Data analytics are still in their infancy for research.Therefore, there is no unified & fully established definition and classification associated with these terminologies.On the other hand, most of the studies related to Big-Data are related to technical aspects rather than implications of Big-Data in strategic management.Therefore, studies based on the relationship between Big-Data and strategic management must be conducted in a multi-disciplinary context to overcome the lack of association between Big-Data and firm performance.Therefore, this study uses the framework of Dahiya et al (2021) and Ramadan et al (2020) to formulate a unique research model based on serial mediation.The model is especially significant in relating Big-Data with sustainable innovation and a firm's performance through the serial mediation model to generate sustainable competitive advantage (SCA).
Decisions that are based on data are termed data-driven decisions, to make these decisions there is also a need for data analysis.These decisions are found to play a positive role for the company through optimizing measures e.g., returns on assets, returns on equity, market value & asset utilization, etc. (Jelonek, Stępniak & Ziora, 2018).Although, different forms of business collect information from different forms of sources to create consumer perfect e-picture through accessing browsing and purchasing information (Lambrecht & Tucker, 2015& Lambrecht & Tucker, 2016).
Thus, data is termed as one of the most effective tools for marching progressively towards better decisions, especially in the field of e-commerce (Seetharaman et al., 2016).Although research in this vein is still at the initial stage (Yang et al., 2019) and companies are also not required to focus on big data until and unless they are required to understand proactively their customers, markets, and results of products (Charles & Gherman, 2013).Similar was indicated by McAfee and Brynjolfsson (2012) that companies that are in the list of top three are more inclined towards a data-driven approach and will attain more productivity, sales, and profit as compared to the rest.In recent times there is no company that is striving for the achievement of relevant data to gain a competitive advantage.However, a competitive advantage cannot be achieved without the extraction of value out of raw data (Charles & Gherman, 2013).
On the other hand, competitive advantage is the outcome of resources that are unique and inimitable and this is also the major assumption behind the IT capabilities.Thus, the IT department is required to apply full of their capabilities to induce and reshape organizational strategies for taking competitive advantage.That may lead to an increase in organizational performance (Wamba et al., 2017).Similar research has also been supported by Resource-Based Theory as firms are required to generate more economic benefits as compared to the competing firms through the effective definition of internal and external resources to use them strategically (Gupta et al., 2018).
There are three major types of IT capabilities i.e., value capability, competitive capability & dynamic capability.These three forms of capabilities are based upon the quality of infrastructure, quality of IT business expertise & intensity of organizational learning.Therefore, by combining all three forms of IT capabilities firms are trying to gain an advantage through (BDAC) Big Data Analytic Capabilities (Wamba et al., 2017).This may include reduction in cost, improvement in the firm's offerings, and optimization of decision making.Although there is a severe lack of studies that may explore the benefits of big-data implementation (Huang, Wang & Huang, 2020).
Therefore, the availability of data & data analytics is ranked among the most critical issues, and hence there is also a need to connect all the devices e.g., systems, devices, and sensors, etc., with the internet of things (IoT), to generate data adequately and to assist Big-Data analytics (Ramadan et al., 2020).These practices are effective in producing Big-Data that is enormously large as compared to the other forms of data (Kumar, Kamesh & Syed, 2014).On the other side, businesses are also required to focus on issues related to the environment and sustainability while carrying out operations and practices (Ramadan et al., 2020).In fact, studies indicated the major benefit of Big-Data is to extract knowledge from conglomerate data sets to establish actionable results that may aid in attaining competitive advantage (Ferraris et al., 2018).Though some firms still remain unable to achieve the desired edge through Big-Data (Dahiya et al., 2021).
Therefore, in addition to the data availability and link between Big-Data analytics and sustainability must also be treated as an emerging issue.A sustainable competitive advantage might be defined as the blend of a firm's assets, capabilities, and features that is not easy to copy and replicate by competitors.However, to achieve that, the company has to use its capacity in a fully efficient manner in order to gain valuable insights through Big-Data analytics capabilities.(Ramadan et al., 2020).BDAC also has the capacity to create sustainable value to assist the progress of business but this also needs proper alignment of strategy and technology (Akter et al., 2016) In fact, Big Data Predictive Analysis is perceived as the next big frontier to gauge firm's level of innovation, competitiveness, and productivity (Gupta et al., 2018).
Consumer footprints are the most useful source for the tracing of individual-level data and these external sources like opinions and interests etc. are significantly important for gaining useful insights regarding individual consumers (Lambrecht & Tucker, 2015& Lambrecht & Tucker,2016).These useful insights might be beneficial for managerial decision-making at every level of management and can also aid management at different stages of decision making like identification of the problem, determining the purpose, evaluation of options, forecasting of consequences, selecting the best stage, and improving the sensitivity analysis (Jelonek et al., 2018).Although the increase in volume and velocity of data will also foster the need to develop proper means for gathering, analyzing, and interpreting the data (Davenport et al., 2012).In addition, volume and other characteristics of Big Data are also found to be beneficial in creating new algorithms and also enable IT practitioners, to use databases in a way that was not possible previously.However, there is also a need for a sufficient amount of reliable data and reasonable precision to make stories that are valid and describe consumer decision processes so as to take effective decisions on the bases of data (Charles & Gherman, 2013).Therefore, the use of Big-Data with massive volume, velocity, variety, and Veracity was found to be a much more effective tool for achieving competitive advantage (Charles & Gherman, 2013).
Consistent with the study by Columbus (2014), 85% of the firms believe that within three years, BDA will reshare the competitive landscape of all the operating industries.Not only this but around 90% of firms believe that the adoption of Big-Data is a must and firms that fail to adopt Big-Data till the next year will surely the momentum and market worth (Akter et al., 2016).

Figure.1. Theoretical Framework Source: Author's own elaboration
H1a: There is a relationship between data availability and advanced algorithms H2a: There is a relationship between advanced algorithms and Big-Data Analytic H3a: There is a relationship between Big-Data Analytic and Big-Data Knowledge H4a: There is a relationship between Big-Data Knowledge and Innovative Capabilities

Research Methodology
The study is based upon the theory i.e., Resource-Based (RB) Theory which has also been used by Dahiya et al. (2021), therefore the research is based on a deductive approach (Saunders, Lewis & Thornhill, 2009& Sekaran & Bougie, 2016).Similarly, a study has been done on the indication by Dahiya et al (2021); Ramadan et al (2020), and Yang et al. (2019) to boost knowledge related to the relationship of Big-Data with the sustainable innovation and a firm's performance.Therefore, in accordance with Saunders et al. (2009), the research philosophy to this study is epistemology, as it is the philosophy of knowledge.
Thus, there is a definite need to relate epistemology with this study as there is a lack of studies related to the Big-Data, especially with respect to strategic management etc.As highlighted by Mikalef et al. (2017) andSivarajah et al. (2017).Hence considering all these points, issues, and problems, especially the lack of quantitative studies, this study is related to post-positivism as the philosophical stance as it is effective for both qualitative and quantitative studies (Saunders et al., 2009(Saunders et al., & Žukauskas et al., 2018)).The study is based on primary data as Ramadan et al. (2020) and a questionnaire was used to collect data which is a hybrid of Ramadan et al. (2020); Le and Liaw (2017), through noncontrived settings with individuals as the unit of analysis (Sekaran & Bougie, 2016).
The study follows the sampling type incorporated by Ramadan et al (2020) and Jabeen et al (2021) to incorporate non-probability sampling through quota sampling.The data has been collected from IT experts associated with online businesses in Pakistan therefore the sample size is relatively low as it has been used in quantitative research.However, Ramadan et al (2020) use Power software to calculate sample size in order to assure adequacy and applicability and similar incorporations indicated minimum sample size for the study must be around 135.

Data Analysis
Most of the time research work in the field of marketing is found to be using Structural Equation Modeling (SEM), which is the second-generation multivariate data analysis model.The model is found to be a better choice for analyzing theoretically supportive additive and causal models.In fact, business research becomes much simpler with the ability of the technique to measure unobservable and latent variables.
The technique is based upon exogenous and endogenous variables with the latter being the one that is under the effect of the other.Although there are various methods to use SEM like AMOS, M-PLUS, LISREL, etc., for smaller sample sizes researchers have little knowledge regarding the parameters related to research PLS-SEM (Wong, 2013).Therefore, this study is lured by the use of SEM through SMART-PLS not only because of its association with strategic management.The study also has a relatively small sample size as compared to the other forms of studies and the model of the study is also much more complex about which there is a lack of conclusive studies, especially in the context of strategic management (Sivarajah et al., 2017).

Figure.1. Conceptual Framework Source: Author's own elaboration
Table 1 is indicating that the outer loading for all the elements (indicators) is higher than 0.708 which is the minimal requirement for including any of the elements in the research model (Hair, Sarstedt, Ringle& Mena, 2012).Similar has been posited by Wong (2016), though values higher than the threshold value are deemed more satisfactory and desirable for the purpose of analysis (Khan, Sarstedt, Shiau, Hair, Ringle & Fritze, 2019).

Source: Author's own elaboration
Table 2 is indicating that the value of R 2 is more than 0.50 for all of the cases and these values are equivalent or higher than the moderate level criteria of predictive accuracy (Cheah, Memon, Chuah, Ting & Ramayah, 2018).Therefore, in the light of these parameters, the variance caused by IV in subsequent DVs is effective (Benitez, Henseler, Castillo & Schuberth, 2020).

Source: Author's own elaboration
Table 4 is indicating discriminant validity through HTMT, i.e., the most preferred tool to reflect the discriminant validity (Ab Hamid et al., 2017).Although there is a requirement of values of 0.85 or less in order to assure the criterion (Hair Jr. et al., 2017).Hence the discriminant validity has also been assured for the analysis and the model is legitimate to proceed for inferential statistics.Source: Author's own elaboration

Figure.2. Path Coefficient Source: Author's own elaboration
Table 5 indicated the path coefficient and this has been reflected through two parameters i.e., t-values and p-values.In order to have an impact, there is a need for a t-value of more than 1.97 and a p-value that is less than or equal to 0.05 (Hair Jr. et al., 2021& Wong, 2013).Therefore, according to these parameters, all the relationships are effective i.e., exogenous created an impact over endogenous.

Source: Author's own elaboration
Table 6 is also showing Path-Analysis and for this, the criteria of p-values and t-values are the same as indicated in table 5.However, when researchers analyze serial mediation then it has been noticed that two of the relationships at the bottom of the table are insignificant.The analysis has been consistent with the criterion given by Hair Jr. et al. (2021) as the t-values are lesser than 1.97 in both the cases and the p-values are also greater than 0.05 Analysis of data through descriptive and inferential means indicated that the data is reliable and effective to use for the analysis of impact created by the bigdata on online retail segment and willingness to take sustainable competitive advantage.

Discussion and Managerial Implications
The findings of the study are fruitful as the findings are effective to reflect the linkage of Big-Data with strategic management as well as with sustainable competitive advantage.That is required by Sivarajah et al. ( 2017) and Dahiya et al (2021) and Ramadan et al (2020) respectively.Moreover, the study will also be helpful in making studies of Big-Data active that were found to be in the infancy stage (Sivarajah et al., 2017), also for the relationship between data and data analytics in the field of e-commerce (Yang et al., 2019).The findings of the study are also relevant to Lambrecht and Tucker (2015) and Lambrecht and Tucker (2016), as the data available are effectively related to advanced algorithms that create Big-Data analytics.That means an e-picture of consumer footprints is effectively created through browsing and purchasing information.Moreover, linkage of Big-Data knowledge is required for making innovative moves and therefore the findings are consistent with Davenport et al. (2012).Therefore, the linkage between all the devices must be developed with (IoT), to assist Big-Data analytics effectively, as indicated by Ramadan et al. (2020).
Last but not least the findings are also effective to legitimize the required relation i.e., Big-data knowledge with the attainment of sustainable competitive advantage (SCA).
Although this requires effective use of firm capabilities and resources as indicated by Ramadan et al. (2020) and verified through the variable of innovative capabilities between Big-Data analytics and SCA.However, the insignificant relationship i.e.Data Availability -> Advanced Algorithms -> Big-Data Analytics -> Big-Data Knowledge -> Innovative Capabilities -> Sustainable Competitive Advantage, is there which still assures the indications of Dahiya et al. (2021).That means the indication made by Sultan et al (2021), that there is also a need for skilled data scientists for making Big-data work in an effective way.Although lacking Big-Data skills in the scientists working in the retail industry of Pakistan are disturbing the firm performance.Similar research has been reflected by Jabeen et al. (2021).

Conclusion
This study is fruitful to indicate that Big-Data is effective for boosting innovative capabilities, especially for online retail.However, the relationship of Big-Data through serial mediation is not clear in the attainment of sustainable competitive advantage.Therefore, the role of technical (IT), skills is significant as previously indicated by research work from Pakistan.Hence, legitimate to quote that relationship (serial mediation), would become more vibrant when supplemented with the moderation of skills from IT personnel.

Need for Future Research
The study is one of the initial ones to explore the effect of Big-Data on the attainment of sustainable competitive advantage, especially with respect to eastern and lower-income sides of the world.The other specialty of the study is a quantitative analysis which is a rare element in the studies of Big-Data which became further enlightened through the use of SEM, and SMART-PLS in the field of the online retail segment.The segment is really special not because only of its infancy stage (IBM, 2018) but also due to the continuous increase in the level of preference for online purchases (Mehmood andNajmi, 2017 &Rehman et al., 2019).
Therefore, the study will be fruitful for academicians as well as managers related especially to online retail segments to realize the importance of Big-Data Analytics and Big-Data knowledge along with the skill set of data scientists.Further studies might be conducted by applying moderation of skilled data scientists between data availability & Big-Data algorithms as the skill set of data scientists is not found to be effective in countries like Pakistan.Therefore, checking these forms of complex models might become more realistic and significant with the inclusion of moderation roles of skilled data scientists.

Table . 2. Predictive Accuracy (Quality Criteria)
(MohdDzin & Lay, 2021)lect construct reliability and convergent validity.The reliability has been reflected through three measures i.e., Cronbach Alpha, Goldstein rho, and composite reliability(Wong, 2013).Convergent validity requires three criteria composite reliability, (Average Variance Extracted) AVE and outer loading(Afthanorhan, 2013).The outer loadings are available in the table.1,however,AbHamid, Sami and Sidek, (2017) reflected that AVE can be the legitimate indicator of convergent validity when it is square of outer loading and must be equal to or greater than 0.5(MohdDzin & Lay, 2021).Therefore, in the light of these criteria is it realistic to declare a table is fulfilling the criteria of construct reliability and convergent validity.

Source: Author's own elaborationTable 1 ,
Table 2, Table3, and Table4are used to reflect the appropriateness, reliability, and validity of the data.However, the purpose of Table5 and Table 6is to reflect the impact of variables through inferential statistics.Analysis indicated that only two relationships in the category of serial mediation i.e., Advanced Algorithms -> Big-Data