‘Data over intuition’ – How big data analytics revolutionises the strategic decision-making processes in enterprises

(1)

‘Data over intuition’ –

How big data analytics revolutionises the

strategic decision-making processes in enterprises

A single case study of IKEA

MASTER THESIS WITHIN: Business Administration

NUMBER OF CREDITS: 30 ECTS

PROGRAMME OF STUDY: Digital Business

AUTHOR: Finn Brand, Filip Höcker

TUTOR:Matthias Waldkirch

(2)

Master Thesis in Business Administration

Title: ‘Data over intuition’ – How big data analytics revolutionises the strategic decision-making processes in enterprises

Authors: Finn Brand and Filip Höcker Tutor: Matthias Waldkirch

Date: 2020-05-18

Subject terms: Big data; Big data analytics; Strategic decision-making; Strategy-as-practice, Data-driven decision-making

Abstract

Background: Digital technologies are increasingly transforming traditional businesses, and their

pervasive impact is leading to a radical restructuring of entire industries. While the significance of generating competitive advantages for businesses utilizing big data analytics is recognized, there is still a lack of consensus of big data analytics influencing strategic decision-making in organisations. As big data and big data analytics become increasingly common, understanding the factors influencing decision-making quality becomes of paramount importance for businesses.

Purpose: This thesis investigates how big data and big data analytics affect the operational strategic

decision-making processes in enterprises through the theoretical lens of the strategy-as-practice framework.

Method: The study follows an abductive research approach by testing a theory (i.e.,

strategy-as-practice) through the use of a qualitative research design. A single case study of IKEA was conducted to generate the primary data for this thesis. Sampling is carried out internally at IKEA by first identifying the heads of the different departments within the data analysis and from there applying the snowball sampling technique, to increase the number of interviewees and to ensure the collection of enough data for coding.

Findings: The findings show that big data analytics has a decisive influence on practitioners. At

IKEA, data analysts have become an integral part of the operational strategic decision-making processes and discussions are driven by data and rigor rather than by gut and intuition. In terms of practices, it became apparent that big data analytics has led to a more performance-oriented use of strategic tools and enabling IKEA to make strategic decisions in real-time, which not only increases agility but also mitigates the risk of wrong decisions.

(3)

Acknowledgements

This Master Thesis and the research process would not have been possible to conduct without certain people who deserves the utmost appreciation.

Firstly, an especial thank you goes out to IKEA, and the employees which has provided valuable information by voluntarily participate in the interviews performed in the case study.

Secondly, we would like to express our special appreciation and deep gratitude to their supervisor Matthias Waldkirch for his extensive and valuable support and constructive criticism of this work.

Finally, we do not want to miss the opportunity to thank all the members of their seminar group (Ouafaa Cherradi & Cansu Tetik, Stephanie Muth & Marius Rauscher, and Sandra Henkel & Gesa Köhrbrück) who gave valuable and constructive feedback in each of the seminars.

________________________ ________________________

(4)

Abbreviation List

BD Big data

BDA Big data analytics

ERP Enterprise resource planning

INGKA The largest franchisee taker of the IKEA franchise

MNC Multinational corporation

SD-SD Structured decisions based on structured data SD-UD Structured decisions based on unstructured data UD-SD Unstructured decisions based on structured data UD-UD Unstructured decisions based on unstructured data

(7)

1 Introduction

__________________________________________________________________________________________

Research problem, research purpose, and research question are presented and explained to create an understanding of the concepts discussed later.

Background

The rapid pace the business environment is transforming in is not merely causing ambiguity when looking at future trends, but also, significant challenges are put in front of organisations (Kotter, 2014). By implementing big data analytics within an organisation, possibilities arise to disrupt the structure within senior management along with changing how decision are being made (Gupta & George, 2016; Merendino, Dibb, Meadows, Quinn, Wilson, Simkin, & Canhoto, 2018; Vidgen, Shaw, & Grant, 2017; Wang, Xu, Fujita, & Liu, 2016). Janssen, van der Voort, & Wahyudi (2017) are highlighting the disruptive effect that the usage of big data analytics has had on current ways of working for both directors as well as decision-makers. A trend is identified where there is a knowledge shift, and decision-making mandates are shifted in organisations (Bumblauskas, Nold, Bumblauskas & Igou, 2017). Instead of having experience and intuition as grounds for a decision, data is being used to understand how to make the correct decision (Gupta & George, 2016; McAfee & Brynjolfsson, 2012). Furthermore, Merendino et al. (2018) argues that the rapid emergence of big data analytics is enabling decision-makers to make decisions based on higher quality data, and by doing so, being able to make quicker and most importantly, make better decisions. The tool big data has quickly become among the biggest concepts for companies to apply for optimizing business operations during the 21st century (Grover, Chiang, Liang & Zhang, 2018; Sivarajah, Kamal, Irani & Weerakkody, 2017). As the usage of big data analytics is becoming standard practice, companies are reporting increases in productivity, as well as improving areas within decision-making (McAfee & Brynjolfsson, 2012; Sheng, Amankwah-Amoah & Wang, 2017). Researchers argue that big data analytics has changed and transformed business models as well as enabling decision-makers to act on structural changes promptly (Rehman, Chang, Batool, & Wah, 2016; Sheng et al., 2017). The rapid increase in interest and usage for big data has emerged as e-commerce firms apply big data analytics, and in the process transforming into a data-driven company (Grewal, Roggeveen, & Nordfält, 2017). By applying big data analytics in the decision-making process, possibilities are enabled to easier make the correct decisions and developing strategies that are well-grounded in organisational data (Intezari & Gressel, 2017; Pauleen & Wang, 2017). The net result is being able to construct strategies that have a higher certainty of success

(8)

compared to previous methods (Bumblauskas et al., 2017). With the emergence of big data, companies are enabled to make decisions grounded in information that previously has been difficult to understand and analyse (Akter, Wamba, Gunasekaran, Dubey, & Childe, 2016). By analysing and applying big data within a company, possibilities to identify patterns and factors that are impacting the ways customers navigate and make purchasing decisions (Grewal et al., 2017). However, in the process a company needs to invest in creating digital capabilities, i.e. becoming a data-driven company (Bharadwaj, El Sawy, Pavlou, & Venkatraman, 2013). By creating and developing digital capabilities decision-makers are enabled to understand what they need to do to predict patterns to improve their decision-making abilities (Janssen et al., 2017). However, research also states that there is a gap at the individual level of directors, stating that shortages are found in the “cognitive capabilities” (Merendino et al., 2018, p. 74) which are necessary to be able to grasp the full potential of big data, i.e. shortcomings in digital capabilities (Merendino et al., 2018). Furthermore, research also states that implementing big data analytics in the decision-making process creates a potential risk of “board cohesion” (Merendino et al., 2018, p. 74) to be disrupted which instead of improving decision-making processes, has negative consequences such as creating uncertainty and inertia (Merendino et al., 2018). In the rapidly changing business markets of today, companies cannot afford to disrupt the chain of decision-making, but rather needs to make it more efficient and lean to increase the competitive gap a company might have towards their competition (Grewal et al., 2017). To keep relevance within a market, companies can evolve, develop digital capabilities, and becoming data-driven (Bharadwaj et al., 2013). By becoming a data-driven company, further analysis is made of the data collected from customers and operations to effectively increase the capabilities of managing a company and making quick and agile decisions (McKenzie, van Winkelen, & Grewal, 2011).

Problem

It is imperative for organisations to come closer to its customers and adapt to the changing business environment (Bharadwaj et al., 2013; Sheng et al., 2017). Decisions need to be made faster and with higher accuracy than ever before (Pauleen & Wang, 2017). Furthermore, data has been collected in firms for many years, however, the knowledge of how to interpret it has historically been flawed (Bradlow, Gangwar, Kopalle & Voleti, 2017). Another issue has been the management of the data, whereas the data has not been used to correctly achieve a result, and instead, decisions are being based on intuition or experience alone (McAfee & Brynjolfsson, 2012). This means that organisations that are failing to adapt big data analytics are risking of falling behind by not analysing their operational performance in the same way as the competition (Ghasemaghaei, Hassanein, &

(9)

Turel, 2017; Merendino et al., 2018). The issue many companies currently face is partly the organisational transformation that needs to be done to start capitalizing on the new technology, but also to identify in what way and how an organisation can capitalize to decrease the steps and involvement in decision-making (Akter et al., 2016; Bharadwaj et al., 2013). The risk of making a miscalculated decision can be high by not being able to analyse the collected data in a correct way (Zhao, Fan & Hu, 2014). This can ultimately lead to a company making inappropriate decisions and as a result lose business and fail to capitalize and increase a competitive advantage (Ghasemaghaei et al., 2017).

Purpose

The purpose of this research is to investigate how the operational strategic decision-making process is changing within a firm as big data analytics has been implemented. The aim of this study is to investigate one of the largest retailers of the world, IKEA, to create an understanding of how a major retailer can shift and apply big data analytics to become more efficient when making strategic decisions. It is of high interest to understand how the process of decision-making is changing and how big data analytics impacts the daily practice of decision-makers and strategists. As an internationally operating company in the retail industry, IKEA offers an excellent empirical context for the given thesis, as the multinational group possesses a plethora of transaction-related customer data, which is mainly generated from furniture sales in its worldwide branches and increasingly from online sales through its website and its complex supply chain that generates vast amounts of internal operational data (Ringstrom & Clarke, 2019). In addition, IKEA is monitoring all parts of the value chain from listening to the customers, developing furniture, manufacturing, packaging, and distributing to ultimately create the best service for the customers (The IKEA value

chain, n.d.). Thus, the organisation truly processes big data at its disposal, which is systematically

monitored and has a decisive impact on the strategic orientation of the company. The retailer operates in a competitive industry that is increasingly dominated by big players that also provide additional value-added services besides furniture. Positioning the company as a mere furniture retailer is no longer enough to stay ahead of the competition; creative ideas and innovative products and services for customers are required. IKEA’s vision “to offer a wide range of well-designed, functional home furnishing products at prices so low that as many people as possible will be able to afford them” clearly focuses on customer centricity (The IKEA vision and business

idea, n.d.). The retailer is investing in making all future decisions based on a combination of

(10)

With regard to the aforementioned problem, the current market situation within the retailing industry, company management needs to be rapid when it comes to decision-making to strengthen a company’s position (Bradlow et al., 2017). As research states, big data analytics are enabling tools that management can use to make well-grounded decisions based on data that has been difficult to analyse previously (Merendino et al., 2018). By utilizing this, companies possess the capabilities of becoming more agile, and in the long run increase their market position (Akter et al., 2016; Ghasemaghaei et al., 2017). However, as mentioned, implementing big data analytics within a firm can create a disruptive environment within the decision-making chain, and thereby disorder a firm’s ability to make agile and quick decisions. Should this happen, companies pose the risk of losing their competitive advantage quickly (Merendino et al., 2018). The purpose of this study is to, by support from theory, understand how big data analytics has impacted the practitioners, practices, and praxis at IKEA making strategic decisions on an operational level, and by doing so, creating a generality of how companies can capitalize from big data analytics for their strategic decision-making processes. By investigating how IKEA has acted, this research will understand how companies can shift dynamically and improving their strategic decision-making process by utilizing big data analytics. This research is of importance, not only because the phenomenon of big data analytics is a recent topic, but also to investigate how the tool is simplifying the strategic decision-making process on an operational level. Furthermore, recent research is focused on the overall impact BD has had within firms that apply the technology and generally touch upon organisational performance (e.g. Akter et al., 2016; Côrte-Real, Oliveira, & Ruivo, 2017; Fosso Wamba, Akter, Edwards, Chopin & Gnanzou, 2015; Gupta & George, 2016). Their research conducted in the field is of great value, however, there is a gap in how big data analytics is affecting parts of an organisation in detail, seen from a strategic decision-making point.

Research question

It is commonly known that big data improves the information flow in organisations, but it is unclear how it affects strategic decision-making. Thus, the research question that has driven this thesis focuses on:

RQ: How does the use of big data analytics affect practitioners, practices, and praxis in strategic decision-making on an operational level?

To answer this research question, the following chapters review the literature on big data and big data analytics in business, management, and decision-making research to provide a thorough understanding of the phenomenon and its role in strategic decision-making. The theoretical lens of the given thesis is grounded in the strategy-as-practice (s-as-p) theory as the research attempts

(11)

to identify the influence of big data analytics on humanized management and organisation practices. This thesis attempts to contribute to the research problem by answering the given research question. More specifically, the goals of this thesis are:

• To identify how big data analytics affects the strategic decision-making processes in an organisation on an operational level.

• To analyse the impact of big data analytics on practitioners, practices, and praxis. • To provide practical implications for data-driven strategic decision-making.

This thesis is organised into four parts. First, a literature review on BD and BDA and how it is affecting strategic decision-making processes in enterprises. Continuing with the theoretical framework that has driven the empirical investigation of this study. Thereafter, a chapter touching upon the methodology of the empirical research, followed by findings of the empirical study. Finally, practical implications are discussed, and conclusions are drawn.

Delimitations

Although the need for implementing big data analytics can have a big impact in many industries and organisations, a delimitation of this thesis has been to concentrate on one multinational corporation (MNC) that for many years has collected a lot of data, but never had a deep analysis of it. Firstly, this decision was made partly in with the reasoning of narrowing down the scope of this thesis, as a type of this research by nature can become were extensive. Secondly, this research was delimited to one MNC due to the lack of agility that can be compared with smaller organisations. Also, it is acknowledged that many perspectives could be taken into consideration, as this master’s thesis has been conducted within the field of digital business, it has adopted a combined view of a digital point of view and a managerial point of view. Considering this aspect, the interviewees encountered has been chosen accordingly. Furthermore, due to the scope of the thesis, the study has been delimited to collect data in the form of semi-structured in-depth interviews from a single company. As a result of the delimitations mentioned, difficulties arise to develop general conclusions across industries or companies. However, the aim of this research is to deliver generalities of how big data analytics could affect the strategic decision-making process within a company, and by doing so, generate guidelines for future studies.

(12)

2 Literature review and theoretical framework

__________________________________________________________________________________________

The purpose of this chapter is to provide the theoretical background for this thesis. First, a literature review examines what is understood by the term big data and big data analytics and how it is affecting strategic decision-making processes in an enterprise. Thereafter, the theoretical framework is presented.

Frame of references

Contemporary research on business and management offers evidence that big data and big data analytics are directly linked to strategic decision-making processes within organisations (Grover et al., 2018; Intezari & Gressel, 2017; Janssen et al., 2017; Pauleen & Wang, 2017; Sheng et al., 2017; Wang et al., 2016). The interdisciplinarity regarding the distinct academic domains (i.e. big data analytics, decision-making, and strategy) demand separate disciplinary investigations in the following literature review. Recently published business and management studies investigated the effects of big data analytics on strategy- and decision-making processes (Gupta, Kar, Baabdullah, & Al-Khowaiter, 2018; Intezari & Gressel, 2017; Janssen et al., 2017; Merendino et al., 2018). Further, by complementing the business and management disciplines with research in decision science, the influence of big data analytics on human decision-making processes can be elucidated in-depth (Intezari & Gressel, 2017; Pauleen & Wang, 2017). Research from the aforementioned disciplines frame the scope of references and are presented in the following paragraphs.

Literature collection

To make sure only to have relevant and true literature, the databases JU Primo, Web of Science, and Google Scholar have been used. In addition, when selecting the relevant literature, care was taken to ensure that the academic articles were taken from high-quality journals that are respected in the relevant disciplines. To determine the current state of literature concerning the phenomenon, search syntaxes such as “big data” and “big data analytics” were used in the databases. Furthermore, to specify the search and narrow down the topic, keywords such as “decision-making” and “strategic decision-making” were entered, separately and combined with “big data” and “big data analytics”. To ensure that the correct domain-specific keywords were used, alternative terms and phrases such as “BD”, “BDA”, “data”, “data analytics”, “strategy making”, “operational decision-making”, “operational strategies” were entered. In addition to the keyword search, as soon as relevant literature was found, further searches within the found literature were carried out to expand the collection of sources. An overview of the search syntaxes and the corresponding databases can be found in Appendix A.

(13)

Literature review 2.3.1 Big data

There exists no universal definition of the term “big data” and scholars suggested that big data can be considered as a “moving definition”, which varies in its meaning depending on the ever-increasing nature of this phenomenon (Gupta et al., 2018; Sheng et al., 2017). Due to its ubiquity, the term is used in various disciplines and thus has a wide range of definitions and interpretations (Mikalef, Pappas, Krogstie & Giannakos, 2018). Moreover, there is no fixed technical threshold for measurement of what size and type of data can be treated as big data, but there appears to be a consensus in the literature about its unique characteristics that distinguish big data from conventional data, namely the V’s of big data (Akter et al., 2016; Grover et al., 2018; Gupta & George, 2016; Janssen et al., 2017; Sheng et al., 2017). Although big data has various definitions, for the further elaboration of this thesis the authors refer to a definition widely used in the literature to define big data as:

“[…] extremely large amount of structured, semi-structured or unstructured data continuously generated from diversified sources, which inundates business operations in real-time and impacts on decision-making through mining insightful information from rambling data.” (Sheng et al., 2017, p. 98)

2.3.1.1 The V’s of big data

This differentiation from conventional data helps in understanding and classifying the concept of big data. According to Gupta & George (2016), the term big data was initially coined to reflect the “bigness” or voluminous size of data generated as a result of using new forms of technology. A recently published study by Mikalef et al. (2018) coincide with these assumptions and state that in addition to the volume property two other characteristics, namely velocity and variety met the unique characteristics of big data. Their study reviewed several publications and consolidated various definitions of big data and its associated attributes. The majority of the reviewed publications described its uniqueness with three Vs, namely volume, which “[...] refers to the sheer size of the data set due to the aggregation of a large number of variables and an even larger set of observations for each variable.” (Mikalef et al., 2018, p. 554), velocity, which “[...] reflects the speed at which these data collected, updated, and analysed, as well as the rate at which their value becomes obsolete.” (Mikalef et al., 2018, p. 554) and variety, which “[...] refers to the plurality of structured and unstructured data sources, which, amongst others, include text, audio, images, video, networks, and graphics.” (Mikalef et al., 2018, p. 554). However, due to the ever-changing

(14)

character, big data is constantly redefined, and several scholars have supplemented these three attributes with other Vs. A commonly acknowledged aspect is its veracity, which “[...] refers to the degree to which big data is trusted, authentic, and protected from unauthorized access and modification.” (Mikalef et al., 2018, p. 555). And further, some scholars refer to its value, which “[...] refers to the value it’s creating for organisations.” (Mikalef et al., 2018, p. 555) as the fifth attribute (Grover et al., 2018; Gupta et al., 2018; Janssen et al., 2017; Mikalef et al., 2018; Rehman et al., 2016; Sheng et al., 2017; Wang et al., 2016).

Studies that primarily investigate the influence of big data on strategic decision-making processes within organisations, which is also the focus of this thesis, are mainly concerned with the velocity attribute (Intezari & Gressel, 2017; Merendino et al., 2018). Velocity plays a decisive role in strategic decisions, as the development of technological data-driven infrastructures enables data to be managed in “[…] continuous flows and processes.” (Davenport, Barth & Bean, 2012, p. 23), allowing decisions to be made in real-time (Intezari & Gressel, 2017). The following paragraphs provide a working definition of big data.

2.3.1.2 Definitional perspectives

Wang et al. (2016) distinguish four distinct perspectives for defining BD, such as “product-oriented”, “process-“product-oriented”, “cognition-oriented” and “social movement” perspective (Wang et al., 2016, p. 749). All of which highlight and emphasize different attributes concerning what constitutes BD. The product-oriented perspective highlights the attributes of data regarding their sizes, speeds, and structures (Wang et al., 2016). Definitions that are classified as a process-oriented perspective concentrate on the novelty of processes required and involved in storing, managing, aggregating, and analysing BD (Wang et al., 2016). The cognition-based perspective focuses on the challenges caused by big data concerning their cognitive capacities and limitations (Wang et al., 2016). Finally, the social movement perspective draws attention to the gap between vision and reality, especially the socioeconomic, cultural, and political shifts that underlie the presence of big data (Wang et al., 2016).

Since this thesis attempts to broaden the scope of reviews in the field of business and management by investigating the influence of big data analytics on the strategic decision-making processes in organisations, only the first three perspectives, namely product-oriented, process-oriented, and cognition-oriented, are taken into consideration as relevant definitional perspectives for the given investigation.

(15)

2.3.2 Big data in enterprises

“Big data has rapidly moved to being a mainstream activity of organisations.”

(Janssen et al., 2017, p. 338)

Over the past two decades, the field of big data in the business context has become increasingly important (Gupta & George, 2016). While organisations in the past have primarily focused on enterprise-specific structured data (i.e., data that can be stored in relational databases) to make business decisions, today’s organisations tend to capture every bit of information regardless of the size of data, the structure of data, and the speed at which data are created (Zhao et al., 2014). Hence, structured data is no longer the only type in enterprises – rather more unstructured and semi-structured data is processed to support strategic decision-making processes internally (Sheng et al., 2017). Unstructured data occurs in various forms such as text (e.g. documents), web data (e.g. web usage, web content), social media data (e.g. online platforms), multimedia data (e.g. image, audio, video) and mobile data (e.g. sensor, geographical location, application) (Sheng et al., 2017). However, the mere collection of data alone does not add any value to decision-makers, enterprises require holistic data-driven ecosystems to grasp the full potential (Gupta & George, 2016).

2.3.2.1 Big data value creation

Well-cited academic papers investigated various data-driven enterprises and developed models, frameworks, and so-called blueprints for businesses to turn data into relevant information for decision-makers (Grover et al., 2018; Gupta & George, 2016; Janssen et al., 2017; Wang et al., 2016). Although these frameworks are designed based on varying theoretical assumptions, they provide a solid understanding of data-driven infrastructures and ecosystems that effect decision-making. Moreover, while studying and analysing these models, patterns became apparent, which will be highlighted in more detail in the following part.

These models, i.e. the “Big data chain” (Janssen et al., 2017, p. 340) and the “Paradigm of Big Data

processing” (Wang et al., 2016, p. 751), have similarities in their design. Both models illustrate how

big data can provide information for decision-making. According to Janssen et al. (2017), a big data chain begins with collecting data and ends when decisions are made (see Figure 1).

(16)

Wang et al. (2016) coincide with these notions as their framework encompasses intelligent decision-making based on data as an outcome of their process (see Figure 2). Both models are in affect similar in their design. They start with data collection/data capture steps, which include the gathering and collating of data. Followed by a preparation or curation phase, as the raw data requires a processing measure before it gets analysed in a third step. Lastly, the result of the analysis is interpreted by a decision-maker, who has the knowledge to utilize the results for appropriate decisions (Janssen et al., 2017; Wang et al., 2016). These decisions are dependent on the strategic goals and orientations of the enterprise (Janssen et al., 2017).

An integral part of these models is an analysis step that is indispensable to gain insights for data-based decision making. The following chapter gives an outlook on the practice of big data analytics.

2.3.3 Big data analytics capabilities

The creation of big data analytical capabilities is indispensable for companies that want to translate data into insights (Gupta & George, 2016). In essence, big data analytics encompasses not only the data upon which analysis is performed, but also elements of tools, techniques, skills, and capabilities (Mikalef et al., 2018). Gupta & George (2016) identify three key building blocks of big data analytical capabilities: Tangible resources (i.e., data, technology, and basic resources), human resources (i.e., managerial skills, and technical skills), and intangible resources (i.e., data-driven culture and organisational learning). In the following subchapters, the three categories are examined in more detail to understand the big data capabilities in their entirety.

(17)

2.3.3.1 Tangible resources

Gupta & George (2016) argue, that firms need to invest in advanced technologies (e.g. database management systems, cloud computing, data warehouses, etc.), software (e.g. Hadoop, Microsoft Azure, Tableau, etc.), and programming languages (e.g. Python, R, SQL) to be able to store, process and analyse diverse data sets. Thus, for example, highly decentralized multinational corporations with several international branches, such as IKEA or other players in the online retailing industry, can gain access to databases from multiple branches via cloud-based solutions (Sivarajah et al., 2017). Although these systems and resources are readily available for all firms and are unlikely to provide any competitive advantage on their own, yet they are required to create capabilities (Gupta & George, 2016).

2.3.3.2 Human resources

Further, Gupta & George (2016) emphasize on two categories of human resources that are necessary to build big data capabilities: Managerial and technical skills. Managerial skills demand strong analytical acumen and the ability to put data analytical results into a strategic perspective to make organisational-wide decisions (Gupta & George, 2016). Managers believing in big data should emphasize finding the right human capital including technical, analytical, and governance skills for data analysts (Gupta & George, 2016). Technical skills are highly sophisticated and some of these skills include competencies in machine learning, data cleaning, statistical analysis, and understanding of programming languages (Davenport, 2014). Centralized departments can assist in consolidating capabilities and skills, thus leveraging accumulated know-how. Departments and teams, namely Business Analytics (BA), Business Intelligence (BI), Centre of Excellence (CoE), and Data Science teams process, analyse, and visualize data that build the foundation for decision-makers in various business operations such as process management, market intelligence, sales forecasting, price optimization, and inventory management (Grover et al., 2018; Gupta & George, 2016).

2.3.3.3 Intangible resources

Moreover, Gupta & George (2016) differ between two intangible resources that make data analytics capabilities: Data-driven culture and intensity of organisational learning. To realize the full potential, firms must develop a data-driven culture (Gupta & George, 2016). Given that employees at all levels in an organisation are required to make decisions, it is pertinent to diffuse the culture of data-driven decision-making to all levels such as that organisational members, regardless of their job titles, can make decisions that are grounded on evidence as suggested from

(18)

data (Ross et al., 2013). Once, data-driven infrastructures and big data capabilities are established, managers might concentrate on how to retain it and even sense and seize novel opportunities (Gupta & George, 2016). Firms that can learn and reconfigure their resources according to changes in their external environment will likely have a sustained competitive advantage (Gupta & George, 2016).

As Gupta & George (2016) state, the purpose of processing big data is to exploit knowledge from data to support intelligent decision-making, the following section analyses how big data creates knowledge for decision-making.

2.3.4 The role of big data in decision-making

“Big data’s power does not erase the need for vision or human insight.”

(McAfee & Brynjolfsson, 2012, p. 65)

As the previous paragraphs emphasize, has big data prompted decision-making processes in organisations. Decision-making is central to what managers do (Hickson et al., 1989; Michel, 2007; Stewart, 2006), and it is integrated into all kinds of management functions (Intezari & Gressel, 2017).

While the business and management discipline investigates data-driven decision-making in the realm of enterprise-specific practices, processes, and structures (Grover et al., 2018; Janssen et al., 2017; Merendino et al., 2018; Wang et al., 2016; Xu et al., 2016), little research has been conducted from a psychological perspective to what extent data influences a person’s decision-making (Intezari & Gressel, 2017; Mazzei & Noble, 2017; Pauleen & Wang, 2017). When investigating studies in decision science, it becomes apparent that human knowledge and experience are solely responsible for decisions, therefore negating the influence of knowledge when discussing the impact of big data on decision-making is impossible (Pauleen & Wang, 2017). The knowledge explosion that accompanies increasing access to BD has a major impact on how and what information managers use to inform their decisions (Pauleen & Wang, 2017). Merendino et al. (2018), investigated the influence of big data on board-level decision-making and argue that understanding the transformational process of changing meaningless raw data into knowledge is essential for decision-makers. Hence, before examining the structure and processes associated with corporate decision-making, it is crucial to understand how data creates knowledge and how it is incorporated into decision-making.

The definitional complexity and plurality debates around knowledge lead the following examination to concentrate merely on how knowledge is achieved through big data analytical

(19)

processes, and not on what knowledge is since it has different meanings in different contexts (Pauleen & Wang, 2017).

2.3.4.1 Knowledge creation through big data

Knowledge is central to any discussion around big data – whether the data generated are used in operational, tactical, or strategic business domains, knowledge will guide its use (Pauleen & Wang, 2017). Wang et al. (2016) developed the “The framework of Big data decision making” (Wang et al., 2016, p. 751) that displays the relationships between both phenomena by providing an overview of the inner-human process of decision-making and connecting it to various stages in an organisation (see Figure 3). The authors explain that decisions are made by deriving information from data, obtaining knowledge from information, and then achieving wisdom from knowledge (Wang et al., 2016).

When looking at the model, two aspects, in particular, are striking. First, the decision-making process follows a funnel-shaped approach, starting with a vast amount of raw data and ending with a narrowed knowledge and wisdom peak. Secondly, decisions take place at varying levels within the organisations (i.e., operational, tactical, strategic, systemic). By processing and analysing data at the operational level, decisions at subsequent levels are increasingly based on facts and data-related evidence (Wang et al., 2016). As mentioned, Wang et al. (2016) locate decisions on an operational, tactical, strategic, and systemic level. To understand how big data analytics influences decisions at these levels, the following chapter is dedicated to this subject.

(20)

2.3.4.2 Locus of data-driven decisions

Decisions are made in varying organisational settings and are mostly initiated with particular business objectives in the decision-makers’ minds (Pauleen & Wang, 2017). According to Pauleen & Wang (2017), decisions only become meaningful when they are classified and brought into relation by decision-makers, the underlying concept is called contextual knowledge. Combing it with the three levels of decision-making proposed by Merendino et al. (2018), varying types of decisions favoured by big data become clearer. The authors outline three levels for data-driven decision-making within enterprises: The director/individual level, the board level, and the stakeholders’ level (Merendino et al., 2018). At the director/individual level, managers and professionals utilize data-driven analytics to make operational decisions for day-to-day business matters (e.g., project-based decisions, performance-related decisions, product- and service-related decisions, etc.) (Merendino et al., 2018). At the board level, big data analytics can deliver insights to more complex decisions (e.g., structural decisions, global supply chain decisions, etc.) (Merendino et al., 2018). The increased number of Chief Data Officers (CDO), Chief Information Officers (CIO), and Chief Analytics Officer (CAO) on the boards of directors show that decision-makers at board level are placing greater emphasis on technical and data analytical skills in their decision-making (Côrte-Real et al., 2017; Mazzei & Noble, 2017). At the stakeholder level, companies are expected to respond to environmental changes by correctly anticipating the stakeholders’ needs (Merendino et al., 2018). By providing new insights into environmental trends, big data can empower corporate decision-makers to respond and adapt to current dynamic economic demands (e.g., market decision, governmental restrictions, fiscal policies, etc.) (Erevelles et al., 2016; Merendino et al., 2018).

The following chapter examines how big data is tied to strategic processes. Here, the links between big data analytics and strategic processes are discussed in-depth.

2.3.5 The role of big data analytics in strategic processes

“Big data in effect multiplies the potential of organisational data engagement and the shaping of enterprise strategy processes.”

(Bhimani, 2015, p. 3)

The chapters in the previous sections indicate that data enables business decisions to be more far-reaching and conclusive. At a strategic level, data can be utilized to target more effective interventions in areas that so far have been dominated by gut and intuition rather than by data and rigor (Bhimani, 2015; Constantiou & Kallinikos, 2015; McAfee & Brynjolfsson, 2012). Making

(21)

effective strategic decisions is one of the critical abilities that managers are required to have and develop to lead their organisations in the increasingly volatile and competitive business environment (Intezari & Gressel, 2017). As Porter (1985) emphasizes, the success or failure of a firm relies mainly on the managers’ competitive ability to make strategic decisions. While in the past managers operated under conditions of information scarcity with decisions made with incomplete and often unstructured binary data, big data create conditions of information abundance due to the massive amount of detailed data made available, which enables strategic decisions (Intezari & Gressel, 2017).

2.3.5.1 The changing context of strategy

Strategy and decision-making are closely intertwined, as decisions form the foundation for any strategic initiative (Intezari & Gressel, 2017). Therefore, big data analytics not only facilitates decision-making processes but also has a considerable effect on the domain of strategy (Intezari & Gressel, 2017). The academic discourse provides various reasons for this perceptible impact. As Constantiou & Kallinikos (2015) note, the process of strategy-making is directly linked to models of data collection (e.g. market research instruments and statistical models of inference, classification, or management accounting systems), which indicates a direct link between the two phenomena. These models are loaded with structured (often quantitative) data to provide insights in a persuasive and argumentative way that can be leveraged for strategic measures (Constantiou & Kallinikos, 2015). The underlying data enables constant updateability and real-time responses to organisational activities (Constantiou & Kallinikos, 2015). Real-time updates of operational data sets provide valuable information and direct feedback on strategic implications. However, the growing number of data sources driven by big data efforts poses a challenge for enterprises in that the data arise from wider configurations of information pools (Bhimani, 2015). Big data is not only produced by internal homogenous data sources, but the majority is also a hugely heterogeneous base, actively through user-generated content, which unleashes a diversity of strategic re-orientation possibilities (Constantiou & Kallinikos, 2015; Galbraith, 2014).

2.3.5.1.1 Growing number of relevant data sources

Modern enterprises collect a plethora of data from various sources that are all relevant for strategic decisions. The literature postulates a distinct classification of organisational data sources. Rehman et al. (2016) distinguish direct and indirect data sources. Direct data sources in enterprises generate operational information relevant to supply chain management, production, behaviour analysis of employees, etc. (Rehman et al., 2016). Indirect information includes data that is not generated

(22)

within the boundaries of the enterprises, such as customer-specific data (Rehman et al., 2016). This data can be acquired, for example, from websites by analysing customers’ feedbacks and online product reviews (Rehman et al., 2016).

Likewise, Schmidt & Möhring (2013) followed a similar approach in the classification of data sources. The authors developed a graph that illustrates different data sources, data types, and data volume (see Figure 4). The data sources are grouped into four categories, namely Enterprise Resource Planning (ERP), Customer interactions, Weblogs and Sensor Data, and Social Media (Schmidt & Möhring, 2013).

What is striking, is that both the amount and structure of the data becomes more complex as the data sources change from internal to external, or from ERP to Social Media. This indicates that enterprises that increasingly include external data sources (i.e., Social Media) in their strategic planning must process more complex data sets (Schmidt & Möhring, 2013). Sources of data that were previously not relevant to strategy-making might become more meaningful, depending on the strategic orientation of the enterprise (Schmidt & Möhring, 2013). The literature coincides with this statement, as Constantiou & Kalliniko (2015) argue, that strategy-making through big data is becoming a part of a wider context of social relationships that are shaped in ways that tend to redefine organisational boundaries.

Not only the foundations for strategic processes have changed, but also the directions of strategic processes internally. The following subsection will emphasize on the changed processes.

(23)

2.3.5.1.2 Top-down vs. ad hoc information flow

Many authors perceive a change in the direction of information within data-driven enterprises (Bhimani, 2015; Constantiou & Kallinikos, 2015; Mazzei & Noble, 2017; Merendino et al., 2018; Tabesh et al., 2019). Constantiou & Kallinikos (2015) notice that many of the methods and techniques of data processing are strongly associated with bottom-up procedures and are supposed to discover patterns in huge databases, the top-down, deductive approach to data gathering and utilization, a cornerstone of standard and predominantly prescriptive ways of traditional strategy-making, is challenged. An ad hoc, inductivist way of direction of strategic processes is emerging, which seems to undermine the foundations of predictive models of strategy-making (Constantiou & Kallinikos, 2015). Mazzei & Noble (2017) and Merendino et al. (2018), coincide with these statements. The former state that big data is transforming the strategic thought process of corporate strategists and managers. The authors argue that rather than corporate strategy dictating which data should be collected and analysed, the data collected and analysed is having a dramatic influence on corporate strategy (Mazzei & Noble, 2017). Merendino et al. (2018) also identify a change, as “[…] boards moving away from top-down planning;” (Merendino et al., 2018, p. 86).

2.3.5.2 Beneficiaries of data-driven strategies

Companies that embrace the opportunities for innovation and exploration presented by big data are realizing new value creation and improved firm performance (Sheng et al., 2017). Big data analytics contributes extensively to marketing strategies, as it offers profiling methodologies and recommender systems (Constantiou & Kallinikos, 2015). Online retailers are one of the main beneficiaries of BD (Mazzei & Noble, 2017). According to Mazzei & Noble (2017), Amazon functions as a best-practice example, as the online retailer relies heavily on customer-relevant data that once analysed delivers valuable strategic insights. The e-commerce giant provides an iconic example of how a firm applies data and analytics to evolve strategically. Starting as an e-commerce firm focused on books, Amazon was able to gain information and apply analytics to consumers. The firm captured browsing history, including search terms, books purchased, those placed on wish lists, and the length of time items were viewed. This led to increased selection, improved target marketing, and ultimately an expansion into additional market segments by the e-retailer. Amazon now sells virtually any product on the e-commerce website, including electronics, sports equipment, apparel, and even construction materials (Mazzei & Noble, 2017).

(24)

2.3.5.3 Strategic decision-making and big data

Intezari & Gressel (2017) state that making effective strategic decisions is one of the critical abilities that managers are required to have and develop to lead their organisations in the increasingly volatile and competitive business world. Strategic decisions address ambiguous and complex issues, engage various departments, and involve a high level of organisational resources (Amason, 1996). Because of the extensive uncertainty, ambiguity, and risk associated with strategic decisions (McKenzie et al., 2011), gathering, analysing and considering reliable data and information are critically important in strategic decision-making (Nicolas, 2004). Big data supports strategic decision-making by providing the decision-makers with the relevant information for their decisions (Wang et al., 2016).

Intezari & Gressel (2017) identified four major types of data-driven strategic decisions: Structured decisions based on structured data SD), structured decisions based on unstructured data (SD-UD), unstructured decisions based on structured data (UD-SD) and unstructured decisions based on unstructured data (UD-UD). The authors developed “The decision-data quadrants” (Intezari & Gressel, 2017, p. 78), which is displayed below (see Figure 5).

According to the authors, SD-SD strategic decisions are based on structured data that can be formulated using mathematical modeling (e.g. explanatory models for forecasting or algebraic models for optimization) and advanced analytics to make automated decisions based on gathered and organised data. Operational decisions are a good example for this quadrant as analysed structured data can provide insights e.g. for inventory, sales figures comparisons, and project revues (Intezari & Gressel, 2017).

(25)

Decisions that are made on unstructured data (e.g. Social media) but through pre-defined procedures can be classified as SD-UD. Again, mathematical modeling can be used to transform unstructured data into structured data. Feedback-based decision can be given as an example for SD-UD. Organisations might use pre-defined rules about analysing customer calls and textual feedback to adjust their customer support (Intezari & Gressel, 2017).

UD-SD is when the individual decision-maker or the organisation has access to structured data, but there are little or no clear and pre-defined decision-making procedures to follow, to integrate the data into the decision. Emotional-driven decisions on investments based on structurally analysed reports fall within this quadrant. Tools with reliable and fast reporting systems, are important and can be useful in making UD-SD (Intezari & Gressel, 2017).

Compared to the other decision-data quadrants, UD-UD is the least structured data-based decision. UD-UD relies mainly on human knowledge, experience, interpretation, and insight, rather than mathematical analysis. Social interactions play a critical role in UD-SD and UD-UD types of decision-making; the uncertainty surrounding strategic decisions and the unstructured nature of the data may require more negotiation and discussion between senior managers. Decisions based on documents, such as policies, reports, training material fall in this quadrant (Intezari & Gressel, 2017).

The new way information is generated, aggregated, and presented to managers will significantly alter and reframe the process of strategic decision-making. Regarding the subsequent empirical investigation of this thesis, the theoretical framework is presented in the following part, thus allowing the influence of big data on the strategic decision-making process to be examined in detail. Touching upon this notion, the next chapter provides the theoretical lens to investigate how decision-makers create organisational strategies in the light of big data.

(26)

Theoretical framework

2.4.1 Motivation for the choice of theory

As the previous sections of the literature review illustrate, is there a lack of an analytical lens to investigate the effects of big data analytics on strategic decision-making. Strategic decision-making can be understood as an integral part of the strategy-making process. Strategy and strategy-making itself is a complex phenomenon that presents different levels of academic research streams. The mainstream research in strategic management has dominantly focussed on macro perspectives by conducting multivariate analyses of firms or industry-level effects upon firms’ performances (Jarzabkowski & Spee, 2009). Most related theories in the research streams lack a focus on human actors and their actions, even those that purport to examine the internal dynamics of the firm (Jarzabkowski & Spee, 2009).

The research problem of this thesis requires a theoretical approach that provides a framework for studying the impact of big data analytics on the daily practice of decision-makers and strategists. Strategy-as-practice is in response to growing frustrations with contemporary strategy literature regarding its relevance to practitioners, as it concerns the doing of strategy (Peppard et al., 2014). Research in the s-as-p field emphasizes how people engage in strategy-making and what tools they utilize (Vaara & Whittington, 2012). The centering of the human being in the s-a-p theory corresponds to the demands of the present research problem, since in decision and strategic processes human knowledge and experience play the decisive role, as previous chapters indicate. Therefore, the theory of strategy-as-practice acts as a suitable design of the subsequent empirical

investigation. The exact components of the theoretical framework are explained in the following. 2.4.2 Strategy-as-practice

Practice-based analysis of organisations is becoming increasingly widespread in the management disciplines because of their special capacity to understand how organisational action is enabled (Feldman & Orlikowski, 2011). In addressing strategy-as-practice, the focus is on strategy praxis, strategy practitioners, and strategy practices (Jarzabkowski & Spee, 2009) – the work, workers, and tools of strategy in other words (Peppard et al., 2014). Thus, s-as-p is concerned with the doing of strategy; who does it, what they do, how they do it, what they use, and what implications this has for shaping strategy. From an s-as-p perspective, strategy has been defined “[…] as a situated, socially accomplished activity, while strategizing comprises those actions, interactions and negotiations of multiple actors and the situated practices that they draw upon in accomplishing that activity” (Jarzabkowski et al., 2007, pp. 7–8).

(27)

Furthermore, the s-as-p field has defined its research parameters as studying: practitioners (i.e., those people who do the work of strategy); practices (i.e., the social, symbolic and material tools through which strategy work is done); and praxis (i.e. the flow of activity in which strategy is accomplished) (Jarzabkowski et al., 2007; Jarzabkowski & Spee, 2009; Vaara & Whittington, 2012; Whittington, 2014). These parameters form the theoretical framework for empirical investigation and are explained in more detail below.

2.4.2.1 Practitioners

Practitioners are all those involved in or seeking to influence, strategy-making (Vaara & Whittington, 2012). Jarzabkowski & Whittington (2008) note that strategy practitioners are defined widely, to include both those directly involved in making strategy – most prominently managers and consultants – and those with indirect influence – the policy-makers, the media, and the business schools, who shape legitimate praxis and practices. As previous chapters of the literature review demonstrated, there is an increasing number of technically skilled decision-makers on boards of directors through the advent of big data, providing evidence that boards are placing greater emphasis on technical and data analytical skills in their decision-making (Côrte-Real et al., 2017; Mazzei & Noble, 2017).

2.4.2.2 Practices

Practices refer to the various tools, norms, and procedures of strategy work, from analytical frameworks such as Porter’s Five Forces to strategic planning routines such as strategy workshops (Vaara & Whittington, 2012). According to Jarzabkowski & Whittington (2008), practices involve the various routines, discourses, concepts and technologies through which this strategy labour is made possible – not just obvious ones such as strategy reviews, but also those embedded in academic and consulting tools (i.e., Porterian analysis, hypothesis testing, etc.) and more material technologies and artifacts (i.e., PowerPoints, flip-charts, etc.). Authors agree that the growing number of relevant data sources require significant modification, if not replacement of existing strategic models and tools (Constantiou & Kallinikos, 2015). Novel software, such as visualization tools (e.g., Tableau, Microsoft Power BI, and Qlik Sense) convert the data (i.e., symbols, mostly numbers) into visual models (e.g. Graphs, diagrams, scatter plots) and thus assist greatly in strategic decision-making. The use of these tools in IKEA’s strategic decision-making processes would confirm the authors’ assumptions.

(28)

2.4.2.3 Praxis

Praxis refers to the activity involved in strategy-making, for example, in strategic planning processes or meetings (Vaara & Whittington, 2012). As there are different nuances in the way that the term praxis is used, Jarzabkowski & Spee (2009) identify three levels of praxis within the strategy-as-practice field. Micro refers to strategy praxis at levels of the individual or group’s experience of a specific episode, such as a decision, meeting, or workshop. Meso refers to strategy praxis at the organisational or sub-organisational level, such as a change program, or a strategy process, or a pattern of strategic actions (Balogun & Johnson, 2005). Macro refers to strategy praxis at the institutional level, which is most typically associated with explaining patterns of action within a specific industry (Lounsbury & Crumley, 2007). These three levels are consistent with the three levels of decision-making proposed by Merendino et al. (2018), which were highlighted in the literature review.

2.4.2.4 Strategy-as-practice and big data

The strategy-as-practice research field has been reaching out for a greater understanding of materiality (Vaara & Whittington, 2012), a domain in which technologies, such as big data and big data analytics are particularly relevant (Whittington, 2014). The joint research stream focusses on the roles of technologies in strategizing practices, as Whittington (2014) outlines. In particular, it examines how technologies influence the strategy-as-practice framework of practitioners, practices, and praxis (Whittington, 2014). This perspective helps us to better understand big data analytics in the strategy as practice context. Henfridsson & Lind (2014) contributes to the joint stream with their notion of ‘technology-mediated practices’, while Arvidsson et al. (2014) consider the role of Information Technology ‘within’ practices. Practitioners are also impacted by big data analytics, as the increasing number of Chief Data Officers (CDO), Chief Information Officers (CIO), and Chief Analytics Officer (CAO) makes clear (Côrte-Real et al., 2017; Mazzei & Noble, 2017). Although this joint research stream calls for further research, little research has been done in this area so far (Whittington, 2014).

(29)

3 Methodology

___________________________________________________________________________

This section presents this study´s research philosophy, research purpose, research approach, research strategy as well as the methodological techniques used to gather and analyse data. Lastly, credibility and trustworthiness are elaborated on.

Research philosophy

One can divide research philosophy in two paradigms, positivism and the interpretivist paradigm (Saunders et al., 2015). The interpretivist paradigm is mainly used as a research philosophy viewed as apt for social science which follows the philosophy that all answers cannot be binary, but rather up for interpretation (Saunders et al., 2015). The philosophy suggests that questions concerning the social reality cannot be based completely on raw data, but rather include the values and beliefs of the authors and of the population that the data has been collected from (Saunders et al., 2015). The interpretivist view is based on the assumption that as all researchers are human, perceptions and interpretations will be different between different actors as all researcher are social actors, meaning that some answers to research questions are up for interpretations, i.e. all answers are not black and white (Ketokivi & Choi, 2014; Saunders et al., 2015).

On the other hand, the positivist research philosophy is aimed towards the objective aspects of research (Saunders et al., 2015). Although, all research should aim at being as objective as possible, the positivistic view is focusing on testing hypothesis, to confirm or deny a research question (Saunders et al., 2015). The difference in the collection of data is that the positivist view subtracts all values and beliefs and rather focus on independent data that only can be established by testing various hypothesises (Saunders et al., 2015). Saunders et al., (2015) further argues that sample selection and choice of methods becomes easier once having developed a suitable research philosophy.

As this research is focusing on understanding a phenomenon, a change in the strategic decision-making process, the interpretivist paradigm has been chosen. It is identified that by only taking raw data into account when researching the decision-making process, difficulties will arise when trying to interpret the data collected since it will not be binary. This is mainly due to the fact that the research problem is not binary but include values and beliefs within the employees at IKEA. By following the interpretivist view, higher degree of subjectivity can be used and by doing so account for the human aspect in the result to a larger extent (Saunders et al., 2015). Additionally, by using an interpretivist approach, the usage of a qualitative research is encouraged to be able to collect a higher depth of the data. Following the interpretivist approach, the researchers will be able to make

(30)

their own interpretations of the data and account for soft values in the interviews, such as tone of voice, body language and facial expression (Saunders et al., 2015). Furthermore, as there are few binary values other than identifying how the mandate of making a decision has changed, the philosophy of this thesis needs to be interpretivist in order for the researchers to be able to correctly analyse the phenomenon from various points of view.

Research purpose

Defining the purpose of a research is important to understand how the research has been conducted. According to Saunders et al. (2015), one can define three different research approaches that helps understanding the conduction of the research. The three approaches defined are descriptive, explanatory, and exploratory (Saunders et al., 2015). Descriptive research can be viewed as a prior step to the other two approaches of research due to the fact that it is not essential to finalize with a discussion (Saunders et al., 2015). An exploratory research is used to explain a new phenomenon or attempting to understand a situation that previously has not been researched (Saunders et al., 2015). However, the final objective with an exploratory research does not have to end up with a final answer, but rather open to new questions or areas in which could be subject to research (Saunders et al., 2015). The final approach, explanatory research, aims to understand a problem and study a specific phenomenon to decide why a specific cause is happening as a result of an action (Saunders et al., 2015).

The research at hand has the purpose of understanding how big data analytics has changed the strategic decision-making processes within an organisation. The approach will include conducting a qualitative research to understand what impact the implementation of big data analytics has had for an organisation. One could therefore argue that the nature of this research becomes exploratory, as the research intends to identify changes around a phenomenon and create a discussion regarding how practitioners, practices, and praxis are affected as big data analytics has been utilized in company’s strategic decision-making procedure. As this research is of exploratory nature, these areas will be investigated to collect data regarding how IKEA has been impacted when transforming into a data-driven company.

Research approach

The research approach can be understood as the plan and ways of working throughout a research (Saunders et al., 2015). It decides the approach that the researchers will utilize and sets the framework for how the research will be conducted (Saunders et al., 2015). One could argue that there are two major research approaches, with a third approach forging a bridge between the two

(31)

main approaches (Saunders et al., 2015). The deductive approach is preferred when applying a test to a research question or wanting to test a hypothesis (Saunders et al., 2015). Collins & Hussey, (2014) explains the deductive theory as the “testing theory” which is the preferred method used in studies that laws stands as the ground of reasoning. On the contrary, the inductive approach can be utilized as the researchers aim to firstly collect data to progress a theory that is resulting of the research conducted (Saunders et al., 2015). Saunders et al. (2015) explains a practical example of utilizing the inductive approach by conducting a study at a company where employees are being shadowed or interviewed to collect data. The author further explains how the data is analysed to be able to categorize issues and various phenomena within that company (Saunders et al., 2015). The final step would be to develop a theory regarding the identified phenomenon (Saunders et al., 2015).

Abduction is the third research approach that can be adopted whilst planning for a study. As previously mentioned, the abductive approach is bridging the gap between the two earlier mentioned approaches (Saunders et al., 2015). One could describe the differences between the three by stating that as the deductive theory aims to assess both theories and hypothesis to get a definite answer (Ketokivi & Choi, 2014). The inductive approach tires to develop new theories and solve problems identified within various areas (Ketokivi & Choi, 2014). The abductive approach, however, has the objective of testing current theories whilst resulting in creating a new hypothesis to a research question or further discuss current theories (Ketokivi & Choi, 2014). One could argue that this research intends to use elements of both the inductive research approach as well as the abductive approach. One the one side, this research is utilizing an exploratory research that intends to investigate how a company’s strategic decision-making process has changed since implementing big data analytics. The objective will be to investigate a theory, i.e. the strategy-as-practice theory, and establish what has changed and how this change can be performed efficiently. On the other hand, the research intends to develop a hypothesis with collected data from only one company. Due to the singularity of companies’ interview, the objective of this research will be to create a generality of how the decision-making process has changed, rather than how it is. By doing so, the research creates further hypothesis of how the decision-making process is changing in other sectors and organisations. Furthermore, the contribution to science of this thesis will be to shed a light on the existing research and offer insights into a phenomenon as parts of both the inductive and abductive approach are implemented.