Exploring Mid-Market Strategies for Big Data Governance

Exploring Mid-Market Strategies for Big Data

Governance

Section 1: Foundation of the Study

Background of the Problem

Data generation has increased exponentially in recent years (Bello-Orgaz, Jung, &

Camacho, 2016), with no signs of this trend stopping. Bello-Orgaz et al. (2016) estimated

that globally, 2.5 exabytes of new data are being generated per day, and the Government

Accountability Office (2017) estimated that by 2025, there would be between 25 and 50

billion devices connected to the Internet and generating data. Even with the vast amount

of data available to them, organizations effectively use less than 5% of their available

data (Zakir, Seymour, & Berg, 2015). This emergence of big data has introduced data

management challenges involving processing speed, data interpretation, and data quality

for organizations that wish to consume complex information (Lee, 2017). The traditional

methods, frameworks, strategies, and tools for data governance and analysis are outdated

and no longer adequate for processing the vast amount of data available to organizations

today, thus making current strategies ineffective for handling big data (Bello-Orgaz et al.,

2016). Big data analytics challenges arise from issues relating to data that are too vast,

unstructured, and moving too fast to be managed by traditional means (Zakir et al., 2015).

Companies of all sizes are dealing with this new big data environment and

struggling to determine how best to analyze big data as a critical driver of strategic

business decisions. Larger companies have more resources to direct toward this problem,

but mid-market organizations face similar issues with less available capital to apply to the

problem. To remain competitive, however, it is just as critical for them to find ways to

address this data deluge that faces companies of all sizes.

Problem Statement

Contemporary outdated data processing systems are unable to handle the data

deluge of exponentially increasing amounts of data that we are generating daily

(Sivarajah, Kamal, Irani, & Weerakkody, 2017). More than 40% of organizations are

currently challenged to attract and retain skilled data scientists, while by 2020, the U.S

alone will need more than 190,000 skilled data analysts (Mikalef, Giannakos, Pappas, &

Krogstie, 2018). The general IT problem is that there is a lack of knowledge regarding

data governance principles for analyzing big data. The specific IT problem is that some

data scientists lack big data governance strategies that provide a holistic perspective of

those data for making decisions.

Purpose Statement

The purpose of this qualitative multiple case study was to explore the big data

governance strategies employed by data scientists to provide a holistic perspective of

those data for making decisions. The population for this study was data scientists

employed in three mid-market companies in the greater Salt Lake City, Utah area who

have strategies to govern big data. The findings from this study could potentially affect

social change by creating additional jobs in mid-market companies. Job creation would

be accomplished through increased capability of these companies to have a better

understanding of meaning of data which they gather and use to make strategic and

operational business decisions daily. This increased and more targeted information could

lead to increased profitability for these mid-market companies, which in turn could lead

to increased jobs available in these communities. These new jobs could elevate the

midmarket organization’s community by providing better pay for residents and improving

their lives in general.

Nature of the Study

This study used a qualitative method to explore the big data governance strategies

employed by data scientists to provide a holistic analysis of those data for making

decisions. Qualitative studies are useful when investigating peoples’ views or human

behavior, or when attempting to find out how well something is performed, but for which

quantitative data does not offer a complete picture (Kelly, 2017). Because strategies for

data governance involve understanding how well those governance activities are

performed, the qualitative method was appropriate. Quantitative studies can demonstrate

trends and correlations, but are not as capable of providing explanations or reasons

(Kelly, 2017). Quantitative studies are focused on predictions and theories based on

cause-and-effect relationships, which are testable and generalizable (Gehman et al.,

2018). Since this study was not attempting to prove a hypothesis, a quantitative approach

was not appropriate. A mixed methods approach would involve combining quantifiable

data with an in-depth understanding of participants’ behavior and actions (McCusker &

Gunaydin, 2015). Because strategies related to data governance do not involve

quantifiable statistical data, a mixed methods approach was not appropriate.

I used a multiple case study methodology for this study. A case study

methodology allows a deep examination of an issue or phenomenon in its natural setting

(Arseven, 2018). Because this study focused on strategies used within organizations that

are using big data for decision-making, this design approach was warranted. An

ethnographic approach requires embedding the researcher in a cultural context to observe

the behaviors of a people in a particular cultural context (Watson, 2018). I did not select

the ethnographic design because this study was not investigating any specific culture,

customs, or social group. Thus, it was not appropriate for this study. The

phenomenological design was not appropriate because it involves how participants live

through, interpret, or experience the topic under research (Sousa, 2014). Because this

study was not investigating perceptions of participants, the phenomenological approach

would not have been appropriate.

Qualitative Research Question

What big data governance strategies do data scientists employ to provide a holistic

perspective of those data for making decisions?

Interview Questions

Demographic Questions

1. What is your role in data gathering and analysis?

2. Please describe your background concerning big data analysis.

3. How long have you been involved in big data analysis?

Interview Questions

1. How do you ensure that you are analyzing all available data for any particular

model?

2. How do you determine appropriate sources of data for any specific model or

analysis?

3. What processes are in place to ensure that all applicable data are available to you

when needed for analysis?

4. What countermeasures are in place for risk mitigation regarding the validity of the

analysis that you perform (i.e., the risk of not having current data, or of incorrect

correlation of data from various sources)?

5. What have you found to be most successful in building data models based on big

data from various sources?

6. Why do these processes give you confidence that you are analyzing all applicable

data?

7. How do these processes ensure the protection of the data?

8. What controls are in place to ensure the privacy of the data?

9. What other issues are there that we should discuss as relating to big data analysis?

10. What other areas are there left to discuss?

Conceptual Framework

I selected the organizational information processing theory (OIPT) as the lens

through which I examined data governance strategies used by data scientists to provide a

holistic perspective of those data for making decisions. Jay R. Galbraith introduced the

OIPT in 1974 (Galbraith, 1974). The central concept of the theory is that organizations

need relevant information to make decisions. Three concepts primarily comprise the

ability to use that information for decision-making efficiently: the need within an

organization to process information, the organization's inherent ability to process

information, and the gap between the two (Premkumar, Ramamurthy, & Saunders, 2005).

The OIPT aligns well with this study because of the direct association between the

amount of information that is available to an organization and the ability for that

organization to process that information to improve the decision making within that

organization. Data scientists are starting to employ novel strategies in terms of

governance and analysis of big data to transform large quantities of information into

valuable insights for strategic business decision-making processes. By using the lens of

the OIPT, I was able to assess strategies used by data scientists to determine whether

these strategies are increasing the ability of the organization to process information and

thereby reduce slack resources within the organization. By viewing governance of big

data through the lens of the OIPT, I focused on strategies of data scientists to properly

govern the vast amount of data available to them as a mechanism to reduce the

uncertainty of information analysis to enable better decision-making throughout their

organizations.

Definition of Terms

Data/Information Governance: An emerging discipline involving managing,

processing, and controlling information throughout an organization. It encompasses

processes, procedures, and rules involving the management of information throughout an

organization (Mullon & Ngoepe, 2019).

IT Governance: The set of processes and oversight within an organization that

ensures that investments in IT generate business value, mitigate risks associated with IT

initiatives, and operate with accountability and traceability for those who are funding the

IT resources (Grossman, 2018).

Assumptions, Limitations, and Delimitations

Assumptions

Walsh (2015) described assumptions as unconscious beliefs that the researcher

accepts as accurate which can produce biases in terms of both perceptions and thoughts. I

assumed that participants answered my questions to the best of their abilities. I also

assumed that individuals within organizations that I interviewed were subject matter

experts within their organization based on job title and work performed. I assumed that

documentation these organizations provided to me was official and reflected their current

processes and procedures.

Limitations

Teichler (2014) defined limitations of a study as those aspects of the study that are

beyond the control of the researcher, despite their best efforts. Limitations of this study

are primarily centered around the fact that I performed a qualitative case study, and as

such, findings may not be able to be extrapolated beyond organizations selected for the

case study. Qualitative studies in general have inherent limitations since qualitative

assessments are limited by incomplete understanding of judgments made by the

researcher and lack of consensus among researchers regarding the rigor of deriving those

judgments (Cook, Kuper, Hatala, & Ginsburg, 2016).

Delimitations

Delimitations are boundaries defined by the researcher that guide the study itself

(Kongnso, 2015). This study was geographically limited to the greater Salt Lake City,

Utah, area. Additionally, participants for this study were limited to those who were

involved directly in data modeling and data analysis strategy and processes within case

study companies, as well as those who were involved in the design of data governance

processes and practices. Specific roles or job titles for this responsibility may differ

between organizations. Finally, the organizations that I selected for the case study were

required to have some experience analyzing big data.

Significance of the Study

Contribution to Information Technology Practice

This study may provide strategies to data scientists in mid-market companies who

lack strategies for governing big data. This study may provide value to the overall IT

practice by uncovering insights about successful data governance practices that are in use

within mid-market organizations. Overall contributions to IT practice may potentially

include additional insights into best practices that mid-market organizations have adopted

and from which they have found success as they struggle to deal with the deluge of

information that is available to them for making critical strategic business decisions.

Kemp (2014) indicated that nearly 85% of Fortune 500 organizations are unable to use

their data effectively, and a structured approach to data governance, policy, and strategy

is a critical factor for project success. Mid-market companies have a similar need to use

data available to them for strategic business decisions effectively, and they can typically

react and implement new strategies more rapidly than their larger competitors.

Implications for Social Change

The potential for social change involves the increased ability of companies to

meet the needs of their customer base due to an increased understanding within the

company of the meaning of data which they gather and use to make strategic and

operational business decisions daily. This improved customer focus could lead to higher

profitability, which in turn would enable these organizations to hire more employees and

increase wages. These overall improvements in businesses could lead to overall elevation

of residents in these communities by allowing them to earn more money and improve

their way of life. Also, better overall management could lead to better employment

opportunities for all types of employees, as well as improved customer experiences for

customers of mid-market companies. Effective use of big data could benefit customers by

providing more specific and targeted solutions to their needs for small companies that are

part of their local communities. In sum, the ability for data scientists to effectively use big

data in mid-market companies could be an equalizing force that could benefit customers,

employees, and owners of all types of mid-market companies.

A Review of the Professional and Academic Literature For this study,

I focused on three basic concepts when researching the professional and academic

literature: issues involving the adoption and use of big data as a reliable and viable

technology, applicability of the OIPT as a conceptual model for applying big data

governance in the enterprise decision-making process, and the use of the case study

methodology to study this topic. I searched for academic literature with Google Scholar,

Science Direct, IEEE Explore, Research Gate, and the Walden Library as primary sources

for academic research. I used the following search terms to initiate my search at these

websites: big data governance, big data challenges, big data decisions, big data benefits,

big data roles, big data security, big data ethics, small business big data, data governance,

big data analytics, data scientist roles, data analysis roles, big data job description

expectations, OIPT, organizational information processing theory, Galbraith, enterprise

decision, small business decision making. I also employed the suggested search links that

appeared in the search results on these various sites to continue my search. As I studied

the literature, I frequently used reference lists of cited articles as alternate sources. I

focused my research on contemporary publications, with the intent of using articles

published in 2016 or later. I also researched seminal articles and publications when

researching the OIPT and case study topics.

The content of the literature included in this review includes history,

contemporary discourse, and contrasting views of the OIPT as well as all aspects of big

data governance. I selected seminal sources where applicable and available. I focused my

review of literature on the data governance aspects of big data, including expectations of

big data analytics, quality and volume of data, challenges due to evolving from traditional

to big data management, ethical and privacy concerns, organizational and human resource

concerns, and IT and data governance in general. I intentionally left the discussion of

specific toolsets such as Hadoop and R out of this study as these are analytic technologies

and not explicitly related to the overall governance of big data. I reviewed more than 250

sources in total, eventually eliminating more than 110 sources as not relevant,

duplicative, or repetitive. In my review, I did not encounter any studies that specifically

addressed the governance of big data and its effect on strategic decisionmaking in the

manner that I am studying.

To document this literature review, I start with a discussion of the conceptual

framework and then proceed to relevant literature regarding big data. The review of big

data literature starts with a definition of big data, leading to a discussion of its benefits

and uses. I then discuss some of the challenges associated with big data, which leads to a

discussion of related technologies such as the Internet of Things (IoT) and cloud

computing. I end the literature review by discussing IT and data governance as relates

specifically to big data initiatives.

This literature review consists of 105 articles relating to the OITP and big data and

the associated technologies such as the IoT, cloud computing, data governance, and

information governance. Eighty-two percent of articles reviewed were published between

2016 and 2020. Ninety-one percent of these articles were peer-reviewed as verified

primarily by searching Ulrich’s Periodicals Directory and reviewing journal information

directly in many cases.

Application to the Applied IT Problem

The purpose of this qualitative multiple case study is to explore big data

governance strategies employed by data scientists to provide a holistic perspective of

those data for making decisions. In the next sections, I will provide a thorough

discussion of the conceptual framework and current literature regarding big data.

Conceptual Framework

The conceptual framework for this study is the OIPT. I will discuss the OIPT

indepth as well as provide an analysis of both supporting and contrasting theories.

OIPT. OIPT was a seminal theory on organizational design, which contributed to

Galbraith later becoming known as the father of modern organizational design theories

(Galbraith, 2017). It was Galbraith’s mentor, James D. Thompson, who first introduced

him to the concept of viewing organizations as systems and studying the inherent need for

organizations to process information (Galbraith, 2017). Galbraith’s theory was also

heavily influenced by Herbert Simon, a Nobel Prize-winning economist who, in

conjunction with James Marsh, introduced the concept of slack resources within an

organization to aid in information processing (Galbraith, 2017). The basis of the OIPT is

that organizations are comprised of individuals who must deal with both predictable and

unpredictable information (Galbraith, 2017). Tushman and Nadler (1978) added that the

OIPT directly applies to the gathering, interpretation, and synthesis of information within

an organization for decision-making. Feurer, Schuhmacher, and Kuester (2019) explained

that the fundamental basis for OIPT is centered around the uncertainty and complexity of

tasks to be completed in a task workflow and is based on the concept that organizations

should be designed to enable decision-making. All of Galbraith’s major writings also cite

the contributions of Alfred D. Chandler, which demonstrates that Chandler’s work

influenced Galbraith’s connection of an organization’s strategy with its structural form

(Galbraith, 2017).

Zelt, Recker, Schmiedel, and vom Brocke (2018) articulated that the OIPT

describes organizations as information processing systems that collect, process, and

distribute information, and Allegrini and Monteduro (2018) added that greater uncertainty

is associated with the need to process more information during task execution. Galbraith

(1974) explained that the greater the uncertainty of a task, the more information must be

processed by decision-makers to execute that task. Gupta, Kumar, Kamboj, Bhushan, and

Luo (2019) explained that a lack of information within an organization leads to

uncertainty, and Tushman and Nadler (1978) explained that uncertainty was the

difference between information processed and required to complete a task. Foerstl,

Meinlschmidt, and Busse (2018) theorized that the OIPT evolved in response to

organizational design problems stemming from the size-induced complexity of

organizations in the 1970s.

Tushman and Nadler (1978) identified management as collecting, processing, and

combining relevant information for organizations. Bartnik and Park (2018) added that

management’s role is to reduce both uncertainty and equivocality by clarifying mission

statements and organizational priorities for organizations. Gao, Liu, Guo, and Li (2018)

explained that merely increasing information available does not resolve uncertainty as

individuals do not have unlimited processing capacity and will reach information

overload.

Zelt, Recker et al. (2018) clarified that the OIPT builds on prior contingency

theories by aligning with the concepts that the organizational structure needs to fit

variables both inside and outside of the organization. Duvald (2019) added that the OIPT

is one of many contingency theories that identify that while there is no one best way to

organize, not all ways of organizing are equally valid. DiMaggio and Powell (1983)

established institutional theory in 1983, which establishes that processes that can establish

practices, policies, and norms. Chrisman, Chua, Le Breton-Miller, Miller, and Steier

(2018) indicated that governance in general is usually the domain of institutional theory.

The application of institutional theory focuses mainly on coercive, mimetic, and

normative pressures that shape behavior within an organization and not on the effect on

strategic decision making (Chrisman et al., 2018).

Duvald (2019) explained that equivocality arises from the existence of multiple

and often conflicting interpretations of data, which can only be mitigated by additional

processing mechanisms rather than simply gathering more data. Zelt, Recker, et al. (2018)

added that increasing equivocality leads to a need for systems to interpret and assign

meaning to the information being processed.

Galbraith (1974) theorized that as task complexity increases, organizations can

deal with increased uncertainty of task completion in only four ways, two of which

address the reduction of the information needing to be processed, and the other two

addressing the possibility of increasing the overall capacity to process information (see

Figure 1). Premkumar et al. (2005) added that the OIPT has three primary elements: the

need within an organization to process information, their inherent ability to process

information, and the gap between the two. Galbraith (2012) explained that organizations

can deal with this gap in one of two ways: they can either increase the capacity of the

organization to process more information, or they can decentralize the interdependence

on that information. Cao, Duan, and Cadden (2019) defined the information processing

capability of an organization as their capacity to capture, integrate, and analyze data and

information and use those insights in the context of organizational decision-making.

Wang, Xu, Fujita, and Liu (2016) added that combining heterogeneous sources of data

into new information sets is one way to reduce uncertainty. Gao et al. (2018) theorized

that using the lens of the OIPT, it becomes clear that increasing access to information will

only serve to create information overload rather than benefiting decision-making

processes. Ever-increasing amounts of information coupled with the inability of the

organization to process that information lead to organizations adopting two typical

strategies for coping with the inherent uncertainty that ensues: developing buffers to

reduce the effect of uncertainty and implementing structural mechanisms and

information- processing capabilities to enhance information flow and thereby reduce

uncertainty (Premkumar et al., 2005).

Figure 1. Organizational design strategies. From "Organizational Design, an Information

Processing View" by J. R. Galbraith, 1974, Interfaces, 4(3), p.30. Reprinted with

permission.

While the concept was quite simple, Burton, Obel, and Håkonsson (2015)

indicated that Galbraith’s subsequent work that built upon this theory became the

foundation for how to make matrixed organizations work well. Because of the lack of

vertical IT systems to analyze information, the only option was to create matrix

organizations to process information and make business decisions. Feurer et al. (2019)

explained that one of the general concepts of the OIPT is that as lateral organizational

relationships are established, the number of decisions referred upward is reduced, which

then leads to more effective decision-making. Obel and Snow (2014) posited that the

theory of organizational design allowed many areas of research to be brought together to

create a model of organizational design with significant predictive power and that it is

even more relevant in today’s world of the dramatic increasing availability of information

in the form of big data. Intezari and Gressel (2017) added that the ability of

organizations to make strategic decisions amidst uncertainty and ambiguity is predicated

their ability to learn and reconfigure their knowledge base continuously.

Additional research regarding the OIPT led Galbraith to investigate matrix

organizations as a means to decentralize decision making, which set the foundation for

Galbraith’s most famous theory on organizational design, named the star model

(Galbraith, 2017). Galbraith (2014) demonstrated the potential impact of big data on an

organizational design using the star model, as shown in figure 2. The star model included

five dimensions of organizational structure: strategy, structure, processes, people, and

rewards (Miterev, Turner, & Mancini, 2017). Obel and Snow (2014) indicated that the

star model became the most recognized and widely accepted organizational design model.

Galbraith based the star model on his belief that organizations are inherently made up of

individuals who need to process information to accomplish the work of that organization

(Galbraith, 2017).

Figure 2. Big data’s impact on an organization. From "Organizational design challenges

resulting from big data" by J. R. Galbraith, 2014, Journal of Organizational Design, 3(1),

p.12. Reprinted with permission.

Contemporary view. Contemporary thought regarding the OIPT shows the

theory has remained relatively consistent since its inception with incremental additions

and developments. Kroh, Luetjen, Globocnik, and Schultz (2018) explained that the

evolution of IT systems was evidence that the original theory accurately described how

better processing of information within an organization could benefit task completion,

and the increased ability of employees to process information using technology provides

a mechanism for improving decision-making within an organization due to the improved

ability to process more copious amounts of information farther down within the

organization.

Obel and Snow (2014) posited that Dr. Galbraith’s theory of organizational

design is even more relevant today than it was when Galbraith developed it in 1974

because of the dramatically increasing availability of information in the form of big data.

Park, Sawy, and Fiss (2017) highlighted that the ability to process information for

organizational decision making had been studied extensively, but little focus has been

placed on the details within the IT component of that decision-making process. Cao et al.

(2019) articulated that Galbraith’s work was later adopted to address decision-making

within an organization. Jia, Blome, Sun, Yang, and Zhi (2020) clarified that researchers

Tushman and Nadler built upon Galbraith’s work by interpreting that said organizations

are inherently information- processing systems that are intrinsically- programmed to

manage uncertainty by gathering, processing, and acting on information from within their

environment. Hwang, Kim, Hur, and Schoenherr (2019) added that through the OIPT

Galbraith highlights the essential function of an organizational structure, which is to

facilitate the collection, analysis, and distribution of information to reduce uncertainty

within the organization.

Figure 3. OITP key concepts.

Galbraith (2012) hypothesized that big data would add another dimension to the

organizational information processing needs within organizations due to the constant

interplay of increasing complexity and interdependence of information and systems

within that organization. Zelt, Schmiedel, and vom Brocke (2018) articulated that the

application of OIPT leads to an understanding of the factors influencing an organization's

information processing capacity. Galbraith (2012) predicted that big data would become

that technology that would finally allow for the increased information processing

alternative to becoming an effective alternative. Galbraith became fascinated by the way

companies were using big data to gain valuable insights into both their organizational

decision-making capability and to learn more about their customers (Galbraith, 2017). He

believed that big data would become the next significant dimension in organizational

design by providing a digital function similar in power and importance to organizations

existing organizational structure (Galbraith, 2014).

The literature identified some limitations of the OIPT. Haußmann, Dwivedi,

Venkitachalam, and Williams (2012) identified that Galbraith focused mainly on

technical aspects of information processing within an overall organizational design and

did not address the individual and interpersonal effects of information processing

thoroughly. Kroh et al. (2018) articulated that the OIPT emphasizes structural

mechanisms when addressing information flows and task uncertainty. Haußmann et al.

(2012) further clarified that the reduction of uncertainty is not the only factor to consider

when assessing IT adaptation for improvement of decision making within an

organization. Kroh et al. (2018) also stated that information processing capabilities enable

organizations to receive external input and translate that into usable information within

the organization. Haußmann et al. (2012) clarified that Galbraith mentions the external

perspective but does not include this as essential to the overall OIPT concepts.

Despite the limitations, Haußmann et al. (2012) pointed out that the OIPT is

considered more than a theoretical model to serve academia and remains a theory with

wide adaptation and practical implications within the IS community. The OIPT is very

applicable to big data, as shown in figure 3. Before his death, Galbraith was working on

another book that was to focus on how big data would affect decision making within

organizations (Galbraith, 2017). Obel and Snow (2014) highlighted that his previous

articles on big data provided critical insights about improving the scale and speed of

dissemination of information within an organization. Galbraith believed that there were

three main hurdles to overcome for big data to truly demonstrate valuable insights for an

organization: the first hurdle was the power struggle and friction caused by big data

within the current organizational structure. The second hurdle was the creation of a digital

information and decision process. The final hurdle was the ability of an organization to

invest in digital resources required to develop analytic capabilities across the entire

organization (Galbraith, 2017).

Analysis of Supporting Theories

Absorptive capacity theory (ACT). The ACT was introduced in 1990 by Cohen

and Levinthal as a firm’s ability to identify, consume, transform, and apply valuable

external information (Cohen & Levinthal, 1990). The capacity of an organization to

assimilate such information is positively related to improved performance (Cohen &

Levinthal, 1990). Conceptually, the ACT is similar to the information processing theory

(IPT), but the ACT is focused on the organization level while the IPT is focused on the

individual level. Butler and Ferlie (2019) identified that Cohen and Levinthal have been

criticized for narrowly defining the ACT on R&D only and that subsequent definitions

have expanded it to a broader definition centered around the acquisition of knowledge, in

general, increasing a firm’s overall performance. This lack of clear definition has led to

an ambiguous characterization of the ACT in contemporary terms with some scholars

focusing on organizational learning, others on innovation, and yet others on

resourcebased knowledge views of the theory (Chaudhary & Batra, 2018). While the

ACT provides a lens for organizational learning from external sources, the OIPT is a

more appropriate lens to apply to this study because of the clearly understood definition

of the OIPT and the direct connection to decision making.

Organizational information theory. Organizational information theory (OIT)

was first published in 1979 by Karl Weick and updated in the second edition of his book

published in 2015 (Weick, 2015). One of the central tenets of the OIT was the premise

that organizations that operate in an information-rich environment have a great need to

make sense of equivocal information (Weick, 2015). Weick defined a process called

retrospective sensemaking, which entails seeking to understand what is currently

happening by identifying what was done in the past, understanding why those actions

were taken, and determining if those actions were adequate for the institution’s goals

(Weick, Sutcliffe, & Obstfeld, 2005). Weick introduced the concepts of ambiguity and

uncertainty and posits that for an organization to progress, they must reduce or eliminate

ambiguity and uncertainty (Chadwick & Pawlowski, 2007). This concept of reducing or

removing uncertainty is like the central tenet of the OIPT but without the clarity of how

to reduce that uncertainty. The OIT is focused more on the concept of sensemaking

regarding past decisions made, while the OIPT is more concerned with understanding

how to reduce uncertainty by gathering and utilizing more information.

Analysis of Contrasting Theories

Information processing theory (IPT). The IPT was established in 1956 and

attempted to identify and explain how information is processed, captured, and retrieved

by individuals (Mitchell, Mitchell, & Mitchell, 2009). The IPT attempted to address the

way information is acquired, stored, and retrieved from human memory (Mitchell et al.,

2009). The central concept of the IPT was that as the volume of information continuously

increases, it becomes more difficult for individuals to process due to their limited

cognitive processing ability, and they experience information overload (Gao et al., 2018).

IPT focused more on the cognitive capabilities of the human brain when processing

information, whereas the OIPT focused more on the impact of information processing

within an organization. The IPT may be an appropriate lens to view the creation and

dissemination of information from big data on an individual level, but the OIPT is a much

more appropriate lens for viewing governance of big data at an organizational level.

Organizational knowledge creation (OKC). Nonaka, Byosiere, Borucki, and

Konno (1994) introduced the OKC theory to define how knowledge passes throughout an

organization. The OKC is a theory describing knowledge creation that extends both up

and down the organizational structure from middle management based on both explicit

and tacit communications (Nonaka, Hirose, & Takeda, 2016). Milosevic, Bass, and

Combs (2018) explained that the OKC highlights the interrelated phases of knowledge

emergence and knowledge formalization within an organization. Kemp (2014) addressed

this concept concerning big data when he explained that there is a gap between the

amount of data that organizations can gather and the ability of that organization to

leverage that information in meaningful ways. Akbar, Baruch, and Tzokas (2017) noted

that the OKC focused on the logic of appropriateness, which is a situation-driven

response that evolves through socialization and discovery. While applicable to big data,

the OKC focused more on the creation of knowledge within an organization, while the

OIPT is more focused on the gap that exists between the ability to obtain information and

the needed information to complete tasks within the organization.

Big Data Governance

Definition of big data. The term big data is not yet formally defined and was used

inconsistently in the literature. The actual definition was still evolving with some people

defining big data by what it is, while others tried to define it by what it does (Gandomi &

Haider, 2015), and some defined it by the processes required to leverage big data

(Mikalef, Pappas, Krogstie, & Giannakos, 2017). Wang et al. (2016) categorized the

various definitions of big data as being oriented around the product, process, cognitive, or

social aspects of big data. As a case in point, Wamba et al. (2017) indicated that big data

is actionable and delivers sustained value, measured performance and provides a

competitive advantage, while Sivarajah et al. (2017) defined big data by the 7 V’s of high

volume, high variety, low veracity, high velocity, high variability, effective visualization,

and high value. Lee (2017) referred to additional dimensions of big data as including

complexity and decay, meaning that often, data elements of big data require immediate

processing to be of value. Kemp (2014) also referred to the value of the data as a critical

element to the definition of big data. Herschel and Miori (2017) stated that big data refers

to those activities around capturing, storing, analyzing, and acting upon information that

is gathered by both humans and devices, where networks transmit that information.

The most common definition of big data had less to do with the actual amount of

data, and more to do with the attributes of those data. Most researchers agreed that this

includes the following elements, often referred to as the 5 V’s: huge volume, high

velocity, wide variety, low veracity, and high value (Jin, Wah, Cheng, & Wang, 2015).

Bello-Orgaz et al. (2016) stated that these attributes were first identified in 2001 as the 3

V’s, namely: high volume, high velocity, and a wide variety. Gandomi and Haider (2015)

added that other attributes of veracity, variability, and value were added later. Song and

Zhu (2016) proposed that while variety and veracity are challenging to deal with, the

value was by far the most critical dimension of big data. Gartner added a qualifier to this

definition in 2012 that indicated that big data also required new forms of processing to

gain insights from those data (Bello-Orgaz et al., 2016). Sivarajah et al. (2017) indicated

that the most recent definition now has 7 V’s: Volume, Velocity, Variety, Variability,

Veracity, Visualization, and Value.

Big data definitions in literature were also commonly used in conjunction with

either the size of the data set or with the technology used to enable its utilization. Patel,

Roy, Bhattacharyya, and Kim (2017) indicated that big data are anything over gigabytes

of data, while others such as Bello-Orgaz et al. (2016) indicated that big data refers to

terabytes or petabytes in size. Lee (2017) stated that the threshold for Big Data is a

minimum of 1TB but added the qualifying sentiment that the minimum threshold would

evolve as technology is developed to handle more copious amounts of data. De Mauro,

Greco, and Grimaldi (2015) pointed out that the term big data is also often used in

conjunction with the technology, such as Hadoop, that is used to enable the use of big

data. Patel et al. (2017) defined big data simply as data sets that are too large to be

handled by old traditional methods and technologies, while Bello-Orgaz et al. (2016)

defined big data as data sets that are so massive as to extend beyond the ability for

traditional database and software tools to manage and analyze them effectively.

A proper definition for big data must extend beyond the size or attributes of the

data and should also include the concept of the usefulness of the raw data. Bello-Orgaz et

al. (2016) indicated that the definition of big data should include the ability to make sense

of those data and exploit their value. De Mauro et al. (2015) articulated that one of the

fundamental reasons for big data existing is the ability that it provides to make informed

decisions from both structured and unstructured data and failing to include that aspect in

the definition would not provide a full understanding of what big data is intended to be.

De Mauro et al. (2015) proposed the following formal definition of 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 transformational value” (p.

103). Alternatively, Kemp (2014) offered the following definition of big data: “Big data

is shorthand for the aggregation, analysis and increasing value of vast exploitable datasets

of unstructured and structured digital information” (p. 483). Additionally, Wang et al.

(2016) offered the following definition: “Big data start with large-volume, heterogeneous,

autonomous sources with distributed and decentralized control and seek to explore

complex and evolving relationships among data” (p. 750). Since I will be focusing on the

governance aspect of big data, rather than the unique tools required for the analysis of big

data, I will use the Wang definition for this study.

Challenges with big data. While there were varying definitions of precisely what

constitutes big data, the literature covered the challenges around gathering and analyzing

big data in great detail. De Mauro et al. (2015) identified that the combination of

structured, unstructured, and semi-structured data is challenging to handle with traditional

systems. Yang, Huang, Li, Liu, and Hu (2017) pointed out that due to the heterogeneous

and unstructured nature of big data, it is challenging for traditional systems to manage

and analyze those data effectively. Grover, Chiang, Liang, and Zhang (2018) estimated

that unstructured data make up 95% of big data. Siddiqa et al. (2016) clarified that

traditional infrastructures are not designed to handle the distributed computing

functionality required to analyze the large quantity and variety of elements found in big

data. Lee (2017) also pointed out that managing data at velocity strain contemporary data

center architectures which are not sufficient to sustain the volume and velocity of

heterogeneous data elements of both personal and corporate data. Due to the velocity

attribute associated with big data, Lee and Kang (2015) stated that contemporary

analyzing methods such as storing data elements in a data warehouse for later analysis is

no longer sufficient. Lee (2017) also pointed out that big data often must be processed

immediately, or the value will be lost, such as in the case of patient monitoring or

environmental safety-related systems. The volume and variety attributes of big data make

it very difficult to determine the veracity of those data (Matthias, Fouweather, Gregory, &

Vernon, 2017).

Complexity and velocity. Identifying and categorizing big data at both velocity

and volume poses problems for contemporary systems. Kemp (2014) articulated that

identifying the big data engine holistically across the entire enterprise for input,

processing, and output is significantly different than what traditional data systems have

required. Yang et al. (2017) pointed out that metadata can provide significant benefits for

normalizing unstructured data elements for traditional structured data systems, but this

brings about another challenge regarding how to automate the generation of metadata for

such data elements. The high velocity poses an additional problem of lack of storage

capabilities for the transient data as it flows through the analyzing systems. Yang et al.

(2017) further described that maintaining traditional storage systems with existing

redundancy technologies, such as a redundant array of independent disks (RAID), breaks

down at the scale needed for big data.

The complex nature of structured and unstructured heterogeneous data elements is

adding strain to existing analytic systems. Jin et al. (2015) articulated that data scientists

currently lack an understanding of the relationship between data complexity and

computational complexity as relates to big data processing. Yang et al. (2017) further

noted that traditional database management systems lack the scalability for managing and

storing the unstructured elements of big data. Jin et al. (2015) pointed out that among

industry experts, there is not a good understanding of how to deal with the complexity

that comes from complex data structures, complex data types, and complex data patterns

inherent with big data. When considering the velocity associated with big data, Yang et

al. (2017) pointed out that it is clear that high-dimensional data cannot be processed

efficiently within the time constraints required to process big data effectively. Jin et al.

(2015) articulated that data complexity is one of the significant challenges with big data

as compared to traditional data. Yang et al. (2017) argued that traditional algorithms

require structured homogenous data to be effective, and they tend to break down when

applied to the heterogeneous environment of big data.

Data quality is also quickly becoming an issue with analyzing big data. Data

quality is especially problematic because, as Herschel and Miori (2017) pointed out,

organizations tend to have a false sense of confidence in those data simply because of the

hype around big data. Yang et al. (2017) stated that due to the heterogeneity and

complexity of big data, the accuracy, completeness, redundancy, and consistency of those

data become very difficult to manage, thus reducing data quality. Hashem et al. (2015)

added that the multiple sources of big data are less verifiable than was the case with

traditional data, which calls into question the quality of some of the data. Umer, Kashif,

Talib, Sarwar, and Hussain (2017) stated that determining the origin and transformation

of data elements contained in big data is one of the primary challenges of big data

currently, and Cervone (2016) added that external data sources are less likely to have

internal data stewards or data governance processes established to govern its use within

the organization.

Poor data quality can lead to ineffective decisions within an organization. Begg

and Caira (2012) argued that poor data quality is disruptive and dangerous to a business.

As both the variety and volume of data increase with big data from IoT devices, Lee

(2017) pointed out that the quality of those data tends to decrease. High velocity and

realtime analysis can create a timing issue, since not all the needed data may be available

simultaneously for a specific analysis algorithm. Janssen, van der Voort, and Wahyudi

(2017) argued that incomplete data can result in partial analysis, which can paint an

inaccurate overall picture. Yang et al. (2017) pointed out that the veracity attribute of big

data calls for preprocessing to improve the quality of the data to be processed. This need

for real-time processing, in turn, creates an additional challenge regarding the current

processing power of existing systems.

Volume and security. The resource requirements associated with the analysis of

big data are stretching traditional systems to their computing limits. Yang et al. (2017)

articulated that the processing needs for big data currently exceed the capabilities of

traditional systems, and Lee and Kang (2015) concluded that the limitation of processing

power poses a significant challenge for analyzing big data. Intezari and Gressel (2017)

added that to extract information from big data, new techniques, and advanced tools are

required. Jin et al. (2015) identified that this is a problem at the architectural level, as the

system architecture of traditional systems lacks the operational efficiency to analyze big

data effectively. Lee and Kang (2015) added that this problem extends to the software

layer as contemporary event processing engines do not support the parallel computing

needs required for managing complex event processing associated with big data. While

cloud computing may offer some relief, Yang et al. (2017) pointed out that it also adds

another set of challenges, such as networking bandwidth needed for processing such large

amounts of data.

Security and privacy are also challenging for big data analytics. Lee (2017)

pointed out that big data can include many elements of personal information, which

Sivarajah et al. (2017) indicated can raise grave concerns for individuals, companies, and

governments, such as location-based information, which is found within many forms of

big data. Gao et al. (2018) noted that individuals disclose very personal information on

social networks often unconsciously, and Flyverbom, Deibert, and Matten (2019) added

that this information now allows corporations to encroach on the private lives of

individuals at an unprecedented scale. Metcalf and Crawford (2016) added that there are

so many publicly available data sets about individuals that they become highly

identifiable when correlated together.

Utilizing big data is potentially damaging to overall security and privacy. Herschel

and Miori (2017) found that data sources that had not previously had an issue with

privacy may end up threatening privacy once combined with other data sources in the big

data construct. Yang et al. (2017) demonstrated that in the cloud computing environment,

data owners have significantly less control over data confidentiality and integrity due to

the virtualized environment of cloud systems, and Ahmed et al. (2017) added that

security concerns are a significant detractor to faster adoption of IoT devices.

Carbonell (2016) added that there is a big data divide between people and the information

collected about them through IoT, and pointed out that individuals are rarely granted

access to their data.

Big data require changes to traditional security controls. Lee (2017) argued that

weak security controls around big data can lead to financial loss and reputational damage.

Duncan, Whittington, and Chang (2017) identified that because big data originates from

various potentially insecure sources, it is inherently insecure. Yang et al. (2017) stated

that traditional methods of encryption break down in the high-velocity and high-volume

environment of big data analysis, and performance and scalability start to become an

issue when applying contemporary encryption algorithms to big data. Sivarajah et al.

(2017) identified that ubiquitous data sources associated with big data magnify the lack of

advanced security controls within traditional infrastructure systems.

Privacy issues relating to big data are significant because of the predictive

analytical power that is associated with big data. Herschel and Miori (2017) indicated that

combining data mining and analytics with big data creates new opportunities for

organizations to generate new data with significant predictive accuracy relating to

specific consumers. This concept was first publicly documented in a now-famous article

in which Target used a teenage girl’s purchase history to predict that she was pregnant

before she informed anyone (Mai, 2016). Herschel and Miori (2017) pointed out that the

Target incident demonstrates the new reality that organizations may be privy to

information that consumers never intended to share with them. Baruh and Popescu (2017)

argued that the lack of control over information has the potential to create more harm to

consumers than the benefits that the consumer might obtain from predictive analytics,

such as targeted marketing to meet their current needs. With the proliferation of IoT

sensors, social media, and other big data sources, people today are revealing very

personal information daily, often without their knowledge or consent (Mai, 2016), an

activity that Jain, Gyanchandani, and Khare (2016) defined as passive data generation.

Mehmood, Natgunanathan, Xiang, Hua, and Guo (2016) added that passive data may

easily be combined with other publicly available sources to generate new information

about the individual, which that person may not have intended to disclose, nor had

permitted the organization to obtain or retain.

Ethical issues. Passive data generation leads to significant ethical issues for

organizations that analyze big data. Herschel and Miori (2017) posited that big data

changes the nature of ethical debates at a fundamental level by redefining the power of

information and the extent to which free will can guide users' actions. Carbonell (2016)

identified three distinct classes of individuals involved in big data: those who generate the

data, those who can collect it, and those who have the expertise to analyze it. Mai (2016)

added that the primary ethical issue is not whether to collect the information, but how and

when it is ethically responsible for analyzing that information. Carbonell (2016) posited

that companies gathering big data are gaining a privileged position that provides them

with unique insights into individuals activities of which individuals themselves may not

be aware.

The use of big data is also raising new consumer privacy concerns. Herschel and

Miori (2017) identified that predictive analysis of passive data can not only generate new

insights, but it can also violate the privacy of consumers by re-identification of data sets

that had previously been de-identified. Metcalf and Crawford (2016) pointed out that

publicly available data can pose significant and severe risks to individuals and

communities when combined with other data sets. One of the first instances of this

occurred in 2006 as Mehmood et al. (2016) described when AOL release 20 million

search queries that had been de-identified, only to have researcher re-identify those data

within just a few days. Metcalf and Crawford (2016) described a similar case where

public information regarding New York taxicab rides was combined with other sources to

identify private information about specific cab drivers, their religious practices, and even

to track celebrities that used the cabs in question.

Proactive efforts to obfuscate data become less effective because of big data.

Herschel and Miori (2017) pointed out that re-identification of previously obfuscated data

has been documented in such cases as anonymous health information shared between

providers, and anonymous dating profiles were combined with public Facebook profiles

and facial recognition. In both cases, the consumers thought their data were anonymous,

but big data analytics re-identified those data revealing private information that the

organization did not have permission to obtain. Rocher, Hendrickx, and de Montjoye

(2019) found that 99.98% of Americans could be re-identified with any data set with as

few as 15 demographic data elements. Mai (2016) stated that this concept calls into

question the perception that has been applied to consumer data privacy since the 1970’s

that consumers provide informed, rational consent for their data to be used by the

organization that they are providing access to their data. Herschel and Miori (2017)

explained that organizations analyzing big data have an ethical responsibility to

understand how to use the data while protecting the privacy and confidentiality of those

data.

The fundamental concept of individual control upon which privacy is based is

called into question by the existence of big data. Mai (2016) pointed out that when

combining and analyzing individual data points from multiple sources reveals new

information about individuals, and those consumers are neither able to restrict or control

access to their data. Jain et al. (2016) defined privacy as the ability for a person or group

to stop information about themselves from becoming known to individuals or

organizations that they have not provided specific authorization to possess that

information. Baruh and Popescu (2017) referred to this as privacy as an individual good

where users are free to negotiate their acceptable levels of privacy. Herschel and Miori

(2017) stated that when secondary data sources can be combined to reverse-engineer

private information about consumers, as happened with both Target and AOL, the current

concept of privacy control as dictated by consumers is not enough to truly protect

individual privacy.

Organizational challenges. In addition to the technical challenges around big

data, some companies are also experiencing organizational challenges when introducing

big data analytics. Wang and Hajli (2017) explained that to leverage big data initiatives

effectively, organizations must be willing to address the managerial and organizational

changes required. Kemp (2014) explained that there is a gap between the amount of data

that organizations can gather and the ability of that organization to leverage that

information in meaningful ways. Lee (2017) pointed out that many big data projects fail

to achieve their designed outcomes, thus degrading management confidence in big data

and potentially stifling further investment in big data initiatives. Kemp (2014) warned

that demand for the rich information that arises from big data analysis is fueling a

bottom-up approach to data governance, which may lack sufficient legal, compliance, and

regulatory controls. Kemp (2014) added that a top-down approach to data governance of

big data may result in a lack of responsiveness and flexibility.

The demand for big data analytics combined with the new skills needed to analyze

big data properly is creating another challenge around skilled talent within the industry.

Lee (2017) identified that because of the complexity of analyzing the unstructured and

heterogeneous attributes of big data, there is a shortage of qualified data scientists to

accomplish this task. Jin et al. (2015) added that contemporary data analysts lack a solid

understanding of how to deal with the complexity found with complex big data structures.

In a study of 430 firms with advanced analytics capabilities, Lee (2017) identified that

approximately 66% indicated that they are not able to employ enough qualified analysts

to extract the value from the data that they have gathered. Lee (2017) added that the

McKinsey Global Institute estimated that an additional 140,000 – 190,000 big data

analysts will be needed just in the U.S. to meet the demand for big data analytics within

corporations.

The velocity, complexity, and volume of big data, as well as the decay of the

usefulness of those data, pose challenges for disseminating the analysis of big data in

ways that can help meet the business need. Gupta et al. (2019) argued that escalating

information availability and visibility improves decision making, and Mahdi, Nassar, and

Almsafir (2019) asserted that knowledge dissemination increases teamwork. Kemp

(2014) warned that getting that information from the data analysis to the people within

the organization who need it in a timely and efficient manner remains a significant

challenge. Further, Yang et al. (2017) pointed out that due to the heterogeneous attribute

of big data, the visualization of those data in a human-understandable manner poses a

significant challenge.

The value of big data is hampered by the lack of big data expertise and

governance within the organization. Matthias et al. (2017) identified that effectively

utilizing big data to make decisions poses a considerable challenge for companies. Paul,

Aithal, and Bhuimali (2018) stated that big data analytics rely on fuzzy logic and

inductive statistics, and Matthias et al. (2017) explained that improper or incomplete

analysis of big data can lead to misinterpretation of those data, and lead to incorrect

business decisions. Herschel and Miori (2017) added that because big data is by

definition obtained from multiple sources, and those sources are not always known or

validated, those data sets are often incomplete or inaccurate. To be able to access the

benefits from the extreme amounts of data available to them, Ahmed et al. (2017)

articulated that companies will need to develop and implement new platforms for

analyzing the extreme volumes of data generated by connected devices.

Opportunities for decision-making. One of the more important themes in the

literature involved how big data can affect decision making. These decision-making

opportunities include the IoT, the impact of big data on decision making, and big data

analytics. Each of these are described in the following sections.

IoT and big data. One of the primary contributors in the context of big data is the

IoT. Lee (2017) pointed out that in 2016, 5.5 million new devices were connected each

day to the Internet to share, process, and analyze information, and that number was

expected to grow to 20.8 billion by 2020. Ahmed et al. (2017) stated that the number of

human beings globally has now been surpassed by the number of connected devices, and

the overall number of connected devices was expected to reach 50 billion by 2020, and

Herschel and Miori (2017) added that the number of network connections was expected

to reach 18.9 billion, which amounts to approximately 2.5 connections for every person

on earth. The Organisation for Economic Cooperation and Development estimated that by

2020, the average family of four would have 50 internet-connected devices (Government

Accountability Office, 2017). Bilal, Khalid, Erbad, and Khan (2018) warned that IoT data

analysis is driving new challenging requirements for traditional systems due to the high

degree of mobility, real-time data analytics, and extreme latency sensitivity.

One of the results of the proliferation of IoT devices is the phenomenon of

merging the physical and virtual worlds. Because of the amount of data generated by IoT

devices, Paul et al. (2018) explained that some data scientists claimed that big data

provides a direct link between cyber and physical systems. Dourish and Cruz (2018)

utilized the term datafication to describe the transformation of physical, social action into

quantifiable digital data, which enables real-time predictive analytics. Flyverbom et al.

(2019) added that datafication implies that social lives have a digital counterpart in the

form of digital traces left by merely participating in social networking. Galliers, Newell,

Shanks, and Topi (2017) articulated that one of the critical issues with this collision of

physical and virtual worlds is the fact that the minutia of everyday life suddenly becomes

part of big data that is then algorithmically assessed without further human interaction,

often without the individual's knowledge or permission. Herschel and Miori (2017) added

that the analysis of this minutia leads to organizations generating new insights into

individual consumer's behaviors and activities, often without the consumer's consent.

The amount of new information generated daily continues to increase. Lee and

Kang (2015) explained that extreme amounts of data were already being generated even

without considering the IoT, as evidenced by Google which was generating about 25

petabytes of data daily, and Sivarajah et al. (2017) highlighted that estimates are that

globally 2.5 quintillion bytes of data are generated per day. Bughin (2016) pointed out

that various reports predict that data creation would continue to increase by between 40%

and 60% per year. Herschel and Miori (2017) stated that projections indicated that the

total volume of data would increase to 40 zettabytes by 2020.

Analytics of those vast data sets can have significant predictive power. The

Government Accountability Office (2017) identified that new big data analytics

technologies would be capable of extracting hidden patterns and correlations from these

extensive data sets. Ahmed et al. (2017) stated that combining IoT and big data analytics

promises to improve decision making within businesses significantly. Carbonell (2016)

described the power shift that arises as organizations start utilizing big data to predict

future events and even intervene before those events are set in motion. Chauhan and

Sangwan (2017) added that the advantage of big data over traditional data analysis is

better accuracy, which leads to higher confidence in data-driven decisions. This

confidence comes because, as Wang and Hajli (2017) identified, big data analytic systems

provide insights into hidden correlations between data elements that were previously

impossible to determine with traditional data analytics tools. The Government

Accountability Office (2017) predicted that businesses that could analyze data from IoT

devices could build competitive advantages over their competition. In the manufacturing

industry, for example, Popovič, Hackney, Tassabehji, and Castelli (2018) identified

multiple case studies that revealed that big data analytics has already improved

manufacturing operations due to the insights gained from that analysis. Baruh and

Popescu (2017) highlighted one example of this change is the fact that automobile

insurance companies have leveraged big data from GPS sensors to shift from a risk-based

model to a habit-based model using real-time monitoring of individual consumers.

Impact on decision-making. Harnessing predictive analytics with big data

provides significant competitive advantages. Intezari and Gressel (2017) articulated that

the success of an organization is primarily affected by the competitive ability of managers

to make strategic decisions in the face of uncertainty and ambiguity. Braganza, Brooks,

Nepelski, Ali, and Moro (2017) predicted that the economic value of organizations could

be elevated by effectively utilizing the deeper insights available from big data. Cao et al.

(2019) found that by improving information processing capabilities, the decision-making

effectiveness within the firm increased and created competitive advantage, especially

when those capabilities were rare, valuable, and inimitable. Grover et al. (2018) agreed

and clarified that big data analytics were likely inimitable because they are so closely tied

to the decision-making culture and leadership within the firm. The Government

Accountability Office (2017) added that decision making could be enhanced by analyzing

aggregated data from IoT devices. Popovič et al. (2018) added that the distribution of big

data analytics within organizations had a direct effect on improved business decisions

within the organization.

Competitive advantage from predictive analytics derives from the ability of

organizations to quickly adjust to trends uncovered by those analytics. Ahmed et al.

(2017) explained that companies can use big data analytics to drive consumer behavior

by watching for real-time trends, such as an empty parking lot, and reacting in real-time

to promote specials that bring customers into the store. The Government Accountability

Office (2017) suggested that advanced data analytics would be able to draw conclusions

automatically from various discreet data sources what used to require human intervention,

such as suggesting alternative routes in real-time due to traffic patterns, weather, and

other conditions. Lee and Kang (2015) further added that the availability of personal

location information from big data, such as real-time commuter information, was

projected to save $600 billion annually by 2020.

Big data analytics. Big data analytics and effective decision making are closely

correlated. Braganza et al. (2017) articulated that effective analysis of big data enabled

more evidence-based decisions, while Wamba et al. (2017) added that big data enabled

data-driven decision making. Paul et al. (2018) further clarified that analyzing those data

had become a critical need for every business sector. Janssen et al. (2017) stated that

effectively analyzing big data had the potential to change the way that organizations make

decisions fundamentally. Intezari and Gressel (2017) articulated that due to the

uncertainty and ambiguity related to strategic decision-making organizations needed to

reconfigure and reassess their knowledge base continuously. Zakir et al. (2015) observed

that companies that oriented their decision making around fact-based data analytics

outperformed their counterparts in the market. According to Jin et al. (2015), big data has

had a significant impact on almost every business sector and significant industry already,

and to remain competitive, Lee (2017) stated that all companies would need to build big

data expertise. Günther, Rezazade Mehrizi, Huysman, and Feldberg (2017) pointed out

that there is quite a bit of evidence currently that big data can have a significant impact

on an organization's business models. For organizations to extract value from big data,

Günther et al. (2017) stated that it is critical for them to continually realign work

practices, organizational designs, and stakeholder interests.

Organizations are still struggling to understand how to effectively utilize big data

analytics to guide real-time decisions. According to Patel et al. (2017), big data was the

most trending contemporary topic in the computer science field. Wang et al. (2016)

pointed out that between the years 2000 and 2016, there were a total of 7,121 publications

relating to big data, including 2,924 articles published in over 1,000 different journals.

Matthias et al. (2017) discussed multiple case studies of organizations utilizing big data

analytics to provide evidence that significant operational insights are now obtainable

from big data that were previously unavailable from traditional analytics. Janssen et al.

(2017) posited that the real value of big data is found in its potential to improve decision

quality. Gandomi and Haider (2015) identified that one of the most prolific uses of big

data is predictive analytics, where data analysts use a variety of techniques to attempt to

predict future outcomes based on patterns found within big data. Janssen et al. (2017)

warned that care must be given, however, since decision quality may be degraded if the

decision-maker does not understand the relationships between the various data elements

used as part of the analysis.

Organizational and employment impact of big data. It is not yet clear how to

most effectively organize to best utilize big data within the organization. Günther et al.

(2017) discussed the fact that scholars were actively debating what the most appropriate

organizational model would be to most effectively take advantage of big data analytics.

According to Janssen et al. (2017), the collection and processing of big data were quite

often accomplished across departments and even across various companies, requiring

collaboration and partnerships that do not exist with traditional data. Braganza et al.

(2017) added that this lack of role clarity across the organization was inhibiting the ability

to use required big data resources efficiently.

The cross-departmental collaboration required for big data to be successful causes

challenges for contemporary firms. Because of the value and direct relationship between

big data analytics and improved quality of decisions within the organization, Popovič et

al. (2018) identified that power shifts had been documented within organizations that are

effectively utilizing big data analytics to make fact-based and real-time decisions.

Braganza et al. (2017) added that big data processes required new roles and organizations

to interact, which had not previously needed to work together. While there was no

consensus regarding the impact of big data analytics from IoT on employment, the

Government Accountability Office (2017) predicted that it was likely there would be a

combination of both new job creation and job loss.

The need for new organizational behavior also brings the need for new

governance processes. Ahmed et al. (2017) articulated that integrating and analyzing big

data were revolutionizing business processes by turning big data into valuable insights

not available with traditional technology. Janssen et al. (2017) added that the quality of

decisions from big data was not influenced solely by the amount of data, but more

importantly, by how those data were collected and processed. Matthias et al. (2017)

pointed out that without appropriate analysis, it did not matter that large amounts of data

were collected. De Mauro, Greco, Grimaldi, and Ritala (2018) agreed that companies

needed to acquire the right expertise to deal with these technological advances. As stated

by Ahmed et al. (2017), value was extracted from big data when information was

transparent and usable to the organization.

Cloud computing. The adoption of cloud computing capabilities enhances the

adoption of big data within organizations. Hashem et al. (2015) articulated that big data

and cloud computing are conjoined technologies. Because IoT devices are generating

much of the data, Marjan et al. (2017) highlighted that cloud computing is expected to

provide lightning-fast analysis of fast-moving data. Nadjaran Toosi, Sinnott, and Buyya

(2018) added that cloud computing systems are the preferred architecture for hosting

these data-intensive applications. Although big data systems are lagging behind IoT

development, Marjan et al. (2017) clarified that the two should be developed

simultaneously due to their inter-dependence. Ahmed et al. (2017) added that combining

IoT with big data analytics allows for the processing of data closer to the source of that

data.

Cloud computing architecture provides the foundation for big data analytics.

According to Zakir et al. (2015), cloud storage systems offered capabilities that were not

available in their on-premises counterparts, such as the pay-as-you-go payment model,

the elasticity of capacity, and risk mitigation through business continuity and long-term

retention. Varghese and Buyya (2018) pointed out that data centers house most systems

that encompass cloud computing, yet Stergiou, Psannis, Kim and Gupta (2018)

highlighted that the intensive computations required to analyze big data, and the mass

storage needed to support it, were often inefficient even in contemporary cloud systems.

According to Marjan et al. (2017), big data analytics required the same or

fasterprocessing speed of traditional analytic systems at similar or less cost and operating

on much more volatile data sets.

Contemporary cloud systems are still lagging behind the requirements for big data

analytics. Nadjaran Toosi et al. (2018) identified that one of the most challenging scaling

issues relating to contemporary data analysis solutions is the location of the data relative

to the computing resources, with transfer latency of data often exceeding the

computational analysis time. Bilal et al. (2018) added that the IoT is creating new

challenges for cloud architectures when dealing with data systems due to the latency

associated with these large data sets. Stergiou et al. (2018) suggested that significant

changes to the hardware and software architecture will be needed in support of a new

generation of cloud services to support the real-time analytics of data from the IoT.

Varghese and Buyya (2018) suggested that these changes include the decentralization of

data center cloud services by incorporating edge devices in cloud computing architecture.

Most contemporary big data cloud solutions assumed a centralized data center cloud

services model, but Varghese and Buyya (2018) suggested that soon, these architectures

would need to include the use of edge computing such as cloudlets, fog computing, and

mobile cloud computing.

IT governance. Without proper governance practices, technological solutions to

big data problems will not be enough to solve the current problems. Watson and

McGivern (2016) noted that business intelligence initiatives had a low chance of success

without governance practices that provide access to high-quality, secure, and

wellmodeled data. Braganza et al. (2017) added that extracting value from big data

requires not only the appropriate technological solutions but also requires that big data

analysis becomes part of the fabric of an organization. According to Alreemy, Chang,

Walters, and Wills (2016), IT Governance played a critical role in creating this fabric by

establishing a link between the business and IT that has the potential to provide a

competitive advantage by minimizing expenditures and making better use of time.

Repeatable and efficient processes are required to extract value from any data

initiative. Wilkin, Couchman, Sohal, and Zutshi (2016) articulated that IT governance is

the set of systems that direct and control IT operations to meet stakeholder expectations

regarding financial and environmental operations. Alreemy et al. (2016) defined IT

Governance as the process that is established to control investment in IT and to guide

those investments toward achieving business objectives. Wu, Straub, and Liang (2015)

noted that efficient utilization of IT within an organization requires effective IT

governance. Braganza et al. (2017) stated that the repeatability of big data processes

requires reliable and sustainable business processes.

Effective IT governance is critical for IT organizations to effectively meet

stakeholder expectations in all areas of information systems. Wu et al. (2015) stated that

the single most critical predictor of value from an IT organization was an effective IT

governance process. Braganza et al. (2017) suggested that effective business processes

could release value from big data initiatives that have, as of yet, been challenging to

reproduce consistently. Wu et al. (2015) identified that IT Governance was critical due to

the significant investment that is generally associated with IT within an organization.

Mahdi et al. (2019) added that the processes within an organization that create, share, and

use knowledge were highly critical strategic capabilities that create a competitive

advantage to the organization.

IT organizations need to implement formal IT governance practices to align

outcomes with stakeholder expectations. According to Alreemy et al. (2016), the most

important critical success factor for IT governance is the alignment between the business

and IT. Wilkin et al. (2016) noted that alignment between IT and business initiatives was

the most widely realized benefit of IT governance processes. Alreemy et al. (2016)

articulated that appropriately implemented IT governance practices could mitigate

business risks. Wilkin et al. (2016) added that business alignment and risk mitigation

were the primary driving forces for implementing formal IT governance practices.

Smaller organizations are less likely to formalize their governance practices than

their larger competitors. Although both smaller and larger organizations understand the

strategic need for IT governance, Wilkin et al. (2016) noted that larger organizations were

more likely to implement formal IT governance practices than smaller organizations.

Wilkin et al. (2016) added that smaller organizations faced significant challenges in

managing technology assets so that they could sustain a competitive advantage. Begg and

Caira (2012) also added those smaller companies tended to hold the belief that they must

first implement better overall IT governance before they can begin implementing data

governance.

Leadership involvement with IT governance of big data. Contemporary IT

leaders lack strategies for managing big data initiatives. Braganza et al. (2017) noted that

there did not appear to be evidence for the assumption within the literature that processes

for managing big data exist and that resources are well managed. Mikalef et al. (2017)

added that research had focused on big data technologies for data analysis, storage, and

visualization, but there has been a gap of research regarding how organizations would

need to adjust to embrace these technological innovations. Braganza et al. (2017) stated

that senior leadership needed processes that they could implement to extract strategic

value from big data. When addressing the challenges associated with big data analysis,

Begg and Caira (2012) noted that processes need to be very flexible and comprised of

various roles, decision areas, and assignments of specific responsibilities, such as

technical data steward, business data steward, executive sponsor, and other such roles.

Braganza et al. (2017) articulated that the value of big data would continue to be limited

until businesses could deliver repeated benefits from within the same organization over

time and further clarified that Big data processes needed to be flexible enough to change

due to variations in both internal and external forces.

Data governance. In addition to IT governance, organizations also require

specific data governance practices to effectively manage their big data initiatives. To be

able to extract value from raw data, Song and Zhu (2016) noted that organizations require

a well-defined, systematic approach to governing their data. Mikalef et al. (2017)

clarified that Data Governance was the term used to describe these activities, which

included roles, structures, and decision mechanisms around IT resources. However, Begg

and Caira (2012) articulated that there was no evidence to date of a standard

implementation framework for data governance. Hashem et al. (2015) noted that policies,

principles, and frameworks around how big data are used and leveraged within an

organization were posing enormous challenges for those organizations. Posavec and

Krajnovic (2016) recognized that the lack of formal data governance processes was one

of the primary reasons that organizations failed to utilize their data effectively.

The governance of big data is critical to the success of any big data initiative.

Cervone (2016) articulated that while data governance was crucial in most environments,

it was critical for big data, and organizations could not add it as an afterthought. Wang,

Kung, and Byrd (2018) identified that the governance of big data refers to the how-to

aspects of harnessing the data within the organization. Cheng, Li, Gao, and Liu (2017)

added that data governance referred to the series of policies that an organization used to

define cloud data strategy, management, operations, and optimization. Chrisman et al.

(2018) further clarified that governance referred to both formal and informal rules,

practices, and processes that were implemented to direct and control behavior in a

consistent manner throughout the organization. Wang et al. (2016) described information

fusion as the merging of information from various heterogeneous sources toward a

consistently accurate and useful representation. They also clarified that a clear strategy

was required for this process to be successful.

Effective data governance can lead to improved business outcomes. When

organizations implement governance well, Wang and Hajli (2017) identified that big data

initiatives had the potential to provide insights into hidden data correlations for

meaningful decision making. Wang et al. (2018) added that proper data governance

addressed the standardization, accuracy, and availability of those data. When the

organization implemented governance poorly, Wang and Hajli (2017) noted that

organizations experienced substantial financial costs for little value. By effectively

applying governance principles to data governance initiatives, Wang et al. (2018) found

that organizations could better meet organizational goals with lower costs and shorter

timelines. Mahdi et al. (2019) added that to be successful at all, organizations must

methodologically exploit their knowledge assets.

Effective big data governance practices are not easy to implement but can have a

significant impact on the overall business. Hashem et al. (2015) noted that adopting big

data governance practices that provide a balance between value creation and risk

mitigation was imperative to organizations. Mahdi et al. (2019) found that knowledge

management practices within an organization were significantly and positively related to

sustainable competitive advantage. Cheng et al. (2017) added that big data governance

was critical to ensure that data are accurate, shared appropriately, and protected per

company policies. Al-Ruithe, Benkhelifa, and Hameed (2016) further added that data

governance was also vital to demonstrate the ability to manage and monitor data in the

cloud environment where the organization does not control the systems and services. Due

to the high volume and velocity of big data, Siddiqa et al. (2016) stated that key

governance elements such as integrity, auditability, accountability, stewardship, and

change management become extremely challenging.

Effective data governance within the IT organization can help elevate the IT

organization to more strategic alignment with the rest of the business. As more

organizations rely on data as a critical differentiating asset, Watson and McGivern (2016)

identified that governance around the data becomes more critical to the success of the

company, which in turn shifts IT and business intelligence into a more strategic alignment

with business initiatives. Shao (2019) added that alignment between overall information

systems strategy and business strategy is a critical precursor to overall profitability and

competitive advantage. Data governance refers to the people, processes and technologies

that are used to leverage data to create value for the organization by managing, protecting

and using data (Watson & McGivern, 2016), as well as the control and authority over

data-related rules and accountability of individuals and systems accessing and utilizing

those data (Hashem, et al., 2015).

Small companies are slower to recognize the strategic value of data as compared

to their larger competitors. Begg and Caira (2012) found that small companies did not

appear to have the same understanding of the inherent value of data as to their larger

counterparts. Wilkin et al. (2016) added resource constraints in smaller organizations

tended to limit the organization's ability to be competitive in areas such as IT

competencies, knowledge management, and innovation. Begg and Caira (2012) found

that many small companies still viewed data as a means to an end, as opposed to an asset

with inherent value, and they were hesitant to undertake data governance practices out of

concern for inappropriate data retirement or other possibilities for disruption of live data

needed for operations.

Small businesses face larger hurdles to effectively utilizing big data than larger

organizations. Begg and Caira (2012) predicted that data governance would become an

increasingly critical issue for small businesses as vendors and providers forced them into

a cloud business model. When considering the cost of making decisions from bad data,

Watson and McGivern (2016) stated that the cost of data governance was becoming a

foundational cost of doing business in today’s world. However, according to Wilkin et al.

(2016) IT governance competencies were limited in smaller organizations because of

their reduced capacity to identify, assess, and acquire IT talent, knowledge, and skill set.

To remain competitive, Begg and Caira (2012) added that small businesses would soon

be forced to adopt data governance initiatives.

Roles in big data governance. The lack of skilled and knowledgeable personnel

is of even greater concern than the technical challenges surrounding the adoption of big

data within an organization. Technology is actively being developed to fill the gaps

identified in big data analysis. However, Song and Zhu (2016) pointed out that there was

a recognized shortage of people who can think critically about big data and who knew

how to use these new technologies to extract value from those data. Zakir et al. (2015)

added that this required a skill set that was new to most IT organizations. One such new

role in the industry was the Data Scientist, which Song and Zhu (2016) identified as an

emerging role where people were tackling these big data challenges and was considered

one of the top 25 jobs in America in 2016. However, Hu, Luo, Wen, Ong, and Zhang

(2018) identified that there was no agreed-upon definition of data scientist across

different organizations and Gardiner, Aasheim, Rutner, and Williams (2018) warned that

it was challenging to categorize job functions across different companies based only on

job title due to the very loose interpretation of that role at each company. De Mauro et al.

(2018) pointed out that organizations were still defining the job description of Data

Scientist, and this role currently had unrealistically high expectations for managing big

data. De Mauro et al. (2018) added that extracting value from big data required more than

a single role working together across multiple disciplines.

Because big data initiatives intersect with so many disciplines it is a challenge to

find all of the needed skills in a single person or role. Gardiner et al. (2018) found that

data scientists were required to possess a wider variety of skills than traditional

technologists due to the multi-faceted nature of this new role. De Mauro et al. (2018)

stipulated that the main focus of the role was around the data itself, and the associated

analytical methods for transforming those data into usable insights. However, in their

analysis of contemporary, big data job descriptions, Börner et al. (2018) indicated that

soft skills such as collaboration, communication, and writing and problem-solving were

all required for the role of a data scientist. Chen and Zhang (2017) also found skills

relating to social sciences among the most common qualifications listed for big data roles.

Gardiner et al. (2018) identified that contemporary data scientists required both analytical

and soft skills.

Organizations are currently scrambling to identify individuals with this unique

confluence of skills. Börner et al. (2018) identified 69,405 jobs posted for data scientists

or data engineers within the six years between 2010 and 2016. Gardiner et al. (2018)

stated that McKinsey previously projected that by 2018, there would be a 50% - 60% gap

between the supply and demand for big data analytics talent. As De Mauro et al. (2018)

pointed out, the deep analytical expertise required of most data scientists was not

requisite for maintaining the real competitive advantage expected from big data. Gardiner

et al. (2018) added that the skills required to do so include such diversity as mathematics,

probability, statistics, programming, data management, data visualization, data mapping,

to name just a few.

Relationship of the Study with Previous Research

Existing literature on the topic of big data primarily focuses on two overarching

topics: the analytical value of big data and the challenges associated with utilizing big

data as a source of information. These two high-level themes categorize the various

studies and other literature described in the previous sections. The literature regarding

data governance does not apply solely to big data, but rather the available studies focus

on the benefits of data governance in general terms and on the benefits of data

governance in general. This study adds to the body of research by focusing specifically

on the implementation of traditional data governance principles when utilizing big data as

the data source for analytics for decision making within an organization.

Transition and Summary

In this section, I defined my research by introducing both the problem and purpose

statements and introduced my research question and the supporting interview questions

that I used in the interview process. I defined different terms that will be referenced in

this study and explained the limitations, delimitations, and assumptions that apply to this

study. I further demonstrated that big data analytics brings with it both high expectations

for a positive impact on business decisions as well as significant challenges in the overall

governance of those data. In the review of the professional and academic literature, I

described how Galbraith’s work on the OIPT set a solid foundation for the study of big

data in general. I introduced the OIPT and explained how this theory was utilized to guide

this study. The literature review provided an overview of big data in general, as well as

some of the critical challenges that arise from big data analysis. I discussed the topics of

security, the IoT, cloud computing, organizational impacts of big data analysis, and IT

governance. The literature review culminated with a discussion of data governance and

its implementation within mid-size organizations. The following section identifies the

method for performing the study, including the role of the researcher, ethical

considerations, and other vital elements and details around the application of the study

itself.

Section 2: The Project

This section of the paper contains information about the participants, sample,

population, and research methodology. I provide justifications for my decisions regarding

the design of the study and address ethical considerations. I also describe my approach

for data collection and analysis.

Purpose Statement

The purpose of this qualitative multiple case study was to explore big data

governance strategies employed by data scientists to provide a holistic perspective of

those data for making decisions. The population for this study was data scientists

employed in three mid-market companies in the greater Salt Lake City, Utah area who

have strategies to govern big data. The findings from this study could potentially affect

social change by creating additional jobs in mid-market companies and enabling small

and medium-size companies to better cater to the needs of their specific customers. Job

creation would be accomplished through increased capability of these companies to have

a better understanding of data which they gather and use to make strategic and

operational business decisions daily. This increased and more targeted information could

lead to increased profitability for these mid-market companies, which in turn could lead

to increased jobs. Customers would benefit because local small and medium-sized

businesses would now be able to customize their offerings to specifically meet the needs

of their customers in a similar way that large companies with much more extensive

research & development and marketing budgets do.

Role of the Researcher

Qualitative research dictates that the researcher is the principal data collection

instrument, and as such, the researcher maintains as essential a role in the research as the

participants themselves (Draper, 2015). As the primary instrument for data collection, the

researcher is not only an assessor or observer but also the primary analyst for

interpretation and synthesis of data (Cook et al., 2016). My role as the sole researcher in

this study was to perform interviews, gather data, and analyze and present those data in

an unbiased manner. Kelly (2017) explained that in this role, it is crucial to identify

experiences, history, and knowledge to the reader of the study to uncover and disclose

any biases. This is essential so the reader can better comprehend my biases, as well as

where I can apply my expertise within the context of the study. Since 1988, I have

worked in various roles within IT, with much of my early career as a developer and

enterprise architect. In the latter part of my career, I have worked as Chief Information

Officer in various regulated industries, and have focused on process improvement and

interpersonal management of technology employees in various roles within IT. My

experience has not been with big data or data analytics specifically, although I have

managed teams responsible for data analysis. I consider myself unbiased concerning the

governance of big data. I have spent my entire professional career in the same general

geographical area where I conducted this study. As data from multiple sources were

analyzed, it was my responsibility to search for similarities and differences among those

data while making sure to avoid any bias.

To mitigate bias, the researcher must strive to be reflexive throughout the study,

continuously asking if the results are consistent with the research question and whether

the researcher has conducted the analysis fairly (Kelly, 2017). The researcher must keep

records and transcriptions of interviews, as well as documentation of their feelings and

impressions from interview sessions (Kelly, 2017). While performing interviews, I kept

this in mind and continually sought fairness as I conducted the interviews and listened to

responses. I thoroughly documented my feelings and impressions as well as respondents’

answers. I also employed member checking to ensure that I accurately interpreted

participants’ responses. I watched the body language of participants to identify any

questions that may have made them uncomfortable and to attempt to identify any areas of

potential participant bias.

Studies involving human participants require ethical approval (Kelly, 2017). Ethical

considerations require that no harm will come to participants as a result of their

participation in this research (Kelly, 2017). The Belmont Report outlines ethical

principles and guidelines for protecting human subjects while conducting research,

explicitly identifying beneficence, justice, and respect for all persons in the study (U.S.

Department of Health & Human Services, 1979). I achieved beneficence, justice, and

respect for all participants by treating all participants fairly and with respect and both

clarifying and protecting information they provide to me as part of this study. I adhered to

the process outlined in the ethical research section of this study.

Mitigating bias during data collection includes carefully planning interactions with

participants. An interview protocol can assist in the development of an inquiry-based

conversation (Castillo-Montoya, 2016). The researcher should make the process of data

collection as explicit as possible by establishing a formal interview setting (Ponelis,

2015). Semi-structured interviews provide a means for the researcher to ask specific

questions to elicit open-ended responses that allow participants to guide the dialogue in

ways that the researcher could not have anticipated a priori (Brown & Danaher, 2019).

Interview protocol should be structured to ask multiple participants the same starting

questions as a means to maintain a steady target for data saturation and mitigation of bias

(Fusch & Ness, 2015). Because interactions between the researcher and participants add

value to the research, the researcher can strategically arrange the interview setting so

participants feel invited or authorized to share interpretive knowledge that would not

result otherwise (Malli & Sackl-Sharif, 2015). I worked to ensure that any potential

biases were mitigated by following a planned interview protocol (see Appendix B) and

watching for participant reactions throughout the semi-structured interview process.

Semi-structured interviews provide a setting for thoughtful reflexivity regarding the

overall research question (Brown & Danaher, 2019). The research design must link the

research question and overall purpose to processes for data collection and analysis to

draw conclusions from those data (Ponelis, 2015). The data collection protocols include

the types of people who will be interviewed, documents that must be analyzed, and

observations that the researcher must make (Yin, 2017). While the researcher should

make the data collection process as explicit as possible, a semi-structured approach is

extremely valuable to allow for the exploration of new information that will arise during

the interview process (Ponelis, 2015).The interviewer needs to ask well-structured

questions while listening to and interpreting answers to find a balance between passively

listening and overdirecting the interview (Ponelis, 2015). I followed a semi-structured

interview approach by asking questions as outlined in Appendix B and interacted with

participants by asking for more elaboration when responses involved new data related to

the research question (see Appendix B). I maintained field notes, including observations

during the interview process.

Participants

Participants of qualitative studies are typically selected based on their involvement

with the topic being studied based on the research question (Fink, 2000) and their ability

to provide detailed descriptions regarding the study topic (Draper, 2015). Well-informed

interviewees can provide insights into history and decision-making that influence

organizations to end up with models that are currently in use within their organizations

(Yin, 2017). Participants were individuals who were employees or contractors working

within the partner organizations and had intimate knowledge of data management and

analytical strategies of the organizations, had been in a data scientist type of role for a

minimum of 5 years, were data analytic solutions architects at these companies, and had

no recurring working relationship with me.

A gatekeeper can help the researcher by providing access to key individuals

within an organization (Fusch & Ness, 2015). I used a gatekeeper within each

organization to assist in identifying participants and selecting appropriately private

locations for interviews. I met with these individuals to discuss the research question and

overall strategy for completing this study, and we reviewed criteria for participation

together. I used these interactions to learn about the organization as well as help clarify

participant criteria, and I selected participants that met the criteria for this study based on

this information. I then directly contacted participants to obtain their informed consent

and schedule interviews with them.

It is critical for the researcher to both build trust and to establish a rapport with

participants when using a semi-structured interview process in a qualitative method

(Brown & Danaher, 2019). Rapport comes as the researcher and participant move in and

out of empathetic moments during the interview process (Prior, 2017). Response bias can

arise from participants when rapport leads to behaviors that support the researcher’s

perceived beliefs (Latkin et al., 2016). I did not select participants with whom I had a

prior relationship to mitigate response bias, but rather, I established rapport with

participants by learning about the culture of their organization before meeting with them

and adapting to their dress and behavioral standards. I used video conferencing

technology to allow the participants to choose a location for the interviews that

simultaneously allowed the participant to be in a comfortable location while still

maintaining their privacy. After obtaining their agreement with the informed consent

form, I engaged with the participants and briefly explained to them my background,

history, and experience with data governance and big data. I did not go into too much

detail to mitigate any potential bias. Finally, I ensured that all participants understood that

my role was to gather information and not to judge the validity of their architectural

decisions.

Research Method and Design

Research Method

A qualitative method was most appropriate for this study. The central defining

capability of the qualitative approach is to answer how and why questions (Ponelis,

2015). These studies are particularly applicable to applied disciplines since the findings

are more easily applied to improve specific processes, programs, or problems in practice

(Ponelis, 2015). Qualitative studies require the researcher to possess both emotional

maturity and solid interpersonal skills to recount the experiences of the participants

accurately in their own words (Collins & Cooper, 2014). Qualitative studies are most

useful when investigating how well people perform something (Kelly, 2017). Because

this study required an in-depth exploration of the strategies and methods employed by

data scientists, a qualitative approach was most appropriate.

Quantitative studies are focused more on correlations and trends than on providing

explanations or reasons (Kelly, 2017). Quantitative studies are used when the study

intends to predict outcomes or explain theories which are based on cause-and-effect

relationships and to determine the correlation between variables (Gehman et al., 2018).

Quantitative studies are not as capable of providing explanations for processes (Kelly,

2017). Since there was no hypothesis associated with this study, a quantitative approach

was not the most appropriate methodology. If this study had been investigating

correlations between big data analysis and specific outcomes, then a quantitative

approach may have been more appropriate. I considered using a quantitative approach but

deemed a qualitative approach to be more appropriate due to the research question for this

study.

Mixed method studies are completed by using both qualitative and quantitative

data in the same study (Halcomb & Hickman, 2015). A researcher could employ a

mixedmethod design by performing the qualitative and quantitative methods of research

in either a sequential or concurrent process (Imran & Yusoff, 2015). A researcher should

ask themselves whether the mixed-method approach will add more value than using a

single research method (McKim, 2017). Adding a quantitative design element to my

study would not have yielded any additional insights into the research question since

there are no statistical or quantitative elements to be measured and no hypothesis to be

tested. Since a quantitative approach alone was not appropriate for this topic then adding

a quantitative element by employing a mixed-method approach was also not

inappropriate for this study.

Research Design

I used a multiple case study methodology for this study. A case study

methodology allows for a deep examination of an issue or phenomenon in its natural

setting (Arseven, 2018). Because this study involved an exploration of strategies used

within organizations that are using big data for decision making, this design approach was

warranted. This design methodology is most useful when situations or phenomena are

investigated for exploration to gain specific insights or knowledge (Elman, Gerring, &

Mahoney, 2016). The case study methodology also allows the researcher to observe in

person the operation of the governing body to witness the various attributes of the

successful data governance program in the mid-market organization. This opportunity for

personal observation was critical to this study because considerable research has been

conducted regarding information technology governance programs but there is no

agreement on what specific attributes constitute an effective information technology

governance program (Alreemy et al., 2016). It was important to use the case study design

to personally observe and document the strategies used by data scientists using big data.

A case study design is appropriate when the study is intended to be inductive and

attempts to make broad generalizations (Kelly, 2017). A multiple case study tends to be

more generalizable than a single case study would be (Gehman et al., 2018). Because this

study was investigating different strategies used by data scientists a multiple case study

was more appropriate for uncovering strategies that could be generalized to other

organizations.

An ethnographic design would require embedding the researcher in a cultural

context to observe specific behaviors of a group of people in a specific cultural context

(Mannay & Morgan, 2015; Watson, 2018). The researcher becomes part of the participant

group in ethnographic design (Kassan et al., 2018). An ethnographic approach would

have been appropriate if this study intended to identify the specific culture, norms, or

customs of data scientists. Since I studied the strategies used by data scientists and not

their norms or customs an ethnographic design was not appropriate.

The phenomenological design focuses on how the participants live through,

interpret, or experience the topic under research (Adams & van Manen, 2017; Sousa,

2014). A phenomenological design is intended to explore how the participants make

sense of a phenomenon (Adams & van Manen, 2017). Since this study was not

investigating the perceptions of participants, the phenomenological approach was not

appropriate. In this study, I investigated the strategies employed by data scientists and I

was not looking specifically at their reaction to a specific phenomenon. The

phenomenological design was not appropriate for this study.

Data saturation is achieved when no new information is being uncovered, when

there is enough information to replicate the study, and when further coding is not feasible

(Fusch & Ness, 2015; Lowe, Norris, Farris, & Babbage, 2018). Data saturation means

justifying that the amount of data obtained is not more or less than the data needed for

research (Saunders et al., 2018). To achieve data saturation, I needed to interview

participants until no new information was being revealed. Data saturation was determined

by assessing the responses as I received them and adjusting my questions in real-time to

identify any possible new information. Data saturation was achieved by interviewing all

willing participants within the multiple organizations that were involved in the process of

data analysis concerning big data.

Population and Sampling

This qualitative case study’s population consisted of data scientists that utilize big

data analytics to analyze data for decision making within their organization. I selected

organizations that are either based in the greater Salt Lake City, Utah area, or that have a

technical presence in that geographical area. It can be challenging to gain access to

SME’s unless the researcher has a network to leverage (Ponelis, 2015). To mitigate this

challenge, the population for the study was drawn from companies that are known to the

researcher, either by the association in the Utah Chapter of the Association for

Information Management or through membership in the Big Data Meetup group.

Participant case study organizations were those that were currently implementing or using

big data to make business decisions. The population for this study had the knowledge and

expertise in the area of big data analytics and the associated governance around those

data. The population for the in-person interviews was made up of data scientists within

the organization that were directly involved in the gathering, analysis, and modeling of

big data elements.

Fink (2000) indicated that the number of participants in a qualitative study should

be as many as needed to uncover the information needed to answer the research question.

Fusch and Ness (2015) added that when performing a case study, the researcher cannot

identify a specific number of interviews required to reach saturation; however, the

researcher will take what they can get from the case study organization. For this study, I

used purposeful sampling to select the participants from the population. The sample

should be determined by identifying individuals within each organization that have a high

level of interaction with that population (Whittingham, Barnes, & Dawson, 2016). For

this study, I determined the sample by working with the gatekeeper to identify individuals

within the organizations who had a high level of interaction with big data. I planned to

interview between five and 10 participants from two organizations. The final number of

participants was dependent on achieving data saturation. I completed the study with 10

participants from three different organizations. While different studies use distinct

inclusion criteria, it is acceptable to select participants based on their presence in a

specific context (Fink, 2000). Once the population was defined by the selection of the

case study organizations, the number of participants was be determined relative to the

number of data scientists involved in the data modeling and analysis process within those

organizations and when data saturation is achieved.

Semi-structured interviews can take a variety of forms that can include face-

toface, telephone, webchat, or email (Brown & Danaher, 2019). Typically case study

interviews are conducted in an interactive setting where the participants can feel at ease

responding to the questions (Malli & Sackl-Sharif, 2015) and where the interviewer can

establish rapport with the participants (Brown & Danaher, 2019). Interview settings

should be carefully considered to ensure the comfort of the participant while

simultaneously reducing distractions and background noise (Dikko, 2016). I conducted all

of the interviews using video conferencing, thus allowing the participants to interview in

a location that is familiar to them. The interviews typically lasted for less than 60minutes.

Ethical Research

Ethical research practice includes acquiring informed consent and voluntary

participation of participants (Eisenhauer, Tait, Rieh, & Arslanian-Engoren, 2019). The

researcher must protect the privacy and confidentiality of participants (Yip, Han, & Sng,

2016). All participants should enter research voluntarily and with enough information for

them to decide whether to participate (U.S. Department of Health & Human Services,

1979). I explained the interview process to each participant before providing them with

the informed consent form for their signature. In the email that I sent them with the

informed consent document, I outlined the purpose of the study along with the benefits,

risks, and nature of the study (see appendix C). I also informed them of my plans to

record the interview, and I also asked for their consent for this recording at the start of the

interview. Before acceptance into the study, all participants were required to respond to

the email with their consent. This study was conducted ethically, adhering to all ethical

guidelines as outlined by the university as well as best practices within academic

research.

The Institutional Review Board is responsible for overseeing human studies

research, ensuring that the researcher conducts the study according to generally accepted

ethical standards and that it adheres to federal and state regulations as well as with

university policies and procedures (Caldamone & Cooper, 2017; Cseko & Tremaine,

2013; Qiao, 2018). I obtained approval from the Walden Institutional Review Board

(IRB) before contacting any participants or collecting any data. By the informed consent

process, all participants signed a form that informed them of their rights as a study

participant, including the method for withdrawing from the study. Participants have

anonymity in this process, which was enhanced by the private setting in which the

interviews took place. Participants were also asked to maintain the confidentiality of the

study by refraining from discussing the names of those involved in the individual

interviews or discussing the content of the interview itself with anyone else while the

study was underway. All efforts were made to protect all participants from any harm as a

result of their participation in this study.

Names of the organizations, as well as the names of the individual participants,

are masked in the study documentation. I maintained a spreadsheet with the names of the

organizations and participants in which I assigned generic names to each one such as

organization I, participant I, participant II, for as many organizations and participants I

have included in the study. This spreadsheet is the only place where the participant's

names are associated with the masked names. This spreadsheet is stored with the other

study documentation in an encrypted OneDrive folder, and I did not include them in the

final study documentation. This process provides an audit trail and a simplified method of

referring to each participant in an unidentifiable manner throughout the study

documentation. Participants were allowed to decline to participate in the process or

withdraw at any time. The process for withdrawal from the study was to inform the

researcher that they are not willing to participate in the interview process. Upon receiving

their notice to withdraw, I removed them from the interview process. There were no

repercussions to the participant for their request to withdraw.

Following the Walden University process, data from this study will be maintained

for five years after the study, and I will keep them in an encrypted OneDrive folder

during that period. After five years, I will destroy all the data from this study. Participants

will also have no monetary or non-monetary incentives offered for their participation in

this study.

I interviewed participants in a one-on-one setting, with member checking utilized

to ensure that I had documented their responses in alignment with their intent. I did not

use the participant's names in the study and only referred to them by a number. I asked

that each participant maintain confidentiality and that not discuss our interactions with

any other employees at their company.

Data Collection

Data Collection Instruments

In qualitative research, the researcher is the primary data collection instrument

(Clark & Vealé, 2018). Researchers should present their findings in a way that readers

can understand both their approach and their findings (Sutton & Austin, 2015).

Qualitative studies commonly use these four methods of data collection: interviews, focus

groups, document reviews, and observation (Gill, Stewart, Treasure, & Chadwick, 2008).

As the researcher of this qualitative study, I was the primary data collection instrument,

and I used interviews and operational document analysis as the other data collection

mechanisms. I followed the defined interview protocol (see Appendix B) when

conducting the interviews. Methodological triangulation was used to correlate these

multiple sources of information, and from which I generated the study findings. Fusch

and Ness (2015) stated that triangulation goes a long way to achieving data saturation.

Process documentation and transcripts from the interviews were coded and correlated to

uncover trends and standard practices.

As the primary data collection mechanism, the influence of the research cannot be

removed entirely from the investigative equation in a qualitative study (Draper, 2015).

Data collection in qualitative case studies may take several forms, including interviews,

documentation review, on-site observations, and illustrative materials (Yin, 2017). Case

study data collection should follow a formal protocol, but the specific information that

may become relevant is not always known upfront (Yin, 2017). Establishing a rapport

with the participants can set them at ease, which will also increase the effectiveness of the

semi-structured interview process and generate mutually beneficial outcomes (Brown &

Danaher, 2019). Interviews in a qualitative study can provide rich and informative data

that expand the topic (Sutton & Austin, 2015). During the interview process, it was also

essential to ask multiple participants the same questions to achieve data saturation and to

avoid trying to hit a moving target (Fusch & Ness, 2015). Semi-structured interviews help

guide the conversation by allowing the interviewer to ask additional questions based on

responses from the participant that enable reciprocation between the interviewer and the

participant (McIntosh & Morse, 2015). Semi-structured interviews provide a mechanism

for participants to be open to revealing additional relevant information that may not

otherwise be identified (McIntosh & Morse, 2015). I used a semi-structured approach to

the interview process. I did not interrupt responses and allowed the participants to share

as much information as they felt was needed. I asked clarifying questions as needed

throughout the interview.

Document analysis can be used by qualitative researchers to gather pertinent facts

about the organization (Dunne, Pettigrew, & Robinson, 2016). Document analysis

provides a mechanism for reviewing and evaluating documents from the organization

being studied (O’Connell, Mc Carthy, & Savage, 2018). Sometimes the document

analysis may reveal relevant information about the topic that the researcher would not

otherwise discover (Dunne et al., 2016). Analysis of documentation can provide the

researcher with insights into how the organization presents itself as a business (Lawson,

2018). Document analysis can also be used to corroborate information obtained from

interviews to further validate the study (Dunne et al., 2016). The use of triangulation

during data analysis significantly increases the reliability of the results by using multiple

data sources to reach conclusions about the data (Fusch & Ness, 2015). Methodological

triangulation is the specific type of triangulation that is utilized to correlate information

gathered from multiple data collection sources (Fusch & Ness, 2015). Triangulation not

only enhances reliability, but also helps to ensure data saturation, and is the method in

which a researcher explores different levels and perspectives of the same phenomenon

(Fusch & Ness, 2015). I used operational document analysis in addition to interviews in

this study. By working with the gatekeeper, I gained access to the process and standards

documentation that is used by the organization to govern their big data procedures. I

evaluated the documentation for applicability to this study. I used methodological

triangulation and member checking as validation mechanisms.

Data Collection Techniques

Data collection in qualitative research occurs primarily through unstructured

methods of verbal interaction and observations (Morse, 2015). Thorough qualitative

analysis requires that data be gathered from multiple sources to enable triangulation

(Twining, Heller, Nussbaum, & Tsai, 2017). Using multiple data sources will allow for

the methodological triangulation of those data points to enhance the validity of the study

(Morse, 2015). For this study, I gathered data from two distinct sources, including

semistructured, face-to-face participant interviews, and a review of procedural

documentation. I conducted in-depth interviews of individual participants using the

questions in the Interview Questions section and recorded the session using GoTo

Meeting software on a laptop computer. I also took notes where I recorded my

impressions of the responses as well as observations about the participant's body language

and other environmental details. The GoTo Meeting recordings of the interviews were

transcribed into Word documents using a transcription service. I had a confidentiality

agreement signed with the transcription service before the use of that service.

I gained access to participants and documentation through a gatekeeper within the

organization. I ensured that each participant understood and agreed to the informed

consent process, and I conducted the interviews in a private and comfortable setting for

the participants. I acquired IRB approval before initiating contact with any participants.

Once I had their consent to participating in the study, I worked with the

participants to arrange a time for the video conference interview, which I enabled a

remote face-to-face setting within the greater Salt Lake City, Utah area. I worked with the

participants to identify a date and time that fit within their schedule. I ensured that the

interview took place in a location that is free from distractions and with which the

participants are comfortable and familiar, and which protects their privacy. These details

are essential because the location of the interview can influence the content and data

collection process itself, and distractions should be minimized as much as possible

(Dikko, 2016).

I informed the participant that the interview was being recorded for accuracy and

later transcription. I also informed them that they can request that I turn off the recording

at any time. I also informed the participant that I will be taking notes during our interview

and that I would be recording both their responses and my impressions of the interview. I

ensured that they were comfortable with the process before proceeding.

One of the dangers of qualitative research is when data collection is superficial,

such as taking participant's comments at face value (Twining et al., 2017). Participant or

member checking is used in mitigating this danger by allowing the participant to clarify

intent and comment on my interpretations of their input (Twining et al., 2017). A form of

member checking can also be used during the process of the interview by asking for

clarification and more depth of specific responses by the participants during the interview

process (Morse, 2015). Qualitative researchers need to be both gathering and analyzing

data concurrently in an iterative process during the interview (Twining et al., 2017).

During the data analysis phase, my summary of the interview was provided to the

participant for their review, and edits were made accordingly. This process continued

until the participant was satisfied that the transcript reflected their responses accurately.

No online or web-based software was utilized in this process. All digitized data relating

to this study is stored on my encrypted OneDrive and will be maintained there for the

required five-year period. The digital data will then be deleted.

Data Organization

The protection of the anonymity and confidentiality of participants is a critical

aspect of ethical research (Castillo-Montoya, 2016). It is also critical to quality research

to ensure that the data are transparent (Bengtsson, 2016). The use of a spreadsheet for this

purpose makes the data readily accessible (Shapiro & Oystrick, 2018). I used a

spreadsheet to categorize and organize the data and findings from this study. This

spreadsheet is stored on an encrypted OneDrive for both security and recovery purposes. I

used participant numbers to identify the participants both within the spreadsheet as well

as in the physical folders to ensure that the participant names are not associated with their

data in any way. Before being imported into the computer-assisted qualitative data

analysis software (CAQDAS), the documentation for each participant was stored in the

folder on the OneDrive that is associated with their organization. I used the folder to store

interview notes, emails, consent forms and other documentation from the participant or

that was provided to me by that participant.

Due to the remote nature of the interview process, there was no physical

paperwork collected during this study. All electronic data, including a data backup of the

CAQDAS application, are stored in the same OneDrive folder location. This process

ensures accessibility to the data while protecting both the security of the data and the

ability to restore in the case of accidental deletion. Participants were not identified in any

way other than by their assigned numbers to protect anonymity. After five years, the

information on the OneDrive will be deleted.

Data Analysis Techniques

After collecting the data, it must be analyzed to find meaning within the data that

is relevant to the research questions and the overall purpose of the study (Bengtsson,

2016). Data analysis is the process by which the researcher reviews the collected data

repeatedly to uncover either the correlation or contrast that provides meaning for the

study (Fusch & Ness, 2015). Grouping data into common topics and themes allows for

better data analysis, as it focuses the large data sets into smaller and more isolated

elements (Maguire & Delahunt, 2017). My review focused on the strategies, processes,

and procedures identified within the organization for overseeing the governance of big

data. I reviewed the documentation, interview notes, transcripts, and other data sources

multiple times, searching for common themes and descriptions with each review. I

reviewed and analyzed the operational documents as a mechanism for validating the

descriptions from the in-depth interviews. I identified major themes from the various

sources and used color coding of the data to identify and correlate significant themes.

I reviewed the data gathered repeatedly to uncover meaningful information to

inform the research. I specifically searched for data that answered my central research

question of what big data governance strategies do data scientists employ to provide a

holistic perspective of those data for making decisions? Identified themes were

highlighted in various colors on the digital documentation in the CAQDAS.

Software packages can be used by qualitative researchers to assist in the analysis

of qualitative data to look for dimensions, categories, and subcategories of analysis

(Freitas, Ribeiro, Brandão, de Almeida, & de Souza, 2019). Software packages such as

this enable the researcher to relinquish the merely technical tasks relating to the

organization of the findings that do not require intellectual effort (Freitas et al., 2019). A

Computer-Assisted Qualitative Data Analysis Software (CAQDAS) tool is a

computerbased software tool that can automate the qualitative data analysis process

(Moylan, Derr, & Lindhorst, 2015). Some researchers argue that using a CAQDAS tool

for qualitative data analysis provides better results than manual analysis would (Moylan

et al., 2015).

For this study, I utilized a CAQDAS tool called QDA Miner. This software tool

was used for the text-based analysis of interviews and open-ended questions. QDA Miner

was used for coding of the various data sources from my data collection. I entered the

information from the various data sources into QDA Miner and used the coding feature of

the software to help me to analyze the information and categorize the data into themes,

groups, and categories. This software helped me to discover relationships among the data

elements. I was looking for reoccurring themes and strategies from the various sources of

data gathered during my data collection phase.

Reliability and Validity

Validity in a qualitative study is the extent to which the instrument of

measurement has measured the concept that it was intended to measure (Dikko, 2016).

Reliability is defined as the ability to consistently and without bias, achieve similar

results (Dikko, 2016). Trustworthiness in a qualitative study is established through four

criteria known as credibility, dependability, confirmability, and transferability (Forero et

al., 2018). Credibility, dependability, and confirmability in qualitative studies, replace

internal validity, reliability, and objectivity (respectively) in quantitative studies (Morse,

2015). Credibility and trustworthiness in qualitative studies are enhanced using analytic

processes around these four criteria (Twining et al., 2017).

Taken together, these criteria attempt to speak to the overall rigor of the

qualitative study, but these four criteria have never been tested or questioned since their

introduction in 1981 (Morse, 2015). While qualitative rigor is the overall goal, validity

may be supported inherently by reliability, so researchers should be most concerned with

overall trustworthiness, and lean more on traditional reliability and validity concepts as

defined in general research terms (Morse, 2015). This rigor can be accomplished by

implementing qualitative comparative analysis, which consists of clarifying the question

of external validity, ensuring internal validity, and explicitly adopting a specific mode of

reasoning (Thomann & Maggetti, 2017).

Credibility

Credibility establishes confidence that the results are accurate, credible, and

believable from the perspective of the participants (Forero et al., 2018). Korstjens and

Moser (2018) explain that credibility demonstrates that the viewpoints expressed by the

participants and aligns with the researcher’s representation of those viewpoints. Ensuring

that the viewpoint of participants is accurately reflected in the study findings aids in

producing research that is both ethical and reliable (Cypress, 2017). The researcher

achieved credibility by prolonged and varied engagements with each participant, and by

documenting the interview process and protocol (Forero et al., 2018).

To achieve credibility, I ensured that I had enough time with each participant to

gather the needed information, and I thoroughly documented the interview process and

protocol. Korstjens and Moser (2018) identify triangulation as one of the critical methods

of achieving credibility. Qualitative researchers often use triangulation to increase the

probability that their study findings will be found credible (Nowell, Norris, White, &

Moules, 2017). I utilized both methodological and data triangulation in this study to

enhance credibility. I also employed member checking to enhance the credibility of the

study. Korstjens and Moser (2018) explain that the use of member checking to gather

feedback regarding the interpretations, conclusions, and analytical categories will

enhance credibility. This process builds confidence in the accuracy of the responses.

Ensuring the accuracy of responses helps to ensure the validity of the overall study and

will help determine the reliability of the content.

Dependability

Dependability ensures the findings are repeatable if the study were repeated with

the same set of participants, coders, and context (Forero et al., 2018). Korstjens and

Moser (2018) define dependability as the stability of the findings over time. It is achieved

when the researcher explains the logical and traceable research process and when the

process is documented (Nowell et al., 2017). Dependability is also achieved by providing

a detailed description of study methods and ensuring that a stepwise replication of the

data is documented (Forero et al., 2018).

To achieve dependability, I provide a detailed description of the study methods

used. I clearly described the process that I followed to both collect and analyze data, and I

ensured that I documented a stepwise replication of the data. Providing an audit trail of

the research process enhances dependability (Korstjens & Moser, 2018; Nowell et al.,

2017) by providing steps for other researchers to follow if they attempt to replicate the

study. I provide an audit trail within my findings that will help the reader to understand

not only how the data were collected but also to clearly understand my thought process

during analysis, my interpretations and impressions during the interview process, and the

logic that I employed to reach conclusions from the data.

Confirmability

Confirmability provides confidence that the results would be confirmed or

corroborated by other researchers (Forero et al., 2018). This confidence is primarily

achieved by data analytics processes such as methodological triangulation (Abdalla,

Oliveira, Azevedo, & Gonzalez, 2018; Forero et al., 2018). Nowell et al. (2017) state that

confirmability is achieved when credibility, transferability, and dependability are all

reached (Nowell et al., 2017).

To achieve confirmability, I ensured that credibility, transferability, and

dependability were all achieved. I utilized both data and methodological triangulation to

gather and analyze the data for this study. I also clearly described my thought process and

the logic that I used to make decisions regarding this study. I documented the interview

process, including thoughts, impressions, choices, and decisions made during the data

gathering and analysis phase of the study.

Transferability

Transferability provides confidence that the study results could be generalized or

transferred to other contexts or settings (Forero et al., 2018; Nowell et al., 2017).

Korstjens and Moser (2018) explain that transferability refers to the ability of the reader

to assess whether the findings apply to their environment or situation, also known as the

transferability judgment. The clear implication is that the reader determines

transferability rather than the researcher (Korstjens & Moser, 2018).

As the researcher, I cannot know how individual readers will want to transfer the

findings to their environment. The only mechanism that a researcher can use to enhance

transferability is to provide detailed and thick descriptions of the participants and research

process so the reader can determine applicability to their setting (Nowell et al., 2017).

Transferability is primarily achieved through the analytic process of participant selection

and ensuring data saturation (Forero et al., 2018). I documented the selection of

participants and provided detailed descriptions of their selection and participation criteria

for this study to provide as much clarity to the reader as possible.

External validity increases when findings from the population can be generalized

beyond the individual case studied (Thomann & Maggetti, 2017). Care must be taken in

the selection of the population for the case study. Organizations must be selected that will

provide in-depth knowledge that is relevant, critical, and feasible for answering the

research question (Thomann & Maggetti, 2017). Additionally, Welch and Piekkari (2017)

point out that case studies generalize to a conceptual theory, rather than to a population.

By generalizing to the OIPT, this study gains external validity by a selection of a

population that is utilizing big data analytics, and the findings of the case study could be

generalized based on the OIPT.

Obtaining feedback on the interview protocol can enhance its reliability and

trustworthiness as a research instrument (Castillo-Montoya, 2016). The interview

protocols defined in Appendix B were reviewed by my committee to ensure the reliability

of the protocol and enhance the trustworthiness of my research. Pre-defined interview

protocols enhance reliability and increase the quality of data obtained from research

interviews (Castillo-Montoya, 2016).

Transition and Summary

This section restated the purpose of the study and provided information regarding

the methodology, participants, and decisions relating to the design of the study. Also

discussed were the ethical considerations, researcher role, and approaches for data

collection and analysis. Finally, I discussed validity and reliability. This section details

how I conducted my study relative to my selected methodology and design.

The next section will include a presentation of the findings of my study, as well as

the application to practice and future research. Social change will be discussed, as well as

personal reflections on this overall study and process.

Section 3: Application to Professional Practice and Implications for Change

The focus of this study was data governance strategies that data scientists are

using in practice for big data. I will present the findings from my data collection and

analysis and explain how these findings may contribute to the academic literature. I will

also explain how findings may affect social change and professional practice as big data

analytics matures as an industry practice. Finally, I will make suggestions for future study

and present my reflections on the overall study.

Overview of Study

The purpose of this qualitative multiple case study was to explore the big data

governance strategies employed by data scientists to provide a holistic perspective of

those data for making decisions. The data for this study came from semi-structured

interviews that I conducted with data scientists currently working on big data solutions in

the greater Salt Lake City, Utah area. The documents analyzed for this study are 10

participant interview transcripts and four process documents from three organizations.

This section begins with a brief overview of the study methodology, and then I present

the findings from the study.

Presentation of Findings

I began this study with the intent of answering the following research question:

What big data governance strategies do data scientists employ to provide a holistic

perspective of those data for making decisions? In this section, I will present the findings

from my data gathering and analysis relating to this question. The study consisted of 10

participants from three different organizations with an average of 9.4 years of experience

working in big data analytics. I used methodological triangulation by obtaining process

documentation from each organization in addition to the interviews held with the

participants. The following four major themes emerged from the analysis: ensuring

business centricity, striving for simplicity, establishing data source protocols, and

designing for security. I will present each of these themes in this section.

Theme 1: Ensuring Business Centricity

Data scientists cannot provide answers to business questions without first having a

solid understanding of the business question they are answering. The better the business

question is understood by the data scientist, the better they can design an analytic model

to answer that question. This theme includes four common subthemes: understanding

business needs, partnering with businesses, maintaining contextual awareness, and

minimizing data noise (see Table 1).

Table 1

Subthemes for Ensuring Business Centricity

Interviews Documents

Subtheme

Understanding Business Needs

Count References Count References

10 26 3 8

Partnering with Businesses

Maintaining Contextual Awareness

Minimizing Data Noise

10 43 3 8

7 21 4 6

10 34 2 4

Subtheme: Understanding business needs. Having a clear understanding of

business needs was a universal theme that came up in every participant discussion. Before

a data scientist can provide answers to business questions, they must understand the needs

that they are addressing. Understanding business needs is an important theme from both

analytic and data governance perspectives. Understanding business needs helps data

scientists identify appropriate data sources and elements that will be needed to

appropriately and completely address the business question. Each participant discussed

the importance of understanding the question they were being asked to answer before

they started assessing data sources and elements for analysis. There were a variety of

different ways through which they engaged with their business partners. Participants 1, 3,

4, 5, 6, and 7 discussed interviewing their business partners before initiating any analysis.

P5 said: “I do the fact-finding into their requirements and also the composition of what

their data looks like...Once I get adequate information … I make a determination on how

to construct and load the model.” P6 said: “Before developing a model, we interview or

have discussions with the business owners that have expertise in that space to tell us the

hypothesis or reasoning …that might have an influence on the decisions that [we need to

make].”

P2 and P8 referred to the long tenure of the data scientists in their organization

and indicated that this provides them with significant business context. Because of their

tenure in the business, they understand the business need immediately when presented

with a new question to be answered. P8 explained that when they receive a request for a

new model, “generally speaking, we know what we are looking for” because of their

experience in their specific business. All 10 participants discussed the importance of

having a clear definition of the business need as a critical step to deciding which data

sources they needed to connect with to get their data. P7 indicated that “the more

definition that we have upfront from the end-user of what they want to get out of it, the

better the process, the smoother it goes, the faster we can deliver that datamart to answer

their questions.”

Gardiner et al. (2018) explained that big data scientists are required to possess a

wider variety of skills than traditional technologists due to the multifaceted nature of this

new role. Alignment between overall information systems strategy and business strategy

is a critical precursor to overall profitability and competitive advantage (Shao, 2019). De

Mauro et al. (2018) found that while the main focus of the role involves the data itself,

associated analytical methods for transforming those data into usable insights are critical

to success because of the complex nature of big data. Data scientists cannot bridge this

gap without having a clear understanding of business needs to which they are responding

when creating their analysis or models from big data.

The concepts of the OIPT identify that pushing decisions down in the

organization yields better decisions by those closer to the tasks. The OIPT directly applies

to the gathering, interpretation, and synthesis of information within an organization for

decision-making (Tushman & Nadler, 1978). Data scientists require a clear understanding

of business needs to be able to interpret and synthesize big data for strategic decision-

making purposes. The increased ability within an organization to process information

using technology provides a mechanism for improving decisionmaking within that

organization due to the improved ability for employees to process more copious amounts

of information farther down within the organization (Kroh et al., 2018 ). It is only by

understanding business needs within the organization that data scientists can build useful

models for processing that information and disseminating it lower in the organization.

Subtheme: Partnering with the business. Maintaining a partnership with the

business was also a universal theme that all 10 participants mentioned as a critical success

factor. This subtheme is closely aligned with the subtheme of understanding business

needs. By creating partnerships between various technical and business roles, data

scientists can learn more about business needs. Developing a partnership between

business roles and data scientists establishes a level of trust and communication that

enables data scientists to better leverage data to help answer strategic business questions.

P4 said:

A few years ago...people weren't as educated about what data we had and what

kind of questions to ask. [In prior years], we used to come up with the questions. They'd

ask me some questions to see if we can get some data to answer that. And then usually we

will try to understand what they're going to do with the data. And most of the time that

really didn't lead to anything… But more lately, I think people are a lot more aware of

what data we have and what kind of questions to ask. P9 similarly discussed the process

of educating their business partners regarding the type of information that was needed and

available to answer their business strategy questions. Partnerships between business roles

and data scientists helps to communicate more clearly about what data elements are

available for analysis to answer business questions.

P8 discussed using a very iterative approach with their business partners as they

build their models, continually having conversations about what needs to be changed,

adjusted, or clarified as they work to answer the questions that are being asked. They

described their partnership in this way:

So if you ask for something today and I deliver it to you tomorrow, the likelihood

is it's not going to fulfill exactly what you wanted, even though you may have

attempted to communicate to me and I attempted to receive what you were saying.

At the end of the day, it's going to take us [several] exchanges back and forth.

‘Well, that's close, but I really wanted to do this.’ ‘Okay, let me go back and do

that...’, ‘That's a lot closer, but now that I'm using it, I see that it needs this other

element.’ So … what I am looking at is kind of bringing people in and working

with people to make sure we get it. You know, we get what has been requested.

We get that

right.

All 10 participants stated the importance of maintaining an ongoing partnership with their

business partners as a critical strategy to defining what data to collect and from which

data sources to collect those data.

Collecting and processing of big data require cross-departmental collaboration and

partnerships that do not exist with traditional data (Janssen et al., 2017). Successful

implementation of big data analytics requires new roles and organizations around the

business to interact, which had not previously needed to work together (Braganza et al.,

2017). Because of these new partnerships, some power shifts have been documented

within organizations that are effectively utilizing big data analytics to make fact-based

and real-time decisions (Popovič et al., 2018). Effectively utilizing big data to make

decisions within organizations is requiring new partnerships between data scientists and

business roles that were not as clearly required previously.

Uncertainty is defined as a key driver of information sharing within the OIPT.

Partnerships must be cultivated with the intent of sharing and utilizing information to

reduce uncertainty around business decisions. The success of any organization is

primarily affected by the competitive ability of managers to make strategic decisions in

the face of uncertainty and ambiguity (Intezari & Gressel, 2017). One of the key tenets of

the OIPT is that the greater the uncertainty of a task, the more information must be

processed by the decision-makers to execute that task (Galbraith, 1974). Since

organizations in the OIPT are simply information processing systems that collect, process

and distribute information (Zelt, Recker et al., 2018), and since that lack of information

within an organization leads to the uncertainty to which Galbraith refers (Gupta et al.,

2019), it follows that the greater the ability to partner across the various departments to

share information within an organization the greater will be their ability to reduce

uncertainty.

Subtheme: Maintaining contextual awareness. Contextual awareness of the

data elements is critical to providing a valid analysis. Seven of the 10 participants

mentioned maintaining contextual awareness as a key strategy for managing information

from big data. For the analysis to be useful to business decision-makers, data scientists

must ensure they understand the context of the data from the various data sources before

building a model from those data. Participants 1, 4, 5, 7, 8, and 9 each discussed

maintaining contextual awareness while gathering, preparing, and analyzing big data. P1

stated that “we record context around the data…there’s a tremendous amount of context”

and clarified that “with additional context, you have a much, much richer fingerprint that

is harder to accidentally duplicate…you have a greater chance of uniqueness when

retaining that context and metadata.”

The specific strategies for how to maintain this context vary by organization. P9

articulated that their organization maintains the data lineage throughout the analysis

process to provide contextual validity to the analysis. They explained that by maintaining

data lineage, they could determine that the “source provided additional information” and

that they could then check “to see if [they] can gather additional information from that

source.” P9 explained further that context can help them to know “if a particular column

changes, what reports downstream of that data are going to be affected and need to be

adjusted, rather than deal with the outcomes after the fact.” They explained that context

helps them to determine the scope of the analysis. Using context, they can determine

whether the report will meet with executive scrutiny as the executives attempt to correlate

information from various reports. Context is important to help those executives accurately

interpret the results of the analysis.

P1 and P5 indicated that metadata can be used to maintain context. P1, P4, and P9

indicated that the data source itself could provide some important context to the data. P7

suggested that correlating contextual data from various sources is “when the analytics

[are] a lot more powerful for us.” P8 mentioned that it is important to maintain

contextual awareness when automating results so that the context is not lost because of

the automated manipulation.

Big data analytics rely on fuzzy logic and inductive statistics (Paul et al., 2018).

Considering that some big data sources may be unknown or unvalidated some big data

analytics may be incomplete or inaccurate (Herschel & Miori, 2017). Without contextual

awareness, big data solutions can quickly become a liability rather than an asset.

Improper or incomplete analysis of big data can lead to misinterpretation of those data,

and lead to incorrect business decisions (Matthias et al., 2017). Alternatively, proper data

governance addresses the standardization, accuracy, and availability of those data (Wang

et al., 2018). Big data governance is critical to ensure that data are accurate, shared

appropriately, and protected per company policies (Cheng et al., 2017). This governance

includes maintaining contextual awareness of both the data elements and the data sources.

One of the central tenets of the OIPT is the need to close the gap of uncertainty by

processing more information within the organization. Duvald (2019) indicated that the

OIPT includes the mitigation of equivocality in addition to uncertainty. Duvald (2019)

also explained that equivocality arises from the existence of multiple and often

conflicting interpretations of data. The information dispersed throughout the organization

to close the gap identified by the OIPT must be accurate, or the analysis will not hold up

to the scrutiny of the business partnership. Maintaining the original context of the data

elements lends credibility to the analysis and reduces both uncertainty and equivocality of

that analysis.

Subtheme: Minimizing data noise. More data does not always mean better

analysis. Eight of the 10 participants discussed the importance of minimizing the noise

created from big data. P3 said:

A lot of the data is really just noise. You need to be sure to pick the things that

actually matter for the analysis... With the amount of data that we collect, you will

never be able to use it all…but a high percentage of the data that we are collecting

is pretty useless ... Just because you're collecting it doesn't mean that you can use

it for anything.

P1 stated something very similar when discussing the amount of information that their

organization gathers. P8 stated that reducing the noise is “a matter of narrowing the

broader data set down to what we're specifically trying to get to.”

P1, P2, P3, P4, and P9 stated that they do not attempt to analyze all of the data for

any particular analysis because there is simply too much of it. P2 said: “I don't think it has

been necessary to look at all the available data.” P3 added: “I would say that you don't

need to analyze all the data necessarily… I don't feel like you need to or should actually

use all the data.” P4 similarly stated: “I don't think we ever are analyzing all the available

or all the applicable data to answer any question. Practically, I don't think it's possible.”

These participants agree that analyzing all of the data is not an achievable goal

and stated that it is more important to narrow the data sets to the most applicable data for

the specific business question before attempting an analysis. P3 stated: “I think that right

now there are so many more questions that can be answered with smaller data.” Simply

adding more data into the analysis does not result in better analytics or better decisions.

One of the key responsibilities of data scientists is to reduce the noise and simplify

the data analysis process. This theme is confirmed in the literature, as several studies have

highlighted the idea that leveraging big data does not simply imply that obtaining more

raw data will lead to better information. The quality of decisions from big data is not

influenced solely by the amount of data collected but more so by how those data are

collected and processed (Janssen et al., 2017). Matthias et al. (2017) pointed out that

without appropriate analysis, it does not matter that large amounts of data are collected.

Value is extracted from big data when information is transparent and usable to the

organization (Ahmed et al., 2017).

The reduction of uncertainty by processing more information is a key concept of

the OIPT. Uncertainty in the context of the OIPT is defined as the difference between the

information processed and the information required to complete a task (Tushman &

Nadler, 1978). Merely increasing the information available does not resolve that

uncertainty (Gao et al., 2018). The success of an organization is primarily affected by the

competitive ability of managers to make strategic decisions in the face of uncertainty and

ambiguity (Intezari & Gressel, 2017). When data scientists have a clear understanding of

the uncertainty that the managers face, they can then design models with minimal data

noise to assist these organizational managers in resolving that ambiguity, thus resulting in

better decisions deeper within the organization. The ability for organizations to make

strategic decisions amid uncertainty and ambiguity relies on their ability to continuously

learn and reconfigure the organization’s knowledge base (Intezari & Gressel, 2017). Data

scientists can help their organizations face uncertainty when they understand the business

need, partner with the business to ask and answer the right questions, maintain contextual

awareness of the data elements, and minimize the overall data noise.

Theme 2: Striving for Simplicity

Because of the volume, variety, and velocity attributes of big data the analytics

can become complex very quickly. Complexity is one of the significant challenges with

big data as compared to traditional data (Jin et al., 2015). All 10 participants mentioned

the importance of simplicity in their responses in some manner. Process documentation

also reminds data scientists to reduce duplication and complexity in their design. The

fundamental basis for the OIPT is also centered around the uncertainty and complexity of

task completion and is based on the concept that organizations should be designed to

reduce that uncertainty and to enable decision-making (Feurer et al., 2019). Striving for

simplicity is a key skill that will help data scientists to deal with the inherent complexities

of big data while simultaneously offering solutions that can be maintained over time. This

theme consists of four subthemes: minimizing the data sources, using new tools,

simplicity of design, and using automation (see Table 2).

Table 2

Subthemes for Striving for Simplicity

Interviews Documents

Subtheme

Minimizing the Data Sources

Count References Count References

9 23 2 4

Using new tools

Simplicity of Design

8 23 0 0

10 28 3 9

Using Automation 8 15 4 11

Subtheme: Minimizing the data sources. Strategically limiting the sources of

data was a common concept that was raised among the various participants. This theme is

counter-intuitive when referring to big data. It is not uncommon for big data to infer

many data sources, but high volume and velocity can also occur with a small number of

data sources. This is particularly true of IoT where a single data source could have a

significant number of endpoints. Big data does not necessarily have to mean copious data

sources but could simply indicate many endpoints generating data. This theme is related

to the number of data sources and should not be confused with the number of

datagenerating endpoints. The reduction of the number of data sources was a concept

outlined as a strategy for maintaining the simplicity of design. This subtheme is also

closely related to the subtheme of minimizing data noise. One potential source of noise is

having too many data sources providing irrelevant or duplicative data elements.

Nine participants mentioned some form of limiting the data sources in their

responses. When explaining how they select appropriate data sources for any model or

analysis P1 explained that “there’s always going to be more data…we don’t have

hundreds of sources. We have a handful of core primary sources of data”. P2 identified a

handful of core sources of data that can provide answers to most of the business questions

asked at their organization. P3 reiterated similarly that “there are a finite set of sources…

we know what's available to us and what's easily accessible”. P6 articulated that they

strategically limit their data sources to only those that are valuable for the specific

business question being analyzed. All of the participants mentioned that there could be

more data available to them from other sources, but that minimizing the sources of data is

one of the key strategies that they follow to deal with the ever-increasing complexity of

various big data sources.

Determining the origin and transformation of various data elements obtained from

the myriad big data sources is currently one of the primary challenges of big data (Umer

et al., 2017). Because of the difficulty of verifying the various data sources associated

with big data the validity of analysis can be questionable (Hashem et al., 2015).

Additionally, external data sources are less likely to have internal data stewards or data

governance processes established to govern its use within the organization (Cervone,

2016). Strategically determining a minimal set of sources for big data analytics within the

organization can provide the data scientist with confidence in both the validity of the data

as well as confidence in the analytics.

The OIPT defines two mechanisms for closing the gap between the need to

process information and the ability to process information. Organizations can either

reduce the need to process information or increase the ability to process information

(Galbraith, 1974). Reducing the data sources in any analysis accomplishes both, and thus

reduces the gap from both sides simultaneously. The primary goal of closing the gap is to

reduce uncertainty. Tushman and Nadler (1978) defined uncertainty as the difference

between the information processed and the information required to complete a task. This

does not imply that simply gathering more information from myriad sources will result in

the reduction of uncertainty. Merely increasing the information available does not resolve

the uncertainty as individuals do not have unlimited processing capacity and will reach

information overload (Gao et al., 2018). By focusing on a limited set of data sources,

each of which can provide a high volume of data elements, data scientists can build

confidence in their analytics while both reducing the need to process more information

from other sources and increasing the ability to process more information by using big

data analytics.

Subtheme: Using new tools. Many new tools have been developed in recent

years to simplify the analysis of big data. These new and advanced tools are required to

extract information from big data (Intezari & Gressel, 2017). These new tools are also

important from a governance perspective as they help data scientists to better interact

with data in ways that traditional tools do not. Eight of the 10 participants discussed the

use of these new big data tools as a strategy for coping with the complexity of big data.

The adoption of these new tools is still an ongoing process. None of the participants

indicated that they are currently using all of the various tools available to them. All eight

of the participants that discussed tools indicated a desire or plan to start using more of

these specialized tools soon to further refine and enhance their data analysis.

P2 and P5 indicated that they believe the new tools will help them become more

efficient at interacting with big data. P2 said:

I think there's a skill gap that we're trying to address to understand what's changed

in…data engineering over the last five to 10 years and how do we adapt and take

advantage of that? We've primarily been stuck in kind of one technology stack for

a long time and we've become very efficient and good at it…so what's compelling

them to want to change and look at other alternatives and options?

P2, P3, P4, and P7 discussed their plans to leverage more cloud technology soon,

and they all expressed confidence that this would bring new capabilities to their ability to

effectively govern big data. P2 indicated that they “eventually want to put our data [in

the] cloud…right now, it is … on-prem. And so how can we leverage some of the new

paradigms and tools to get the flow of data a little bit quicker?” P9 said:

The volume that they talked about were…challenges [for companies] like Google

and Facebook and Twitter [because of] the volume of data they were processing.

We had to create tools to meet those challenges and we were then able to use

those tools. When they're working with a thousand times more data than we have,

and they've built a tool that can solve that solution, it usually handles all our

smaller problems.

As larger organizations adopt and utilize these new tools small and medium size

businesses can benefit from the use of these same tools to solve their big data challenges.

Traditional analytic systems lack operational efficiency to analyze big data

effectively (Jin et al., 2015). Traditional tools for data governance and analysis are

outdated and are no longer adequate for processing the vast amount of data available

today, making current strategies ineffective for handling big data (Bello-Orgaz et al.,

2016). Intezari and Gressel (2017) added that to extract information from big data, new

techniques, and advanced tools are required. Many organizations are starting to analyze

big data with traditional tools but are quickly establishing plans to implement new and

specialized tools.

As organizations learn to process larger amounts of data new tools will play a key

role in that process. Application of the OIPT leads to an understanding of the factors

influencing an organization's information processing capacity (Zelt, Schmiedel, & vom

Brocke, 2018), which include the tools used within the organization to process

information. Galbraith (2012) predicted that big data would become that technology that

would finally allow for the increased information processing alternative to becoming an

effective solution. Galbraith (2012) envisioned the implementation of big data solutions

would be driven by the constant interplay of increasing complexity and interdependence

of information and systems within that organization.

Subtheme: Simplicity of design. Process documentation from one of the

organizations sets a standard of the simplicity of design for data scientists. The

documentation urges data scientists to “make things as simple as possible, but no

simpler”. Another document instructed data scientists to “reduce complexity for greater

flexibility and lower cost”. This concept was universally reiterated by all 10 participants

and was discussed in three of the four process documents. Designing for simplicity is not

easy and should be considered a key skill to develop for any data scientist. P1 expressed

that big data analytics is not being done correctly today particularly because some data

scientists design overly complex solutions. P5 explained:

The way we process [big data] simplifies it and makes it easier to review…I think

the number one thing at play would be to use a simple a process as possible, frankly…I

have so much confidence in our processes... because it's not complex. It's not overly

complex. It's not muddied. There's not stuff all over the place. We keep it clean, lean, and

to the point. Straight forward. The processes and the model itself, it's doing the work. The

data doesn't have to do the work. P6 discussed the importance of keeping solutions

simple yet not shying away from complexity when it is required. Finding this balance is a

key skillset for any data scientist to effectively govern big data for analytic solutions.

The complexity of big data sources creates a strain on traditional analytic systems

because of the nature of structured and unstructured heterogeneous data elements.

Complexity in big data comes from the complex data structures, complex data types, and

complex data patterns that are inherent with big data (Jin et al., 2015). Heterogeneity adds

to the complexity of big data making big data solutions inherently difficult to manage

(Yang et al., 2017). Data complexity also adds to the overall complexity of big data as

compared to traditional data (Jin et al., 2015). Because of the inherent complexity of big

data, the solutions built from multiple big data sources can quickly become very complex.

Maintaining simplicity in the constructs of the data governance will assist in developing

more useful analytic systems. Jin et al. (2015) pointed out that among industry experts,

there is not a good understanding of how to deal with the complexity that comes from

complex data structures, complex data types, and complex data patterns inherent with big

data. Data scientists who can simplify these constructs form advantage over those who

are still struggling to understand the relationship between data complexity and

computational complexity as relates to big data processing

Through the OIPT Galbraith posits that organizations are inherently information

processing systems intrinsically programmed to manage uncertainty by gathering,

processing, and acting on information from within their environment (Jia et al., 2020).

The OIPT builds on contingency theories (Zelt, Recker, et al., 2018) which focus mainly

on the attributes within an organization that shape behavior (Chrisman et al., 2018). The

OIPT specifically focuses on the essential function of an organizational structure, which

is to facilitate the collection, analysis, and distribution of information to reduce

uncertainty within the organization (Hwang et al., 2019). As solutions are designed for

simplicity they will be more readily adopted within that organization. According to the

constructs of the OIPT complex information needs to be made available lower in the

organization to reduce the uncertainty of task completion (Feurer et al., 2019). The OIPT

helps shape behavior within organizations to take complex information systems and

simplify them so that information can be used farther down in the organization by

individuals closer to the task. Simplicity is a key attribute of this overall solution.

Subtheme: Using automation. The use of automation in the overall design can

add significant value to big data governance solutions. Because of the low veracity

attribute of big data, it becomes important to perform preprocessing to improve data

quality (Yang et al., 2017). With the rapid decay of the value of big data the need for

automation becomes clear. Eight of the 10 participants mentioned the use of automation

to assist in the movement, validation, and correlation of data from big data sources. P4

included automation as part of their job description when discussing their role in the

organization. P6 indicated that they “are responsible for developing automated models to

drive our pricing …strategies.” P7 and P10 stated that they rely very heavily on

automated notifications and alerts from their various automated systems that gather data

from the big data sources daily. P2 stated that they “have automated validation tests…

we've gotten more sophisticated over the past couple of years.” Process documentation

from one of the organizations reinforces the concept of automation by discussing the need

for systems to handle fluctuations of batch processing automatically and without

interruption of the overall data flow. P10 indicated that they have “automated processes”

and stated that they “highly rely” on them. Without the heavy use of automation, it would

be extremely difficult to keep up with the high volume, large velocity, and low veracity

of big data and it would be impossible to maintain the validity of the analytic solutions.

The value of some big data elements decays much more rapidly than traditional

data. To extract value from many big data solutions the data elements must be processed

immediately (Lee, 2017). This cannot be accomplished without automation. The volume

and variety attributes of big data also tend to make it very difficult to determine the

veracity of those data (Matthias et al., 2017). The processing of big data and the

validation of information from big data sources must be automated to provide value to the

organization. Attempting to accomplish big data solutions without automation would be a

futile effort.

The central tenet of the OIPT is the need for any organization to close the gap

between their need to process information and their inherent ability to process

information (Premkumar et al., 2005). Kemp (2014) identified a similar gap between the

amount of data that organizations can gather and the ability of that organization to

leverage that information in meaningful ways. Galbraith (2012) clarified that

organizations can deal with this gap in one of two ways: either they can increase the

capacity of the organization to process more information, or they can decentralize the

interdependence on that information. Combining heterogeneous sources of data into new

information sets is one way to reduce the uncertainty identified in the OIPT (Wang et al.,

2016). When considering the temporary value of raw big data and the need to reduce

uncertainty within an organization by processing more information it becomes clear that

automated solutions are required to process that information quickly while it still has

value.

Theme 3: Establishing Data Source Protocols

Clearly defined protocols for gathering and validating data from various sources

are critical strategies to respond to the volume, velocity, and variety attributes of big data.

In the absence of standards and well-defined processes for data validation, a data scientist

could not have confidence in their analysis. Incomplete data can result in partial analysis,

which can paint an inaccurate overall picture (Janssen et al., 2017). This theme consists

of three subthemes: following defined standards, establishing validation processes, and

reducing duplication (see Table 3).

Table 3

Subthemes for Establishing Data Source Protocols

Interviews Documents

Subtheme Count References Count References

Following defined standards

Establishing validation processes

Reducing duplication

8 18 3 17

10 43 3 4

5 16 3 6

Subtheme: Following defined standards. Eight of the 10 participants and three

of the four process documents referenced following defined standards. In this context,

standards refer to both industry standards and internal organizationally defined standards.

P2, P6, and P7 discussed the standard practice of establishing service level agreements

(SLAs) with their data source vendors for both timeliness and quality of data. P2 stated

that they have “strict SLAs around…external data”, and P7 added that they “set up SLAs

with our vendors of when that data needs to be available for us to pull in.” Establishing

SLAs for the timely access and gathering of data from the data sources assures that those

data will be available when needed for analysis.

P4, P5, P6, and P9 discussed various internal standards that have been defined

within their organization for both quality and timeliness of departmental deliverables.

These standards include such practices as peer review of solutions, daily review of

exception reports and alerts, and time commitments of deliverables between departments.

Process documentation from three of the case organizations referenced the flexibility,

economies of scale, and support of heterogeneous environments as benefits of following

established open standards. One of those documents explains that the “use of standards

provides the ability to leverage the knowledge and efforts of others. Risk is reduced,

proven solutions are implemented, and needless diversity and duplication are prevented.”

Another document explains that “standardization helps achieve economies of scale,

reduces complexity, and improves flexibility.” Defining and documenting standards

within the organization provides the data scientist with consistency and reliability in their

big data governance solutions. Adhering to published industry standards around security

and protection ensures the protection of data within big data solutions.

To be successful in the current environment organizations must methodologically

exploit their knowledge assets (Mahdi et al., 2019). This implies the establishment of

formal processes and procedures around their data assets. Organizations experience

substantial financial costs for little value when the organization implements data

governance poorly (Wang & Hajli, 2017). Knowledge management practices within an

organization are significantly and positively related to sustainable competitive advantage

(Mahdi et al., 2019). Establishing and following defined internal standards and adhering

to published standards provide the data scientist with a reliable and secure foundation for

their big data solutions.

The essential function of any organization as highlighted by the OIPT is to

facilitate the collection, analysis, and distribution of information to reduce uncertainly

within the organization (Hwang et al., 2019). The fundamental basis for the OIPT is

centered around the uncertainty and complexity of tasks to be completed in workflow and

based on the concept that organizations should be designed to enable decision-making

(Feurer et al., 2019). Because of the vast amount of data involved with complex big data

solutions organizations cannot provide reliable mechanisms for reducing uncertainty

without formal standards controlling those solutions.

Subtheme: Establishing validation processes. Validation of data sources is a

key protocol that participants identified for multiple data sources due to the low veracity

attribute of big data. The volume and variety attributes of big data make it very difficult

to determine the veracity of those data (Matthias et al., 2017). This subtheme also aligns

well with the subthemes of minimizing data noise and minimizing data sources. This is

because many big data sources are often insecure which leads to potentially inaccurate

analysis (Duncan et al., 2017). Developing processes for validating data sources is a key

governance practice for data scientists.

All 10 of the participants discussed processes that they have in place to validate

the various data sources in their analysis. P1 and P5 discussed applying a scoring

mechanism to each data source. P1 explained it as follows: “We score the data sources,

and we'll pick the [specific data elements] from the most accurate, the most available, and

the most trusted data source and that is adjudicated in real-time”. This scoring algorithm

provides them with confidence in the data source which can help them to determine

which data source to leverage as new questions are posed. The score is updated and

maintained as new information arises about the validity of the information from that data

source.

P2, P3, P4, P5, P6, P7, and P8 all mentioned performing validation on the data

themselves to ensure validity. They each do this in different ways with some performing

a line-by-line review personally and others performing spot checks and validating their

findings with peers and stakeholders. P6 said:

We also go through a process of validating data using multiple sources…we are

getting data from the competitor, but we also review what that data is telling

about the rates that we put in the marketplace…so we have the ability to compare

the results …[through] multiple rounds of validation.

P3 said:

Everything that we do is validated by us personally. We have data coming from

our website, our call center, from our stores, from Salesforce... hundreds of

different sources and anytime that we use this in our analysis, we validate every

column that we're using against some other source that we know is true to make

sure that the data [are] correct.

P9 articulated that they do “peer reviews to have multiple eyes take a look at an

area …We’ll also do two sets of analysis; have two different groups do the analysis and

compare the numbers.” P1 also described an automated validation process when they

said, “We go through a signing process to make sure the data is not being tampered with.”

Regardless of the specific strategy used every participant was concerned about utilizing

some defined, documented, and replicable process for validation of their big data

analytics. P5 said:

In the validation process, if there's anything that needs to be changed, that's where

you should start. You get that part sorted first. You get that approved first as

quickly as possible. If the data simply can't be validated or can't be approved for

whatever reason, then you can deal with it, number one.

Establishing a validation process for data sources and specific data elements is a critical

step in the big data governance process.

Organizations tend to have a false sense of confidence because of the hype around

big data (Herschel & Miori, 2017). Validating the origin and transformation of data

elements contained in big data is one of the primary challenges of big data (Umer et al.,

2017). The validity of big data analytics can be called into question due to the multiple

sources of big data which are less verifiable than was the case with traditional data

(Hashem et al., 2015). Data scientists must establish validation processes for their data

sources and for their analytics that provide confidence to the organization in the accuracy

of their big data systems. Big data governance is critical to ensure that data are accurate,

shared appropriately, and protected (Cheng et al., 2017)

The OIPT includes the mitigation of both uncertainty and equivocality (Duvald,

2019). Management's role in any organization is to reduce both uncertainty and

equivocality for their organization (Bartnik & Park, 2018). Ensuring that the analysis of

information for task completion and decision making is valid, vetted, and accurate is of

critical importance for good decision making. Knowledge management practices within

an organization are significantly and positively related to sustainable competitive

advantage (Mahdi et al., 2019). If the consumers of the big data analysis are not confident

in the results, they will not be reducing uncertainty but creating more uncertainty. The

main goal of OIPT is to reduce uncertainty, not increase it.

Subtheme: Reducing duplication. Another strategy that arose from the study

was that of reducing the duplication of data. This subtheme aligns well with the

subthemes of minimizing data sources, minimizing data noise, and striving for simplicity.

Copying data from one location to another is time-consuming and introduces risk to the

validity of the data itself. P1 and P5 discussed using metadata to refer to the primary

source of the data elements as a key strategy to reduce the duplication of data. P1

described the value of not duplicating data in this way: “We try and not lose contact with

the source data…once you've made a copy of it and moved it, it could have been from

anywhere and anything. And we keep our finger on the pulse at the data source”. P10

articulated a similar concept when they explained that they “make sure that I do not

duplicate the data…the same information should not appear in different tables in the form

of different dimensions and different facts. That is the main thing I would look at. No

duplication.”

Two process documents from two different organizations also mentioned the

importance of leveraging metadata as a key strategy for big data governance. One

document urges data scientists to “minimize redundancy and reduce duplication” because

it “helps reduce complexity and promotes greater efficiency”. Utilizing metadata reduces

risk by leaving the source data untouched and maintaining information about where the

source data reside. P5 reinforces this concept when they explain that “I think that number

one…would have to be metadata and the critical importance of that. It has to be clean,

crisp, consistent information and you need to get as much detail as possible about that

dataset”. By refraining from copying the data and analyzing the data where it resides data

scientists can avoid these traditional challenges.

Challenges arise with big data analytics from the issues relating to data that are

too vast, unstructured, and moving too fast to be managed by traditional means (Zakir et

al., 2015). Using traditional strategies of copying data for analytics purposes creates

significant challenges due to the rapid decay of many big data elements. Leveraging

metadata is one way to normalize unstructured big data elements for analytic purposes

(Yang et al., 2017). Traditional data copy and retention methods break down when

applied to big data because of the attribute of high velocity. Storage of big data becomes

challenging as transient data flows through the analyzing systems (Yang et al., 2017).

Minimizing the copying and duplication of data is a key governance strategy for data

scientists as they transition from traditional analytic methods to big data analytics.

The OIPT directly applies to the gathering, interpretation, and synthesis of

information within an organization for decision making (Tushman & Nadler, 1978). The

gap identified by the OIPT directly relates to the ability of an organization to gather and

analyze data. There is a limit to the amount of data that an organization can gather and

store, but by analyzing the data where it resides organizations increase their ability to

analyze more data. By designing systems that analyze data without requiring that those

data be copied onto local storage devices, organizations increase their ability to process

larger amounts of data than what was previously available to them due to the physical

limitations of local storage. Ensuring data source protocols by following defined

standards, establishing validation processes, and reducing duplication of data can increase

the validity of big data analytics by data scientists.

Theme 4: Designing for Security

Protection of security and privacy are critical data governance concepts for any

solution involving personal data. Data scientists are among the first line of defense when

it comes to protecting both the raw data and the individual privacy of the consumers who

are represented in those data. With the ever-increasing threat of a data breach, any

organization must ensure the protection of the information within its control. This theme

consists of three subthemes: segregation of duties, using encryption, and protecting

private information (see Table 4).

Table 4

Subthemes for Designing for Security

Interviews Documents

Subtheme

Segregation of duties

Using encryption

Count References Count References

9 35 3 17

3 9 4 10

10 15 4 10

Protecting private information

Subtheme: Segregation of duties. Security and privacy are best addressed in the

data governance process before the data scientists have access to those data for analysis.

When asked about the protection of the data nine out of 10 participants discussed the

segregation of duties concerning the data lineage within their organization. Additionally,

three of the four process documents referenced segregation of duties. In every

organization studied there were either data owners or data stewards who were responsible

for protecting access to the information. P2 described that data custodians in their

organization prepare the various data environments for the data scientists to access and

stated that “when the environment is ready for [data scientists] to work in, they wouldn't

see [private] data”. If a data scientist needed additional information to perform their

analysis, they would need to obtain permission from the data owner or data steward

before they were able to access the information. P3 articulated this process as follows: “if

there's something that we want that isn't available, we have to go through steps to get that

from the BI team…[they would] be the ones that need to give us access to that”. Security

controls on the various systems of record were active to prevent such access without

appropriate permission.

P3, P7, P8, P9, and P10 also discussed having role-based access controls within

their systems that restrict access to private data based on role. P3 indicated that the “BI

team…would be considered more of the gatherers, and then on the data science side we

do more of the analysis”. P8 highlighted a role distinction between the data architect and

reporting writing roles as follows: “I'm not a report writer, I'm not an architect... I kind of

come before the report, I pull together the requirements”. Process documentation from

one of the organizations also specifies that “effective [segregation of duties] ensures

access to company…data is restricted to only authorized and appropriate personnel”.

They also discussed the segregation of duties among the IT team with database

administrators controlling access to the data while data scientists, business analysts, and

business intelligence roles all accessing only the information that they require to

accomplish their job.

Maintaining appropriate segregation of duties between those who are responsible

for data access and those who analyze information is critical to the protection of the data.

P6 spoke of the inherent conflict that exists between data owners who want to protect

data and restrict access and data scientists who want to access and review as much data as

possible. P6 said:

There is some conflict between data science and data warehouse on what data is

available to the data science team. The data science team wants access to …more

data, more raw data, more extensive data. There is some unwillingness to provide

access to all of the data that is needed…there are concerns about privacy and

other aspects.

This is a healthy conflict that forces data scientists to justify all of the access to data in

terms of a business question that they are working to answer. This highlights the

alignment between this subtheme and the subtheme of understanding the business need.

The primary ethical issue is not whether to collect the information, but how and

when it is ethically responsible to analyze that information (Mai, 2016). Herschel and

Miori (2017) stated that data sources that had not previously had an issue with privacy

may end up threatening privacy once combined with other data sources in the big data

construct. Big data analytics must be comprised of various roles and assignments of

specific responsibilities, such as technical data steward and business data steward (Begg

& Caira, 2012). Data owners need to stay focused on their role of protecting individual

privacy and data security of the information that they gather. Maintaining segregation of

duties between those who are responsible for gathering the data, those who are

responsible for analyzing the data, and those who are responsible for protecting the data

creates natural boundaries around data governance within the organization that lead

toward better security overall.

Segregation of duties is not specifically addressed in OIPT. However, the

concepts related to pushing decision making down in the organization are relevant to this

theme. The premise of OIPT is to minimize uncertainty within an organization by closing

the gap between the amount of information needed for a task and the ability to process

information around the task (Premkumar et al., 2005). This means that according to the

constructs of OIPT those data need to be made available lower in the organization (Feurer

et al., 2019). Making information available lower in the organization implies having a

larger audience for the information which in turn implies a greater need for security and

privacy controls. By granting information to the individuals who need it when they need

it, role-based access to that information becomes an important security control to limit

access to private information. The net effect of pushing the decision making down within

the organization aligns with the concept of segregating duties and limiting access to

private and protected information to only those individuals who need to see those private

data as part of their role within the organization.

Subtheme: Using encryption. Encryption of data at rest and data in transit has

long been a foundational security control of data governance. Many regulated industries

require that encryption is implemented as one of many security controls to protect data

against unauthorized access. When asked about the protection of the data three of the 10

participants specifically mentioned encryption. P5 indicated that “the data is encrypted.

So, it is not readable or hackable in any way. We keep that consistency and the data is

always protected.” P8 said:

Files don't go out the door without some form of protection. …our file transfers

now happen via SFTP. We also wrap those files in a PGP or GPG encryption…

we also want to make sure… that wherever that data is sitting and wherever it

eventually lands that it is encrypted at rest. So, it is encrypted as it sits, it is

basically double encrypted as it goes. And then it's still encrypted as it sits at its

destination.

Use of encryption is a common practice to protect personal information for both data at

rest and data in transit.

The remaining seven participants stated that their security team has controls in

place for the protection of the data which does not rule out encryption being used as part

of that protection. In addition, three of the five process documents referred to encryption

of the data as a key security control. One of those documents indicated that “it is

important that we comply with security requirements, laws, and regulations as we design

our systems.” P5 also discussed the idea that consumers are concerned about sharing their

data with companies who are not protective of those data and stated that encryption helps

provide a level of comfort for their customers.

Because of the volume and velocity of big data traditional encryption at rest is

inefficient when applied to big data solutions (Yang et al., 2017). Encryption in transit is

also questionable when applied to big data as many data sources associated with big data

lack advanced security controls (Sivarajah et al., 2017). It is because of these challenges

that data scientists must take additional care within their environments to proactively

enforce encryption of data both at rest and in transit to enhance the security of big data

solutions. Encryption technology itself must also improve to handle the speed

requirements that businesses are demanding from big data analytics.

Similar to the theme of segregation of duties the specific concept of encrypting

data is not directly addressed in the OIPT. However, the safeguarding of information is

critical to maintaining confidence in any information-sharing solution within an

organization. The OIPT defined the essential function of an organizational structure to

facilitate the collection, analysis, and distribution of information (Hwang et al., 2019).

Alignment between overall information systems strategy and business strategy is a

critical precursor to overall profitability and competitive advantage (Shao, 2019). Data

governance of that information refers to the control and authority over data-related rules

and accountability of individuals and systems accessing and utilizing those data (Hashem,

et al., 2015). Ensuring the protection of those data through industry-standard security

controls such as encryption is crucial to the ability of an organization to share information

widely throughout the various organizational levels of the organization. This is critical to

the success of the concepts outlined in OIPT.

Subtheme: Protecting private information. Protecting the privacy of individuals

is critical to organizations that are analyzing big data. Individuals are revealing very

personal information on social networks and through IoT devices often unconsciously

(Gao et al., 2018) and without their knowledge or consent (Mai, 2016). When

implementing data governance over big data it is critical for the data scientists to be

aware of personally identifiable information (PII) within their system and to proactively

put controls in place to protect those data.

When asked about privacy 10 out of 10 participants indicated that they are

cognizant of privacy and that it is important to their organization. P4 explained that “we

don't use customer identifying information for any analysis, but …there's one process that

uses that and we hash [the PII to] protect the data.” Additionally, one of the process

documents describes that the system in use in that organization will automatically

determine whether the information being requested is PII and if so, will return obfuscated

data rather than the raw data. P8 discusses the importance of ensuring that appropriate

data sharing agreements are in place before sharing any PII. P1 explained that they obtain

consumers' permission prior to accessing their information: “if I want to look at Patient

X's x-rays, Patient X has to authorize that”. P2 explained that they scrub the PII from the

system on-demand: “we've invested in some scrubbing technology. So, when we, even on

the application development side, if an engineer has to pull data down to a local

environment…the PII’s scrubbed.”

Regardless of the specific manner in which the personal data are protected, it is

critical for organizations to understand what personal information they have access to and

to put processes and controls in place to protect that information. Big data can include

many elements of personal information (Lee, 2017). 99.98% of Americans could be

reidentified with any data set with as few as 15 demographic data elements (Rocher et al.,

2019). Data scientists analyzing big data can no longer assume that consumers have

provided informed, rational consent for their data to be used by the organization (Mai,

2016). This means that organizations analyzing big data have an ethical responsibility to

understand how to use the data while protecting the privacy and confidentiality of those

data (Herschel & Miori, 2017).

Tushman and Nadler (1978) articulated that the OIPT directly applies to the

gathering, interpretation, and synthesis of information within an organization for decision

making. If individuals have not provided their consent for such use the organization could

be at risk. One of the general concepts of the OIPT is that as lateral relationships are

established, the number of decisions referred upward is reduced, which then leads to

more effective decision-making (Feurer et al., 2019). This means that according to the

constructs of the OIPT those data need to be made available lower in the organization

(Feurer et al., 2019). Hwang et al. (2019) added that OIPT highlights the essential

function of an organizational structure, which is to facilitate the collection, analysis, and

distribution of information to reduce uncertainty within the organization. Considering the

privacy requirements enacted in recent legislation in the form of the California Consumer

Protection Act the organization must focus on privacy before providing access to such

information for decision making within the organization. By implementing the strategies

of segregation of duties, use of encryption, and protection of privacy data scientists can

successfully design their solutions for security.

Applications to Professional Practice

By implementing the concepts outlined in these four themes, data scientists can

implement governance practices that will assist them in working with big data. The four

themes described in this study guide data scientists to provide a foundation for big data

analytics that will ease the transition from traditional data analytics to the use of big data

for decision making within the organization. These practices can help provide better

information and insights to those individuals who are closer to the work within the

organization as theorized in the OIPT. The implementation of the practices associated

with these four themes will allow individuals within the organization to work more

autonomously and to close the gap between information needed and information available

for the front-line employees. As described in the key tenant of the OIPT this application

to professional practice should result in better efficiencies throughout the organization.

Data scientists who use the strategies described in these findings could improve

their effectiveness as agents of change for their organization. Adopting big data

governance practices that provide a balance between value creation and risk mitigation is

imperative to organizations (Hashem et al., 2015). Learning and implementing these

practices could make data scientists more valuable to their organization which could

result in an increase in both efficiencies of the organization and improved value of the

data scientist. By implementing the governance practices outlined in this study the data

scientist could improve their position within the organization by bringing them closer to

the business itself. This increased focus on the business will benefit both the role of the

data scientist as well as the efficiency of the solutions that the data scientist can provide

to the organization.

Implications for Social Change

If the implementation of these practices results in the efficiencies theorized in the

OIPT, the result would be better decisions being made by the organization. Better

decision making within the organization would occur because the information would be

made available lower in the organization and closer to those who need the information to

do their job. Making better information available lower in the organization would in turn

result in increased responsiveness and time to market for the organization, which could

provide the overall organization with a significant competitive advantage. Competitive

advantage could lead to increased profitability for the organization. The cumulative effect

of these practices could lead to more employment opportunities for the communities

where these mid-market companies are based.

Implementation of these practices could also result in more efficient use of data

within the organization, resulting in increased profitability for the overall organization.

Increasing the efficiency of data use means that the potential for social change is the

increased ability for companies to meet the needs of their customer base because they

now have an increased understanding of the meaning of the data that they have. These

business improvements could benefit the local communities where the company is

located by allowing the employees to earn more money and improve their way of life.

Improved business performance could also lead to better employment opportunities and

improved customer experiences. Effective use of the big data themes in this study could

benefit customers by providing them more specific and targeted solutions to their needs

by the small companies that are part of their local communities. In sum, the ability to

effectively utilize big data in mid-market companies could be an equalizing force that

could benefit customers, employees, and owners of all types of mid-market companies.

A social change could also result from data scientists enacting the security

strategies outlined in the findings from this study. Weak security controls around big data

can lead to financial loss and reputational damage (Lee, 2017). Implementing the

security controls outlined in the findings can help ensure better security of big data

solutions and thus protect the reputation of the institution and mitigate financial losses.

By implementing the strategies relating to the segregation of duties, encryption of data,

and protection of personal information, data scientists can mitigate the contemporary

concerns relating to the use of private information in big data analytics.

The use of multiple data sources that may themselves be insecure leads to

potentially insecure solutions (Duncan et al., 2017). During the completion of this study,

Simon Weckart tricked Google Maps into reporting that 100 cell phones in a wagon were

a traffic jam demonstrating the ease with which some big data systems can be fooled by

injection of IoT devices (Torres, 2020). Validation of data sources can mitigate this

concern and enhance the overall security of big data analytic solutions, which could result

in less possibility of injection of malicious and falsified data streams into AI-driven

analytic systems. By protecting the privacy of individuals and ensuring the validity of the

data sources, positive social change can be accomplished in the embodiment of trusted

solutions for autonomous artificial intelligence systems.

Recommendations for Action

Policies, principles, and frameworks around how big data are used and leveraged

within an organization are posing enormous challenges for those organizations (Hashem

et al., 2015). The results of this study validate this concept from literature. Data

governance is critical for big data, and data scientists cannot add it as an afterthought

(Cervone, 2016). By following the findings of this study, data scientists can thoughtfully

and proactively implement data governance into their big data solutions from the start. To

be able to extract value from raw data, organizations require a well-defined, systematic

approach to governing their data (Song & Zhu, 2016). Data scientists can leverage the

data governance strategies outlined in the findings of this study to design appropriate

governance models for their big data initiatives.

Data scientists should review the findings of this study to determine how best to

apply these practices into their big data solution architectures. The lack of formal data

governance processes was one of the primary reasons that organizations fail to utilize

their data effectively (Posavec & Krajnovic, 2016). Implementation of these themes can

guide data scientists who are trying to move from traditional data analytics to big data

analytics. To be able to extract value from raw data, Song and Zhu (2016) noted that

organizations require a well-defined, systematic approach to governing their data. The

themes outlined in this study provided a foundation for such a systematic approach to big

data governance. Begg and Caira (2012) argued that there is no evidence to date of a

standard implementation framework for data governance. This study helps to set a

foundation for that standard implementation framework. As data scientists implement

these themes, they will also be establishing a set of best practices within the industry.

Data scientists should review the themes identified in this study and should look for ways

to implement these concepts as they implement big data analytics within their

organizations.

Recommendations for Further Study

This study focused on the general data governance practices of data scientists who

are utilizing big data to help make decisions within their organization. Data governance

of big data was a broad topic about a subject that is in its infancy within current practice.

Big data by itself is of little use when considering the application to an organization. The

power of big data comes when combined with both IoT and Machine Learning. I would

recommend that future studies focus on the combination of these three technologies. IoT

is creating additional data sources for big data, and machine learning is starting to provide

faster analysis of those data. As future studies focus on the powerful combination of big

data governance combined with IoT and machine learning, more insights will emerge

regarding the benefits of big data to the decision-making process within an organization.

Security, privacy, and ethical use of big data are also critical aspects of data

governance when implementing big data solutions. Re-identification of previously

obfuscated and anonymous data has been documented in multiple cases (Herschel &

Miori, 2017). Additional study is warranted in these areas to determine the effectiveness

of the security controls and the extent to which the strategies outlined here are effective in

reducing the ability to correlate big data elements into data sets that contain private

information not previously permitted by the data owner. The security strategies outlined

in this study should be further studied to determine their effectiveness.

Reflections

As I reflect on this process, my thoughts turn first to the current state of the

practice of big data governance. Documentation regarding data governance processes

lacks in all the organizations that I studied. Each organization had documentation

regarding their security processes, and one had design documentation regarding their

overall approach to big data analysis. Still, very little of the documentation was focused

on the data governance aspects of how to handle data within their organization. Most of

the documentation was focused on overall IT practices and only lightly touched on the

data specifically. While design specifications are important, it is also important to

document decisions around the governance process. It is important to have documentation

stating why certain governance decisions were made and how to implement those

decisions in practice.

It was also clear to me that the participants validated the lack of clarity described

in literature around the definition of big data. Some of the participants did not think they

were working with big data until I discussed the definition of big data with them. There

are also still many different perspectives on how to handle big data in the future. One

participant indicated that they should not collect any more data until they can effectively

utilize all of the data collected already, and another participant opined that they should

collect much more data and simply store it until they can determine the best use for it.

While there was very good alignment around the four themes that emerged from this

study, there were also many conflicting ideas about some of the other strategies in use

today.

Another concept that was identified throughout this process is the fact that the

data scientists that I spoke with take their responsibilities very seriously. Each one was

very conscious of the fact that they do not own the data that they are accessing, and the

need to maintain privacy and security is paramount to them. As I reviewed my notes from

the interviews, I noticed that there are several occasions where I noted from their body

language that the participant took their role as data steward seriously and that they clearly

understood that they were interacting with other people's data. The stewardship that these

participants feel became clear early in the process when some potential participants

declined to participate because they were concerned about sharing too much information

about the data that they gather. The concept of data stewardship versus data ownership is

well understood by the individuals that I interacted with for this study.

As I reflect on the process, my thoughts also turn to the environment in which I

conducted the study. I entered the data gathering phase of this study during the COVID19

pandemic of 2020. The environmental changes that came as a result of the pandemic

caused some added challenges to the data gathering process that I had not previously

anticipated. Some of the partner organizations had to delay scheduling interviews because

their employees were busy responding to the challenges of providing support to their

company while employees transitioned to working from home. Others had to delay their

participation because their department had gone through a layoff, and the remaining

employees were dealing with the emotional stress of having lost some co-workers while

having to take on the added work that was left by these departing employees. These

various effects of the pandemic caused the data gathering phase to take longer than I had

initially anticipated.

I was surprised at the difficulty of finding partner organizations willing to

participate in this study. During the proposal phase of my study, I had identified two

organizations that I was planning to use as partner organizations. I spoke with the

leadership of both organizations, and they both seemed excited to participate. Once I

obtained the approval of the proposal and approached these two organizations to obtain

their Letter of Cooperation, both backed out. In one case, the leadership was

uncomfortable with the fact that they would not be able to review and approve the

individual interview content before inclusion in my data analysis. They were concerned

that something might be revealed that might reflect poorly on their company, even though

I assured them that confidentiality would be paramount. As I searched for other

organizations to participate, it was challenging to find organizations mature enough in

their big data processes to participate in this study. In one case, the potential partner

company had been gathering quite a bit of data from IoT devices, but they were not yet

doing any analysis of those data. The difficulty of locating companies that are currently

using big data to make decisions is a clear indication that we are still incredibly early in

the adoption of big data analytics as a true strategic capability within mid-market

organizations.

After I found organizations that met the criteria, I then met with the challenge of

obtaining permission from the potential participants to agree to the interview process.

Several potential participants declined to participate because they were uncomfortable

sharing details of their data governance practices outside of their organization. As I

discussed the process with potential participants, it became clear to me that many of the

individuals involved in the data gathering and analysis process take their role in

protecting information very seriously. Thus, they are more reluctant to discuss details of

their process with others. While not completely surprising, this reluctance to share details

regarding their big data governance processes did pose an additional challenge to the

overall study process.

Summary and Conclusions

The focus of this study was to explore big data governance strategies that data

scientists are currently using in practice. I presented four major themes that provide

insights into those strategies: ensuring business centricity, striving for simplicity,

establishing data source protocols, and designing for security. Implementation of these

strategies can assist mid-market organizations in making the transition from traditional

data analytics to big data analytics, which could, in turn, help those organizations to be

more profitable by gaining competitive advantages. I explained the possibility of social

change in the way that individuals' private information is gathered as part of a big data

strategy. Following the strategies outlined in the four themes of this study for big data

governance can lead to the improved overall protection of individual privacy.

Students also viewed