Content uploaded by Jung P. Shim
Author content
All content in this area was uploaded by Jung P. Shim on Jul 26, 2018
Content may be subject to copyright.
Communications of the Association for Information Systems
Volume 37 Article 39
10-2015
Big Data and Analytics: Issues, Solutions, and ROI
J. P. Shim
Georgia State University, jpshim@gsu.edu
Aaron M. French
University of New Mexico
Chengqi Guo
James Madison University
Joey Jablonski
Dell, Inc.
Follow this and additional works at: h=p://aisel.aisnet.org/cais
<is material is brought to you by the Journals at AIS Electronic Library (AISeL). It has been accepted for inclusion in Communications of the
Association for Information Systems by an authorized administrator of AIS Electronic Library (AISeL). For more information, please contact
elibrary@aisnet.org.
Recommended Citation
Shim, J. P.; French, Aaron M.; Guo, Chengqi; and Jablonski, Joey (2015) "Big Data and Analytics: Issues, Solutions, and ROI,"
Communications of the Association for Information Systems: Vol. 37, Article 39.
Available at: h=p://aisel.aisnet.org/cais/vol37/iss1/39
C
ommunications of the
A
I
S
ssociation for nformation
ystems
Panel Report ISSN: 1529-3181
Volume 37 Paper 39 pp. 797 – 810 October 2015
Big Data and Analytics: Issues, Solutions, and ROI
J. P. Shim
Department of Computer Information Systems, Georgia
State University
jpshim@gsu.edu
Aaron M. French
Department of Management Information Systems,
University of New Mexico
Chengqi Guo
Department of Computer Information Systems &
Business Analytics, James Madison University
Joey Jablonski
Director of Product Management for Analytics & Big Data
at Dell, Inc.
Abstract:
Recently, the topic of big data and analytics has received renewed attention from academia and practitioners. There
has been an increase in demand for skills in big data and analytics due to the increasing speed, variety, and volume
of information. Several research reports have shown that big data and analytics remain top priority for CIOs. A recent
study shows how a company accurately predicted a teen girl’s pregnancy via the company’s big data algorithm.
However, there are dark sides to big data and analytics. A panel discussion addressed topics concerning ho
w
companies ensure that big data projects clearly define measurable goals up front, methods that companies use to
ensure maximum return and most effectively, and ways that companies evolve culture, processes, and technology to
simultaneously maximize return. Most companies are looking at how they can effectively manage their business more
through using their data assets. Companies today target an average return of $3.50 dollars for every dollar spent on
big data projects. However, most are only returning a fraction of that today, which leaves room for improvement and
the possibility that organizations will push back against new analytic technologies. In this paper, we cover these topics
that a panel of researchers at AMCIS 2014 in Savannah, GA, discussed.
Keywords: Big Data, Analytics, Structured and Ill-Structured Data, Specific Issues, Big Data Projects, Return on
Investment (ROI).
This manuscript underwent editorial review. It was received 05/25/2015 and was with the authors 1 month for 1 revision. Matti Rossi
served as Associate Editor.
798 Big Data and Analytics: Issues, Solutions, and ROI
Volume 37 Paper 39
1 Introduction
Big data and business analytics are popular topics gaining significant attention from practitioners and
scholars alike (Chen, Chiang, & Storey, 2012). A recent paper has showed that 2.5 exabytes (2.5 billion
gigabytes) of data is created every day (McAfee& Brynjolfsson, 2012). A great number of firms are looking
at how they manage their business more effectively through using their data assets. Companies today
target a return of $3.50 for every dollar spent on big data and analytics projects. However, many projects
only yield returns of $0.50, which leaves room for improvement and the possibility that organizations will
push back against new analytic technologies. Analytics and Big data can be a revolutionizing force for
numerous industries. Though website analytics itself has been a topic of discussion for over a decade, 90
percent of the world’s data has been created in the past several years. This significant increase is due to
the increased sophistication of mobile technology and social media networks. Further, as these mobile
computing devices expand globally, the volume of this digital content will only increase. Various platforms
allow individuals to be tracked along with real-time updates. The massive amount of customer data that is
generated and collected on a daily basis holds the potential to drive profits. Customers are more educated
than ever before and currently have a vast amount of information readily available. As a result, businesses
are increasingly reliant on big data and analytics.
Big data can have a puzzling element because not all data is easily discernible. One can categorize data
into structured and ill-structured information. Structured data mirrors that of information obtained from a
direct transaction, while ill-structured data comes in a form often obtained from social media, such as
Twitter feeds, “retweets”, Facebook posts, and “likes”. Previous information systems used by
organizations, such as data warehouses and ERP systems, typically processed structured for reporting
and decision making, while semi-structure and ill-structured data was left behind (Negash, 2004). With the
increased sophistication of data storage and processing tools, this data that was once unusable is
providing valuation information and resulting in new business intelligence. As social media platform users
grow, so will the 3Vs (volume, variety, and velocity) of this information. These data components offer key
performance indicators in their own way. Knowing one’s customers, their behaviors, and markets and
altering quickly to accommodate changes are imperative to adapting to instantaneous issues that arise
and to ensure customer satisfaction and increase profitability.
The future of all industries relies heavily on firms positively leveraging big data and analytics. In our digital
society, one can purchase goods and services from a remote location. The way in which firms interact
with customers will have to incorporate specific tailoring of their products, platforms, and internal
management. Such firms can invest in big data, analytics, and other technologies to track the customers,
provide easily accessible platforms on various devices, store this gathered data, and adapt business
practices to create experiences specifically to accommodate various micro-market segments. Leveraging
big data can also increase our ability to research problems at the macro level (society) by evaluating large
amounts of comprehensive data at the micro level (individuals) by using new tools equipped for handling
semi-structured and unstructured data (Agarwal & Dhar, 2014).
In the past years, there has been an explosive growth of user data in terms of volume because of social
media’s proliferation. Such a data avalanche poses serious challenges and significant competitive
advantages to business entities that rely on distributed or cloud computing to handle user requests and
respond rapidly. “Perception is reality” seems to have become the slogan of big data practices, which
differentiates them from traditional database sectors (e.g., relational database) where consistency
requirement is ubiquitous.
However, there are dark sides to big data. While big data has received a lot of attention for its potential,
we must face several of its challenges. For instance, privacy is a major concern in terms of big data. The
massive amounts of data that organizations collect has led to the development of digital dossiers at a level
of detail that we have never seen before. One can use these digital dossiers to uncover intimate details
about an individual such as sexuality, menstrual cycles, and whether a woman is pregnant or not. It raises
ethical concerns whether or not companies have a right to mine for such personal information that can be
used for marketing purposes. Additional concerns could be the misuse of data, which can result in
misleading truths or the introduction of the digital divide 2.0 (i.e., difference between those who have
access to big data and those who do not). In addition, one of the most important issues with regards to big
data is whether companies are seeing a return on their investments (ROI) in it.
This paper proceeds as follows: in Section 2, we discuss the current status of big data and analytics. In
Section 3, we describe how to interpret big data’s benefits. In Section 4, we discuss the dark side of big
Communications of the Association for Information Systems 799
Volume 37 Paper 39
data, including big data cases, digital dossiers, ethics and privacy, and the digital divide 2.0. In Section 5,
we discuss maximizing the return for big data projects, which includes improving ROI for big data projects
and big data project fundamentals. Finally, in Section 6, we concludes with recommendations for future of
big data and analytics.
2 Current Status of Big Data and Analytics
2.1 Timeline and Data Storage with Complexity
The big data timeline illustrates how the first big data problem began in 1890 when the U.S. Government
Census was carried out. In 1965, the first big data center was established when the U.S. Government
needed a place to store 742 million tax returns and 175 million sets of fingerprints. The birth of the World
Wide Web in 1989 was a milestone for the big data that was housed and accessed on the Internet on a
daily basis. Big data was defined around 1997-2001. By 2002, the world contained in excess of three
billion documents with an estimated 80 percent of it being ill-structured and predictions of the number of
ill-structured documents doubling every eight months (Negash, 2004). In 2004, big data tools, such as
Hadoop, helped individuals and firms to actually understand big data and realize its potential (Alteryx.com)
Big data is often characterized by volume, velocity, and variety (the three Vs) (Gillon, Aral, Lin, Mithas, &
Zozulia, 2014; Goes, 2014; Hashem et al., 2015; Lycett, 2013; McAfee & Brynjolfsson, 2012; Wixom,
Ariyachandra, Douglas, Goul, & Gupta, 2014). Researchers have also extended this model to include
veracity (Gillon et al., 2014; Goes, 2014) and value (to make five Vs) (Hashem et al., 2015; Lycett et al.,
2013). In detail:
Volume refers to the amount of data being collected ("how much"). Data sets today can extend
beyond exabytes of data.
Variety refers to the type of data being collected ("in what forms"). Sensor data, mobile
technology, and social networking have increased the types of data to include text, images,
videos, location information, and more.
Velocity refers to the speed at which data is generated ("how fast"). Ubiquitous technologies
have generated continues flows of data at rates well beyond anything seen in history to date.
Veracity refers to the data’s integrity ("how reliable"); that is, its accuracy, truthfulness,
precision, and other such factors.
Value refers to the ability to use the data to extract information of value to the organization
("how valuable") that will result in business intelligence and assist decision making.
Following the hierarchy of data information knowledge intelligence (Goes, 2014), we can classify
the five Vs of big data into two subgroups: big data characteristics and big data processing. One can
argue that we do not know data’s validity (veracity) or value until it has been processed. After data has
been processed, we can extract knowledge and intelligence from the information produced. Figure 1
displays the characteristics of big data and its processing.
Figure 1. Characteristics and Processing of Big Data
800 Big Data and Analytics: Issues, Solutions, and ROI
Volume 37 Paper 39
By definition, big data is characterized by the large volumes of various types of data generated at a high
rate. Just as the name implies, big data is a lot of data. We cannot know the data’s veracity or value until it
has been processed. However, big data with low veracity and low value is still big data. The variety of data
ranges from structured to unstructured (Hashem et al., 2015). The continued growth of illstructured data
increases the difficulty of processing and extracting useful information, which results in text mining and
image processing becoming an important new frontier of research (Agarwal & Dhar, 2014). New data
classification and analysis tools such as MapReduce and Hadoop have been developed to process and
manage these large repositories of data that are too large for traditional storage methods (i.e., relational
database) to handle (Ferrera, De Prado, Palacios, Fernandez-Marquez, & Serugendo, 2013). The open
source solution Hadoop has led to a plethora of big data processing tools such as Sawzall, FlumeJava,
Pig, Hive, Jaql, and Cascading that all contain specialized features ranging from SQL-style data
manipulation to Java-based APIs serving a wide range of users and skills (Ferrera et al., 2013).
Big data comes in several types; 1) Web and social media data, including clickstream and interaction data
from social media; 2) machine-to-machine (M2M) data, including readings from sensors, meters, and
other devices; 3) big transaction data, including healthcare claims, telecommunications call detail
rerecords, and utility billing records; 4) biometric data, including fingerprints, genetics, handwriting, retinal
scans, and similar types of data; and 5) human-generated data, including vast quantities of unstructured
and semi-structured data such as call center agents’ notes, voice recordings, emails, paper documents,
surveys, and electronic medical records (Gartner, 2013). Figure 2 displays various systems along with
information used and the data analyzed as the variety of data and complexity increases compared with
the amount of data being collected (i.e., petabytes, exabytes).
Figure 2. Data Storages and Increasing Data Variety and Complexity
2.2 Real-world Applications
There are numerous real-world big data applications in major firms. Among these, several interesting
cases have been illustrated (Davenport & Dyche, 2013): UPS, a global leader in logistics, has been using
big data for over the past two decades. “UPS is no stranger to big data, having begun to capture and track
a variety of package movements and transactions as early as the 1980s… Much of its recently acquired
big data comes from telematics sensors” (Davenport & Dyche, 2013). In particular, UPS’s on-road
Communications of the Association for Information Systems 801
Volume 37 Paper 39
integrated optimization and navigation (ORION) and DIAD (handheld device) are a legacy to big data at
UPS.
Caesars Entertainment, formerly known as Harrah's, established itself as a leader in big data and
analytics. For instance, Caesars has data about its customers from its "total rewards loyalty program, web
clickstreams, and from real-time play in slot machines” (Davenport & Dyche, 2013). Like most other
entertainment industry, Caesars analyze mobile data for spotting service (i.e., use of real-time or near
real-time trend spotting with visualization tools).
Prior to the development of big data tools, many companies stored data that provided little to no
information. It was reported that one telecommunications provider contained up to 10,000 phone
conversations per day with customers but were unable to evaluate them; they could only measure the end
result displaying if a phone plan was changed or not (Negash, 2004). With new analytical tools to evaluate
this qualitative information, managers would not only know the results of a conversation but also the
underlying data the led to the positive result.
3 Interpreting Big Data Benefits
In general, big data refers to all the data/information a business has. Companies, either profit or non-profit,
have faced the fundamental question “how do I make this work?”. They need data/information that is
relevant to ad hoc and long-term decision making so that organizations are built for data-driven decision
making, which is important because decisions made without data-driven answers will likely fail. Big data
has been around for many years and there are two aspects in terms of how the public perceives it. First,
big data refers to the large amount of data assets that information workers store and manage. Second, big
data means the techniques used to process and analyze these data assets in real-time for valuable
information. In the late 1970s, when point of sale (POS) system was introduced in supermarkets and
supercenters, there was a huge growth of consumer data, which essentially led to major paradigm shifts in
manufacturing companies’ marketing and advertising strategies (see Figure 3) (Fulgoni, 2013). POS data
were obtained and analyzed by constantly evolving software applications (e.g., data warehousing) to
generate business value. In a nutshell, big data’s most fundamental advantage lies in its capability to build
the foundation for information workers who identify interesting patterns among data at an extreme scale in
terms of volumes and formats. Doing the same is difficult and not cost-effective for conventional database
technologies (e.g., relational database management systems (RDBMS)).
Figure 3. Paradigm Shift of Marketing due to POS Data Increase (Fulgoni, 2013)
In recent decades, the world economy has become increasingly dependent on knowledge/business
intelligence and well-informed decisions (Kabir & Carayannis, 2013). This trend has stimulated the rapid
development of computing technologies for collecting and analyzing data, which generate the data
tsunami we witness today. Correspondingly, the challenge has transformed from not having enough data
to dealing with too much data. Although the word “big” stresses the volume characteristic, big data
802 Big Data and Analytics: Issues, Solutions, and ROI
Volume 37 Paper 39
requires more than just volume management. In fact, in Mark Beyer’s (2011) three Vs model (volume,
velocity, and variety), big data displays conspicuous advantages over traditional databases (Stephens,
2013; Vriens & Brazell, 2013).
3.1 Perception is Reality: The Choice between Consistency and Speed
“Reality is merely an illusion, albeit a very persistent one.” —Albert Einstein
Customers demand convenient and customized data visualization that helps them to perform daily tasks
more efficiently. To them, everything happens at their fingertips and that is all that matters. The internal
database structure, storage mechanism, and cross validations of query accuracy are invisible to end users
and, therefore, all considerations should give way to data availability and query response speed. This
notion, therefore, introduces a give-and-take scenario: should we sacrifice data consistency or reliability to
obtain a boost in query speeds? For RDBMS developers, this choice is difficult because the atomic
consistency isolation durability (ACID) requirements guarantee reliability in a rigorous way. Plus, the
benefits of ACID have been touted by RDMBS manufacturers for years. For big data engineers, they have
more flexibility when implementing the solutions thanks to the “eventually consistent” technique. The
following list, although incomplete, describes important advantages of big data’s schema-less design:
Not ACID compliant: non-relational has been proven to be valuable in contemporary Web-
based database environments such as MongoDB and HBase. According to Oracle, any
database that isn’t RDBMS upholds schema-less structures and is generally relaxed on ACID
transactions. The schema-free model promises high availability and support for large data sets
in distributed environments.
Horizontally scalable: big data has progressed hand-in-hand with cloud computing. The
Hadoop file system first maps the data and then reduces it based on nodes to distribute tasks
in a large network of servers. A major advantage of this feature, when combined with schema-
less data structures, is superior performance without committing to a significant system
upgrade. Such an advantage also involves large volumes of structured/semi-
structured/unstructured data and, thus, answers the challenge of working with a large variety of
data.
Eventually consistent: eventually consistent, or lazy consensus, is a popular workaround to
resolve the conflict between availability and consistency in big data solutions. With shared data
(e.g., shards) and distributed systems, big data platforms sacrifice consistency rigor to some
degree because they do not prioritize the demand of serialized operations. In return, the speed
of data generation, query performance, and availability are significantly improved. Once a user
request has been recorded, it takes effect in one of the data replicas and, eventually, such
updates will be validated across all nodes. Some big data flavors prefer consistency over
availability and, according to the CAP theorem, no database solution can guarantee
consistency, availability, and partition tolerance together. Therefore, one should evaluate the
business requirements to plan for two and make the best of the third.
4 The Dark Side of Big Data
While big data has received a lot of attention for its potential, we need to face some of its challenges.
Privacy is one major concern. With the amount of data and rate at which it is being created continue to
increase, organizations are able to create comprehensive digital dossiers for each customer with details
beyond anything seen before. These digital dossiers can be used reveal intimate details about an
individual such as sexuality, menstrual cycles, and whether a woman is pregnant or not. It raises ethical
concerns about whether or not companies have a right to mine for such personal information that can be
used for marketing purposes. Additional concerns could be the misuse of data resulting in misleading
truths or the introduction of the digital divide 2.0 (i.e., difference between those who have access to big
data and those who do not).
4.1 Big Data Cases
There are several cases where big data and the information mined from this data have been intrusive and
even harmful to individuals. The retailer Target had a situation where their data mining proved to be an
invasion of privacy into the life of a minor and informed her father of her untimely pregnancy prematurely
Communications of the Association for Information Systems 803
Volume 37 Paper 39
(Hill, 2012). Through data mining customers’ shopping habits and trends, Target was able to identify 25
products that they classified as a “pregnancy prediction” for their shoppers. When the company identified
customers as potentially being pregnant, it flagged the customer and sent coupons based on their
pregnancy score. When a local teen in Minneapolis received coupons for baby products, her father was
outraged and complained to the store. It turned out the teen was pregnant, but she did not intend her
father to find out that way.
In January 2014, a family in Chicago received coupons from OfficeMax identifying their daughter as
deceased from a car crash one year earlier (Merrick, 2014). Through data collections about their
customers and other data sources, OfficeMax knew when their daughter had died and how. The coupons
being sent were blamed on a data error but the collection of that information and its potential use for
marketing raises serious ethical issues. The amount of data collected by these companies is small
compared to the vast amount of user generated data collected through search engines and social media.
The amount of data being collected and stored by companies such as Google, Facebook and Twitter is
beyond anything that has been seen to date. These companies received scrutiny for using public and
private messages in their data mining for marketing purposes (Compeau, Haggerty, & Fraiha, 2011).
Furthermore, Facebook has received complaints about privacy issues for using its technology called
Beacon, which tracks its users’ activity across the Web. Beacon collects user IP addresses from partner
sites to match with IP addresses used on Facebook. Information provided by partner sites can be
matched to a Facebook accounts, increasing the information known about individuals and the ability to
target them for ads (Martin, 2010).
4.2 Digital Dossier
Companies have been collecting information for many years about their customers, which they use to
develop dossiers that they can use for marketing. The use of digital technology has improved companies’
ability to understand their customers through developing digital dossiers. With the advent of social media
and the amount of information social networking users are willing to provide, these digital dossiers have
reached a level of detail never before seen in history. While this information provides a powerful tool for
marketers to understand their customers and relate to them on a micro-level, they need to consider ethical
and privacy issues.
As the examples in Section 4.1 show, companies can effect negative outcomes by collecting big data.
With Facebook’s purchase of WhatsApp, the amount of information contained in the digital dossier of
Facebook’s user base increased exponentially. Facebook has made a strong attempt to collect their
users’ phone numbers in ways ranging from recommended security features and password recovery
systems to their development of Facebook Messenger, which requires a phone number to register.
WhatsApp messenger requires a valid phone number to register and is used to exchange private
messages among its users. After being purchased by Facebook for USD$19 billion, Facebook obtained
the rights to the company and all its information. By matching phone numbers from WhatsApp to phone
numbers it had obtained itself, Facebook could link the accounts of both systems to create a more
complete window into a person’s life. But the privacy threats extend beyond textual messages posted both
publically and privately.
If a picture is worth a thousand words, then one can imagine how many words can be generated by social
networking users who have posted thousands of pictures to their accounts. With new technology such as
facial recognition software, images shared can tell more of a story than the captions posted to describe
them. New technology described by Facebook called DeepFace accurately has identified matching
individuals 97.25 percent of the time (O’Toole, 2014). This type of technology could match individuals from
the background of pictures taken by others they don’t know. For instance, if person A was taking a
vacation and appeared in the background of a picture taken by person B, then facial recognition software
could identify person A and link them to a location identifying where they were traveling. Google chose not
to include facial recognition software in its Google Glasses due to privacy concerns raised by privacy
advocates such as the American Civil Liberties Union (Bawab, 2014). This technology could allow people
wearing Google Glasses to identify others while walking down the street and conduct Google searches for
more information on them.
804 Big Data and Analytics: Issues, Solutions, and ROI
Volume 37 Paper 39
4.3 Ethics and Privacy
Morris (2002) has reported that the top five sources for big data include social networks, social
influencers, activity-generated data, software-as-a-service (SaaS) and cloud applications, and public data.
Many individuals use these services as part of their daily routine without intending to provide their
personal data for organizations to exploit. Social networking in particular contains information viewed by
the user as both private and public. Wall posts are viewed as public information, while personal messages
sent through these systems are deemed private and personal. However, both private and public data are
used in big data for business intelligence and analytics to increase marketing capabilities, which many
individuals view as invading their privacy. Without formal laws dictating how big data can be used, we
must rely on ethical considerations to find a solution. It would be unethical for a store manager at a
grocery store to listen to their customers’ private conversations to improve marketing capabilities. Despite
customers being in the grocery store, their conversation is still considered private and do not want to be
eavesdropped on.
We can apply this same concept to social networking. There is an intended party that an individual wishes
to receive private messages sent across the network. The intended party encompasses other users on the
social network and not the social network itself. To use these private message for marketing and further
developing digital dossiers is unethical and an invasion of privacy. Due to the recency of which the
technology exists, privacy laws do not exist to protect social networking users. We must depend on the
owners of big data to be responsible with the data they collect and how they use it.
4.4 Digital Divide 2.0
Since the commercialization of the Internet, there has been a digital divide between those who have
access to Internet technology and information and those who do not. Great strides have been taken to
bridge this gap and provide access to even the most rural areas. While there are areas of the world that
still lack the ability to access information through the Internet, businesses worldwide have been able to
overcome this gap and compete on equal ground. With the growth of big data due to social networking,
search engines, Internet tracking, and other data sources, a new digital divide in information is being
formed. Small to mid-sized companies do not have the luxury of collecting data from a half-billion users
who are provided broad ranges of information on various topics. They also do not have the monetary
assets needed to purchase this data to mine and develop their own business intelligence. Large
companies are buying out the smaller ones (e.g., Facebook buying WhatsApp) and building industries that
make it difficult for others to compete, which has developed a divide between those who have access to
big data and those who do not. As this trend continues, a natural monopoly is being formed where smaller
companies cannot hope to compete. To gain access to the data, one has to be willing to provide
information that will contribute to the continued growth of big data by these companies.
5 Maximizing Return for Big Data Projects
5.1 Improving ROI for Big Data Projects
As more companies look to accelerate their growth and better engage with customers, big data has taken
off as a set of common principals for leveraging data to better manage an organization. As with any new
trend, the early adopters were very informal in how they began and executed projects: they knew that
value was there but focused on execution rather than measurement. As big data has become more
mainstream, organizations have looked for ways to measure the impact of investment in big data projects.
Communications of the Association for Information Systems 805
Volume 37 Paper 39
Figure 4. Big Data Return On Investment (Jablonski, 2014)
Senior leadership has said that greater than 70 percent (Bertolucci, 2013) of big data projects fail to live
up to the initial hype because of unrealistic measurements or measurements that are not properly aligned
with the expectations of senior leadership in an organization. Big data projects must both align with the
organization’s needs and have measurements that are unique to the rapidly changing technology and
customer landscape. Some organizations have seen returns as high as ten times their investment (Tata
Consultancy Services, 2013) for big data projects, so it is possible to successfully execute on these
complex engagements.
Figure 5. Big Data Project Fundamentals (Jablonski, 2014)
As organizations look to execute big data projects, they should consider four key areas of planning:
Skill: skills are a key consideration for planning any big data project. Big data projects often
introduce new technologies and methodologies into an organization, and any project should
include resources to ensure proper training is provided for staff to minimize the learning curve
and ensure maximum understanding of new technology.
Costs
Returns
Costs
• Acquisition (Hardware & Software)
• Support
• Skills (Training & Consultants)
• Facilities & Overhead
• Royalties
Returns
• Speed of business execution
• Operational improvements
• Increase business
• Decrease costs
• Product design cycle
Big Data
Technology
Skills
Use
Profile
Measurement
Technology
• Flexibility, Scalability,
Manageability
• Compliments existing, not
replacement platforms
• Evolves as the business
evolves
Use Profile
• Who is the primary user?
• What do they want to
know?
• What is their skill set?
• Where and how do they
work most effectively?
• Start small, single use and
single user group
Skills
• Both deployment & usage
• Combination of training
and fresh blood in the
organization
Measurement
• For all aspects of the
project
• Derived from LoB metrics
806 Big Data and Analytics: Issues, Solutions, and ROI
Volume 37 Paper 39
Measurement: all big data projects should have clear project measurements defined that are
aligned with the needs and daily metrics of the sponsoring line-of-business. All senior
leadership in specific lines-of-business in an organization leverage key performance indicators
(KPIs) to measure the performance of the organization; the most successful big data projects
will align with those KPIs and work to improve them. Since it is very hard to quantify the
benefits of a big data project, this measurement area is critically important to consider.
Technology: big data projects will combine the deployment of new technologies and integration
with existing technology and work flows. Big data projects should consider all necessary
requirements and plan for phased deployments of new technologies to both gather experience
over time and to ensure upfront plans and designs are feasible and can be executed. Ensure
the focus establishes clear lines of sight to technology requirements (Marchand & Peppard,
2013).
Use profile: big data projects affect a variety of staff in an organization including system
administrators, system architects, program managers, and business analysts. Each staff
member has a different skill set and job description that should be planned for during a big
data project. All big data projects should inventory the various types of users that will interact
with the platform and ensure the project accounts for the various needs around interfaces,
presentation, and usability.
Organizations also need to understand how each category affects designs and planning for other
categories:
Technology influences training: the training plan will be heavily influenced by the technology
strategy for a big data project. Any new technologies will impact training that will need to be
provided to staff, both operations and users.
Use profile influences technology: big data projects vary on usage; some are deployed with
developer-centric users in mind, while others are deployed with leadership users as the
primary design target. This use profile influences the technologies that will be used to present
and analyze the results of any data in the big data environment.
Measurement impacts technology: many KPIs used by business are components of time and
cost. These KPIs influence technology choices around the company’s environment’s sizing, the
scalability of the technologies involved, and integration with existing business systems.
Skills impact use profile: the relative skill set of users in an organization will impact the use
profile because of how users will use new technology provided to them and the preferences
they will have when adopting new technologies.
Properly executing big data projects requires a balance between skills, measurement, technology, and
use profile to ensure maximum return. Insufficiently considering or investing in any one of them will
negatively affect the final platform and staff’s ability to leverage the big data investment.
6 Conclusion
In this paper, we present several distinct but interesting perspectives on current issues of big data and
analytics and, specifically, on maximizing return from big data projects. As mentioned earlier, big data and
analytics are an important part of today's business (i.e., analyzing customer information, categorizing
customer responses, and tracking trends and patterns). Further research should explore new and growing
subsets of big data and analytics, such as speech and call analytics, biometrics data analytics, and sensor
data analytics.
While the term big data is relatively new, the concept of big data has existed since the late 1800s when
the U.S. Government conducted the country’s first census. As the capabilities of processing data
continues to increase, the amounts of data being collected also increases. The term big data gained
significant traction with the explosion of social media and user-generated data. The speed at which data is
generated has exceeded the capabilities of current technology to process it. Text mining has created
significant strides in our ability to process and understand big data, but there are still mountains of data in
the form of images and video that can be explored as technology continues to evolve. However, many of
this user-generated data was not intended for corporate use and raises ethical issues about whether or
not it should be processed and mined for monetary purposes.
Communications of the Association for Information Systems 807
Volume 37 Paper 39
Some have argued that Facebook owns the messages sent from one user to another through the social
networking platform because the company owns the platform on which those conversations occur.
However, conversations taken place in a retail store do not belong to the organization despite two
customers having a discussion while in the store. This raises the ethical debate as to who owns the data,
how it should be used, and how it should be secured. However, many technologies are too new to answer
many questions. We need further research to explore ethical issues related to big data and how the levels
of detail being captured should be used. Furthermore, data scientists who have expertise in quantitative
skills (i.e., how to use big data tools for managing huge sets of data like Hadoop and analytics) and
effective at communicating their findings in a manner that functional managers can understand are rare
(Tata Consultancy Services, 2013). Since some industries have invested heavily in IT over several
decades, they have different levels of data intensity.
In this paper, we discuss trends and raise awareness of these issues to stimulate further research into big
data to address big data’s capabilities, privacy issues and ethical concerns, and security issues. This
opens to door to a plethora of research possibilities ranging from new analytical techniques to ethics and
privacy. From the organizational perspective, we need to improve the ability to generate information from
images and videos beyond the meta-data provided by the content developer. Images can contain a lot of
information relating to who, what, when, where, why, and how things took place. User comments provide
feedback to the events taking place in the image that provide additional information. From the user
perspective, we should evaluate ethics and privacy concerns when using this information for
organizational gains. Most users provide content to share with their friends and other social contacts and
not organizations to evaluate and determine the best way to manipulate them into purchasing products or
services.
Another area of research could revolve around mashup research, which involves combining data from
multiple sources to research phenomenon that we were previously unable to evaluate. Many data sources
contain personally identifiable information (PII) that one can use to group individuals based on various
factors such demographic and geographic data. As an example, a researcher could combine social
networking data from Twitter and Facebook with Google Maps to conduct textual analysis across multiple
networks to determine how perceptions in various regions may differ. Organizations continue collected
large amounts of data to answer questions that have yet to be asked or test hypotheses that have yet to
be developed (Agarwal & Dhar, 2014). This is where the creativity of researchers will come in to start
discovering the questions and hypotheses that can take advantage of these data sets as we continue to
advance business and the IS field. While we identify some areas that are in need of immediate attention,
the opportunities for continued research will grow as expeditiously as big data itself.
808 Big Data and Analytics: Issues, Solutions, and ROI
Volume 37 Paper 39
References
Agarwal, R., & Dhar, V. (2014). Big data, data science, and analytics: The opportunity and challenge for IS
research. Information Systems Research, 25(3), 443-448.
Bawab, H. (2014). Privacy concerns raised around facial recognition technology. LinkedIn. Retrieved from
http://www.linkedin.com/today/post/article/20140407061644-14091619-privacy-concerns-raised-
around-facial-recognition-technology
Bertolucci, J. (2013). Big data ROI still though to measure. InformationWeek. Retrieved from
http://www.informationweek.com/big-data/big-data-analytics/big-data-roi-still-tough-to-measure/d/d-
id/1110150?
Beyer, M. (2011). Gartner says solving “big data” challenge involves more than just managing volumes of
data. Gartner. Retrieved from http://www.gartner.com/newsroom/id/1731916
Chen, H., Chiang, R., & Storey, V. (2012). Business Intelligence and analytics: From big data to big
impact. MIS Quarterly, 36(4), 1165-1188.
Compeau, D., Haggerty, N., & Fraiha, S. (2011). Privacy issues and monetizing Twitter. Ivey Publishing.
Davenport, T., & Dyche, J. (2013). Big data in big companies. International Institute for Analytics.
Retrieved from http://www.sas.com/content/dam/SAS/en_us/doc/whitepaper2/bigdata-
bigcompanies-106461.pdf
Ferrera, P., De Prado, I., Palacios, E., Fernandez-Marquez, J.-L., & Serugendo, G. D. M. (2013). Tuple
MapReduce and Pangool: An associated implementation. Knowledge and Information Systems.
Fulgoni, G. (2013). Big data: Friend or foe of digital advertising? Journal of Advertising Research, 53(4),
372-376.
Gartner. (2013). Drive value from big data through six emerging best practices. Retrieved from
https://www.gartner.com/doc/2600415?ref=SiteSearch&sthkw=2013%20big%20data%20volume%2
0velocity%20variety&fnl=search&srcId=1-3478922254
Gillon, K., Aral, S., Lin, C. Y., Mithas, S., & Zozulia, M. (2014). Business analytics: Radical shift or
incremental change? Communications of the Association for Information Systems, 34(13), 287-296.
Goes, P. (2014). Editor's comments: Big data and IS research. MIS Quarterly, 38(3), iii-viii.
Hashem, I., Yaqoob, I., Anuar, N. B., Mokhtar, S., Gani, A., & Khan, S. U. (2015). The rise of "big data" on
cloud computing: Review and open research issues. Information Systems, 47, 98-115.
Hill, K. (2012). How Target figured out a teen girl was pregnant before her father did. Forbes. Retrieved
from http://www.forbes.com/sites/kashmirhill/2012/02/16/how-target-figured-out-a-teen-girl-was-
pregnant-before-her-father-did/
Jablonski, J. (2014). Maximizing return for big data projects. Presentation presented at 2014 AMCIS Panel
Session.
Kabir, N., & Carayannis, E. (2013). Big data, tacit knowledge and organizational competitiveness. In
Proceedings of the International Conference on Intellectual Capital, Knowledge Management &
Organizational Learning (pp. 220-227).
Lycett, M. (2013). “Datafication”: Making sense of (big) data in a complex world. European Journal of
Information Systems, 22, 381-386.
Marchand, D., & Peppard, J. (2013). Why IT fumbles analytics. Harvard Business Review, 91, 104-112.
Martin, K. (2010). Facebook (A): Beacon and privacy. Institute for Corporate Ethics. Retrieved from
http://www.corporate-ethics.org/pdf/Facebook%20_A_business_ethics-case_bri-1006a.pdf
McAfee, A., & Brynjolfsson, E. (2012). Big data: The management revolution. Harvard Business Review,
90, 60-68.
Merrick, A. (2014). A death in the database. The New Yorker. Retrieved from
http://www.newyorker.com/business/currency/a-death-in-the-database
Communications of the Association for Information Systems 809
Volume 37 Paper 39
Morris, J. (2012). Top 10 categories for big data sources and mining technologies. ZDNet. Retrieved from
http://www.zdnet.com/top-10-categories-for-big-data-sources-and-mining-technologies-
7000000926/
Negash, S. (2004). Business intelligence. Communications of the Association for Information Systems, 13,
177-195.
O'Toole, J. (2014). Facebook’s new face recognition knows you from the side. CNN Money. Retrieved
from http://money.cnn.com/2014/04/04/technology/innovation/facebook-facial-recognition/
Stephens, C. (2013). The power of big data and high performance analytics. Finweek, 4-5.
Tata Consultancy Services. (2013). The emerging big returns on big data.
Vriens, M., & Brazell, F. (2013). The competitive advantage. Marketing Insights, 25(3), 32-38.
Wixom, T., Ariyachandra, T., Douglas, D., Goul, M., & Gupta, B. (2014). The current state of business
intelligence in academia: The arrival of big data. Communications of the Association for Information
Systems, 34, 1-13.
810 Big Data and Analytics: Issues, Solutions, and ROI
Volume 37 Paper 39
About the Authors
J. P. Shim is a faculty of Computer Information Systems and Executive Director of Korean-American
Business Center at Robinson College of Business at Georgia State University and professor emeritus at
Mississippi State University (MSU). Before joining at GSU in 2011, he was faculty of BIS and Larry and
Tonya Favreau Notable Scholar at MSU. During the past 27 years at MSU, he was a seventeen-time
recipient of outstanding faculty awards, including 1994 John Grisham Faculty Excellence awardee, 2006
Powe Research Excellence award winner, and winner of 2011 MSU Diversity award. He has published
several books and seventy journal papers. He serves on Wireless Telecommunication Symposium as
Program chair and served on 2013 AMCIS Program co-chair. He has received awards, grants, and
distinctions, including NSF, Microsoft, U.S. Small Business Administration, Japan Foundation, and Korea
Foundation. He has been interviewed by the media (CBS TV, AP, AJC, Global Atlanta) and worked as a
consultant for Booz Allen, U.S. EPA, Rehabilitation Associates, Inc. (in Wisconsin), CYR International, and
Kia Motors. His current research interests are cross-cultural study of BYOD, big data, speech analytics,
and wireless telecommunications.
Aaron M. French is an Assistant Professor of Management Information Systems in the College of
Business at the University of New Mexico in Albuquerque, New Mexico. He received his PhD in Business
Information Systems at Mississippi State University. He is a three-time recipient of the Outstanding
Teacher of the Year Award. His research has been published in the Journal of Information Technology,
Behaviour & Information Technology, Journal of Internet Banking and Commerce, and The Journal of
Internet Electronic Commerce Research. His research interests include social networking, e-commerce,
cross-cultural studies, and technology acceptance.
Chengqi (John) Guo is an Assistant Professor and Madison Research Fellow of Computer Information
Systems and Management Science in the College of Business at James Madison University. He received
his PhD in Business Information Systems from Mississippi State University. He received a Masters of
Operations Management and Information Systems from Northern Illinois University and a BS in
International Marketing from Guangdong University of Foreign Studies, Guangzhou, China. He is a senior
consultant and Director of International Business Development at JDArray Co. Ltd, Beijing China. His
research interests are information systems security, social media, mobile computing, technical innovation,
human-computer interaction (adoption, trust, privacy, and communication), innovative technology in
education, and cross-cultural studies.
Joey Jablonski is currently a VP/Principal Architect, leading the Big Data Practice at Cloud Technology
Partners, a professional services firm focused on strategy, architecture and development of cloud
applications and platforms. Before joining Cloud Technology Partners, he was an Enterprise Technologist
at Dell working as a member of the Office of the CTO. He focused on the architecture and strategy for the
deployment of complex analytic technologies and big data platforms. His technical interests include big
data, high performance computing, low-latency networking technologies and information security. He has
previously held technical and organizational leadership roles at DataDirect Networks, HP, and Sun
Microsystems.
Copyright © 2015 by the Association for Information Systems. Permission to make digital or hard copies of
all or part of this work for personal or classroom use is granted without fee provided that copies are not
made or distributed for profit or commercial advantage and that copies bear this notice and full citation on
the first page. Copyright for components of this work owned by others than the Association for Information
Systems must be honored. Abstracting with credit is permitted. To copy otherwise, to republish, to post on
servers, or to redistribute to lists requires prior specific permission and/or fee. Request permission to
publish from: AIS Administrative Office, P.O. Box 2712 Atlanta, GA, 30301-2712 Attn: Reprints or via e-
mail from publications@aisnet.org.
