Journal of Business Logistics, 2015, 36(1): 120132

Council of Supply Chain Management Professionals

doi: 10.1111/jbl.12082

Data Science, Predictive Analytics, and Big Data in Supply Chain

Management: Current State and Future Potential
Tobias Schoenherr and Cheri Speier-Pero
Michigan State University

hile data science, predictive analytics, and big data have been frequently used buzzwords, rigorous academic investigations into these
areas are just emerging. In this forward thinking article, we discuss the results of a recent large-scale survey on these topics among supply chain management (SCM) professionals, complemented with our experiences in developing, implementing, and administering one of the
rst masters degree programs in predictive analytics. As such, we effectively provide an assessment of the current state of the eld via a largescale survey, and offer insight into its future potential via the discussion of how a research university is training next-generation data scientists.
Specically, we report on the current use of predictive analytics in SCM and the underlying motivations, as well as perceived benets and barriers. In addition, we highlight skills desired for successful data scientists, and provide illustrations of how predictive analytics can be implemented in the curriculum. Relying on one of the largest data sets of predictive analytics users in SCM collected to date and our experiences
with one of the rst masters degree programs in predictive analytics, it is our intent to provide a timely assessment of the eld, illustrate its
future potential, and motivate additional research and pedagogical advancements in this domain.
Keywords: data science; predictive analytics; big data; data scientist; supply chain management; education; curriculum development

INTRODUCTION
A topic that is on the minds of many supply chain management
(SCM) professionals is how to deal with massive amounts of
data, and how to leverage and apply predictive analytics. This
challenge is a direct result of the ease with which data have been
able to be collected via modern information technology, generating unprecedented volume, variety, and velocity of data.
Heralded to revolutionize how SCM is conducted (Waller and
Fawcett 2013b), predictive analytics has the potential for signicant above-average returns (McAfee et al. 2012). Within the context of SCM, predictive analytics can be dened as using both
quantitative and qualitative methods to improve supply chain
design and competitiveness (Waller and Fawcett 2013b, 80).
Predictive analytics is positioned within the overall domain of
data science, which refers to the application of quantitative and
qualitative methods from a variety of disciplines in combination
with SCM theory to solve relevant SCM problems and predict
outcomes, taking into account data quality and availability
issues (Waller and Fawcett 2013b, 79).
Despite the importance and relevance of SCM predictive analytics, there is a dearth of literature on the topic and many questions (Waller and Fawcett 2013b, 77). While articles in
practitioner outlets and consultancy reports are becoming more
prevalent, their content is mostly repetitive, and rigorous scientic investigations into the topic have been absent. In addition,
while companies are experimenting with big data analytics, for
the majority, its application has been elusive or has only
scratched the surface of what may be possible. The need for
well-trained data scientists is thus imminent (Dwoskin 2014).
Further research into predictive analytics is also needed, both

from an academic and a practical perspective. It is the objective

of this forward thinking article to contribute to this domain, and
provide an assessment of its current state and future potential.
We specically aim to answer the following questions: (1) What
is the current extent of use of SCM predictive analytics, and
what are the underlying motivations? (2) What are benets and
barriers to SCM predictive analytics? and (3) How can we train
our next-generation data scientists?
We rely on two primary data sources to develop our insights.
The rst consists of a large-scale survey of 531 SCM professionals (administered in September 2013), inquiring about their current use, associated motivations, benets and barriers to data
science, predictive analytics, and big data, as well as desired skill
sets. With this data set, we provide an assessment of the
domains current state. The second data source is comprised of
insight derived from developing, implementing, and administering one of the rst predictive analytics masters degree programs at a major U.S. research university. Specic data sources
consulted for the development of this article include workshops,
company interviews, and student feedback (both pre- and postgraduation). With these data, we are able to assess the domains
future potential, offering insight into how next-generation data
scientists are trained. This exposition should be especially useful
to those institutions that are contemplating the development of
their own degree programs in predictive analytics. Our discussion
of the development and modications over the last three years,
including lessons learned, should offer further invaluable guidance for these institutions; it essentially represents a blueprint or
template for such a program.

SCM PREDICTIVE ANALYTICS

Corresponding author:
Tobias Schoenherr, Supply Chain Management, Michigan State
University, 632 Bogue St., East Lansing, MI 48824, USA;
E-mail: [email protected]

Extent of use and motivation

To obtain insight into the current state of SCM predictive analytics adoption, we conducted a large-scale survey among SCM

Data Science, Predictive Analytics, and Big Data in SCM

professionals. A total of 531 individuals responded to the online

survey. As a requirement for participation, respondents had to
be in a function commonly associated with SCM (logistics/transportation, operations management, SCM, purchasing/procurement/sourcing, engineering, or research and development). We
asked respondents to indicate their current use of analytics in
terms of whether (1) they currently do not use analytics but plan
to do so in the future, (2) they use analytics to some extent, (3)
they use analytics to a great extent, (4) they currently do not
use analytics and also do not have any plans to use it in the
future, or (5) they are not familiar with analytics. Table 1 presents the results.
Most respondents were in logistics/transportation (45.2%), followed by operations management (20.9%), and SCM (17.9%).
About half of the sample indicated using predictive analytics to
some (27.7%) or to a great extent (13.2%), or plans to use it in
the future (8.7%). The other half of the sample was either not
familiar with predictive analytics (28.4%) or they did not have
any plans to use it in the future (22.0%).
The results offer intriguing insight into the current and
planned adoption of SCM predictive analytics. Encouraging is
the fact that more than 40% of the respondents actively use analytics. However, while some of the remaining sample plan to use
the approach in the future (8.7%), many respondents do not plan
to go this route (22.0%). It was also surprising that about onethird of the respondents were not familiar with analytics (28.4%),
especially in light of a survey conducted two years prior, which
reported only 7% not being familiar with data analytics (Russom
2011). This illustrates the signicant educational work that still

121

has to be conducted in introducing these individuals to predictive

analytics potential.
As our focus was on the use of SCM predictive analytics, we
disqualied all respondents who were not familiar with analytics
or who did not have any plans for its future use. We further
removed all data with missing values. This yielded a total of 212
complete and useable records, which are used for the subsequent
analyses.
One of our objectives was to identify the motivations for using
predictive analytics. To tap into this domain, respondents were
provided with a list of potential motivators, to which they were
asked to indicate the extent that these motivate/inuence their
use of such approaches for SCM. The scale was anchored at
not at all (value = 1) and to a great extent (value = 7). The
list of motivators was derived from an extensive literature review
of primarily practitioner articles on this topic and interaction with
SCM professionals.
Table 2 presents the results, also split by usage group (there
were no signicant differences when contrasting the three usage
groups based on the respondents function). Looking at the overall mean, encouragement by senior leadership and the respondents conviction about the value of SCM predictive analytics
were the most prominent motivators, followed by internal colleagues, competitors, and customers using analytics. Coverage in
the popular press did not seem to inuence the respondents
motivation.
Intriguing are the results based on the comparison by usage
group, which was accomplished by means of analyses of variance (ANOVAs). Most signicant differences across the three

Table 1: Cross-tabulation of function and usage group

Usage group

Function
Logistics/transportation (count)
% Within function
% Within usage group
Operations management (count)
% Within function
% Within usage group
Supply chain management (count)
% Within function
% Within usage group
Purchasing/procurement/sourcing (count)
% Within function
% Within usage group
Engineering (count)
% Within function
% Within usage group
Research & development (count)
% Within function
% Within usage group
Total (count)
% of total

No current use,
but plans for
the future

To some
extent

To a great
extent

No current
use, no plans
for the future

Not familiar
with data
analytics

Total

19
7.9
41.3
10
9.0
21.7
8
8.4
17.4
6
18.2
13.0
2
7.4
4.3
1
4.0
2.2
46
8.7

66
27.5
44.9
27
24.3
18.4
29
30.5
19.7
9
27.3
6.1
5
18.5
3.4
11
44.0
7.5
147
27.7

23
9.6
32.9
9
8.1
12.9
26
27.4
37.1
6
18.2
8.6
3
11.1
4.3
3
12.0
4.3
70
13.2

55
22.9
47.0
32
28.8
27.4
13
13.7
11.1
2
6.1
1.7
10
37.0
8.5
5
20.0
4.3
117
22.0

77
32.1
51.0
33
29.7
21.9
19
20.0
12.6
10
30.3
6.6
7
25.9
4.6
5
20.0
3.3
151
28.4

240
100.0
45.2
111
100.0
20.9
95
100.0
17.9
33
100.0
6.2
27
100.0
5.1
25
100.0
4.7
531
100.0

122

T. Schoenherr and C. Speier-Pero

Table 2: Motivation to use SCM predictive analytics

Usage group

Motivation
My conviction
Internal colleagues
Customers
Competitors
Encouragement by
senior leadership
Suppliers
Press coverage

Overall
mean
4.73
4.47
4.11
4.25
4.78

(1.36)
(1.48)
(1.73)
(1.77)
(1.57)

No current use,
but plans for
the future (1)
4.17(3)
3.94(3)
4.11
4.19
4.42

4.05 (1.70)
3.06 (1.65)

To some
extent (2)
4.55(3)
4.44
3.84(3)
4.03(3)
4.72

(1.36)
(1.55)
(1.69)
(1.58)
(1.52)

4.03 (1.50)
3.25 (1.56)

ANOVA
F-value
and sign.

To a great
extent (3)
5.40(1,2)
4.83(1)
4.62(2)
4.72(2)
5.12

(1.25)
(1.39)
(1.70)
(1.80)
(1.58)

3.97 (1.75)
3.00 (1.67)

(1.33)
(1.54)
(1.72)
(1.75)
(1.54)

12.62**
4.20*
4.06*
3.11*
2.47*

4.22 (1.73)
3.07 (1.67)

0.43
0.32

Notes: This table shows means and standard errors (in parentheses). The numbers in the superscripted parentheses indicate the group score number from
which the group score is signicantly different. Due to unequal sample sizes, post hoc pairwise comparison tests were conducted utilizing the Hochberg
test statistic. **p < .001, *p < .05.

groups existed on the individuals conviction serving as a motivator. The valuable implication derived from this nding is that
an SCM professionals conviction about the value of SCM predictive analytics is one of the primary drivers for early adoption.
The result may also be indicative of users realizing the value of
the approaches once they are actively being utilized, further
increasing the individuals conviction. To a lesser degree, internal colleagues, customers, and competitors serve as distinguishing characteristics being related to more frequent use of

predictive analytics. Two further statements, while serving as a

motivator for all, did not differentiate across the three groups
(encouragement by senior leadership and suppliers).
Benets and barriers
A further primary objective of this research was to identify benets of and barriers to SCM predictive analytics. Similar as above,
respondents were provided with a list of potential benets and

Table 3: Benets of the use SCM predictive analytics

Usage group

Benets

Overall
mean

Better/more informed decision making

5.57 (1.23)
Increased visibility
5.22 (1.36)
Better management of supply chain risk
5.13 (1.33)
Improvement in supply chain costs
5.26 (1.44)
Enhanced bargaining position in negotiations
4.62 (1.60)
with suppliers
Improvement in supply chain efciencies
5.32 (1.32)
Enhanced demand planning capabilities
5.28 (1.29)
Enhanced sales and operations planning (S&OP) capabilities 4.97 (1.33)
Enhanced bargaining position in negotiations
4.56 (1.59)
with customers
Ability to respond faster to changing
5.11 (1.30)
environments
Real-time decision-making capability
5.12 (1.31)
Greater power in relationships with suppliers
4.73 (1.52)
Greater power in relationships with customers
4.83 (1.55)

No current use,
but plans for
the future (1)

To some
extent (2)

To a great
extent (3)

ANOVA
F-value
and sign.

5.36(3)
5.14
4.97
5.19
4.50

(1.44)
(1.29)
(1.21)
(1.28)
(1.63)

5.40(3)
5.01(3)
4.94(3)
5.03(3)
4.38(3)

(1.18) 6.03(1,2)
(1.42) 5.68(2)
(1.38) 5.58(2)
(1.51) 5.73(2)
(1.60) 5.15(2)

(1.09)
(1.17)
(1.21)
(1.26)
(1.46)

6.21**
5.17**
5.11**
4.91**
4.90**

5.28
5.06
4.69
4.31

(1.23)
(1.33)
(1.26)
(1.58)

5.12(3)
5.17
4.88
4.41

(1.34)
(1.29)
(1.31)
(1.60)

5.72(2)
5.63
5.32
4.98

(1.28)
(1.21)
(1.37)
(1.52)

4.15*
3.29*
3.14*
3.12*

4.86 (1.48)

5.03 (1.27)

5.43 (1.18)

2.82*

5.06 (1.29)
4.72 (1.52)
4.75 (1.59)

4.97 (1.34)
4.58 (1.54)
4.72 (1.58)

5.45 (1.23)
5.03 (1.45)
5.10 (1.46)

2.80*
1.80
1.28

Notes: This table shows means and standard errors (in parentheses). The numbers in the superscripted parentheses indicate the group score number from
which the group score is signicantly different. Due to unequal sample sizes, post hoc pairwise comparison tests were conducted utilizing the Hochberg
test statistic. **p < .01, *p < .05.

Data Science, Predictive Analytics, and Big Data in SCM

123

barriers, as identied by the authors via primarily practitioner

articles and interaction with SCM professionals. The same 7point scale as above was used, and similar group comparisons
based on ANOVAs were conducted.
The results for the benets are presented in Table 3, which
also includes a split by usage group (ANOVAs based on the
function again did not yield signicant differences between the
groups). Benets were especially seen in the form of more
informed decision-making capabilities, ability to improve supply
chain efciencies, enhanced demand planning capabilities,
improvement in supply chain costs, and increased visibility.
Signicant differences existed between the three usage groups
for at least six of the perceived benets. In general, the super
users (i.e., those using analytics to a great extent) were much
more convinced about the benets pertaining to better decision
making, increased visibility, better risk management, improvement in SCM costs, the creation of an enhanced bargaining
position with suppliers, and improvement in supply chain efciencies. The implication of these ndings is that with increased
use of predictive analytics, the magnitudes with which these
benets are perceived increases. These benets, which are anticipated by all three groups, thus seem to materialize in even
greater magnitude once predictive analytics are used to some or
to a great extent. This result may be indicative of expectations
pertaining to these benets being surpassed with the actual use
of predictive analytics. For the remaining benets statements, in

general, the extent to which the benets were perceived

increased nominally when moving from the group that has
plans for the future use, to the group that uses predictive analytics to some extent, and to the group that uses predictive analytics to a great extent. However, ANOVA F-statistics indicated
that these differences were not statistically signicant. All postulated benets were however perceived as being favorable, as
indicated by the mean values being above the midpoint of the
scale.
Table 4 offers the results for the barriers to SCM predictive
analytics, again also split by usage group (ANOVAs based on
the function again did not yield signicant differences between
the groups). Primary barriers as perceived by our respondents
include employees being inexperienced (and the need for training), time constraints, lack of integration with current systems,
the costs of currently available solutions, change management
issues, lack of appropriate predictive analytics solutions for
SCM, as well as the perception of SCM predictive analytics
being overwhelming and difcult to manage.
Signicant differences between the three usage groups existed
only for two barriers identied, that is, lack of data, and the
inability to identify data most suitable for predictive analytics.
Both of these barriers were perceived to be more severe by the
group that plans to use analytics in the future, compared to the
group that already uses analytics to a great extent. Efforts thus
need to be undertaken by the former group to identify appropri-

Table 4: Barriers to the use SCM predictive analytics

Usage group

Barrier
Lack of data
Inability to identify most suitable data
Security concerns
Lack of upper management support
Unclear business case or value
Privacy/condentiality issues
Lack of policies and governance structure
Inability to make sense of available data
No need/not necessary/no benet
Overwhelming, difcult to manage
Cost of currently available solutions
Lack of integration with current systems
Employees are inexperienced (need to train)
Change management issues (resistance to
change)
Lack of appropriate solutions for SCM
Current applications unable to meet business
needs
Time constraints

Overall
mean
3.83
3.99
3.84
3.83
3.83
3.80
3.91
3.95
3.30
4.16
4.48
4.61
4.92
4.44

(1.50)
(1.48)
(1.71)
(1.74)
(1.44)
(1.72)
(1.57)
(1.54)
(1.64)
(1.46)
(1.48)
(1.51)
(1.47)
(1.64)

No current use,
but plans for
the future (1)
4.42(3)
4.50(3)
4.36
4.39
4.19
4.28
4.14
4.39
3.67
4.56
4.86
4.92
4.92
4.56

(1.32)
(1.56)
(1.69)
(1.40)
(1.04)
(1.60)
(1.31)
(1.48)
(1.45)
(1.46)
(1.44)
(1.50)
(1.61)
(1.61)

To some
extent (2)
3.79
4.00
3.86
3.83
3.89
3.84
4.03
3.96
3.34
4.15
4.41
4.62
4.83
4.32

(1.35)
(1.41)
(1.69)
(1.79)
(1.41)
(1.70)
(1.59)
(1.44)
(1.60)
(1.33)
(1.38)
(1.45)
(1.47)
(1.67)

To a great
extent (3)
3.55(1)
3.67(1)
3.50
3.52
3.52
3.45
3.53
3.67
2.98
3.95
4.38
4.40
5.12
4.60

(1.79)
(1.50)
(1.71)
(1.78)
(1.65)
(1.77)
(1.65)
(1.71)
(1.80)
(1.66)
(1.68)
(1.60)
(1.38)
(1.59)

ANOVA
F-value
and sign.
3.93*
3.66*
2.91*
2.87*
2.77*
2.71*
2.54*
2.51*
2.07
1.97
1.44
1.33
0.77
0.69

4.33 (1.41)
3.96 (1.49)

4.56 (1.46)
4.14 (1.40)

4.30 (1.33)
3.87 (1.44)

4.25 (1.54)
4.03 (1.64)

0.58
0.54

4.63 (1.37)

4.56 (1.56)

4.71 (1.25)

4.52 (1.49)

0.44

Notes: This table shows means and standard errors (in parentheses). The numbers in the superscripted parentheses indicate the group score number from
which the group score is signicantly different. Due to unequal sample sizes, post hoc pairwise comparison tests were conducted utilizing the Hochberg
test statistic. *p < .05.

124

T. Schoenherr and C. Speier-Pero

ate data. Most likely, some type of data exist that may be suitable for predictive analytics, due to the ease with which data can
be collected todayit just needs to be identied, made palatable,
and worked with. It may also just need an illustration how existing data can be readily used for predictive analyticsthe data
may be there, but individuals may not realize its suitability. For
the remaining barriers, a similar pattern is evident, although not
signicant at statistically detectable levels.

Table 5: Skill sets required for data scientists

Understanding application of qualitative and

quantitative methods of forecasting
Numerical methods of optimization
Broad awareness of many different methods
of estimation and sampling
Determining opportunity cost
Using numerical methods to estimate
functions relating independent variables to
dependent variables
Using probability theory with actual data to
estimate the expected value of random
variables of interest
Quick design and implementation of
discrete event simulation models
Capital budgeting
Managerial accounting
Marketing science

Mean

SD

5.38

1.216

5.19
5.18

1.117
1.113

5.17
5.08

1.224
1.137

5.04

1.196

4.92

1.248

4.88
4.74
4.67

1.269
1.275
1.221

Positioning our ndings within the literature

Academic research into data science, predictive analytics, and
big data in SCM has been scarce. The few papers that exist
introduced and dened the domain, and called for further
research (Waller and Fawcett 2013a,b), stressed the importance
of data quality (Hazen et al. 2014), and tested the relationship
between big data analytics and operational performance (Chae
et al. 2014). It has been noted that big data analytics facilitates
enhanced decision making, increased visibility, better risk management, and overall greater value (Akkermans and Van Wassenhove 2013; Lycett 2013; Chae et al. 2014). However, we
were not able to nd academic research that addressed the current extent of use, motivations, benets, and barriers to predictive analytics. It is in these areas where our contributions lie.
Focused case studies and surveys of predictive analytics in
SCM have been published in practitioner journals and consultancy white papers. For example, Russom (2011) provided a rst
look into the adoption of big data analytics, Cecere (2012a)
offered insight into rms capabilities to use big data sources
and how to get started with data analysis, and Cecere (2012b)
identied big data supply chains as an exciting trend among her
respondents. Manyika et al. (2011) positioned big data as the
next frontier for innovation, competition, and productivity, and
Balboni et al. (2013) identied levers for companies to successfully pursue data analytics. In addition, LaValle et al. (2011)
drew attention to managerial and cultural obstacles to the adoption of big data analytics, and Jaspersoft (2014) found increasing
commitment to big data projects. While the insights provided by
these studies are valuable, our work builds upon these prior ndings by obtaining a signicantly larger number of respondents,
and by providing more detail in our analysis. Our results and

Table 6: Expert interviews on desired skill sets

Expert and title

Company

Tim Rey, Director of

Advanced Analytics

Steelcase

Philip Lear, Manager of

Trade Analytics

Kelloggs

David Dorleans, Manager,

Advanced Risk and
Compliance Analytics
Mike Marshall, Director of
Marketing and Statistical
Science
Richard Rodts, Manager of
Global Academic Programs
for Data Analytics

PriceWaterhouseCoopers

Jeremie Juban, Chief Data

Scientist, Statistics, Data
Mining, Machine Learning

Desired skill sets

Being able to convert data into business gain; being inquisitive about
problems; creativity, having a mathematical slant; statistics; machine
learning; operations research
Critical thinking; mathematics; programming; however, it is not really only
about crunching numbers and getting statistics, but to develop insights from
numbers; understanding the business behind it is thus important; passion
Ability to analyze the data, but then also to convey useable results and
implications to executives (communication skills)

J.D. Power & Associates

Quantitative skill sets, ability to nd and see patterns, passion for

discovering things; inquisitive mindset; technical capabilities and skill sets

IBM

Understanding what questions to ask (not necessarily with a big technology

background); being able to address business needs; leverage technology to
look further into data to facilitate better decisions; communication skills
(need to tell a story about why the data matter); mathematics; sociology
Being able to spend time with the data, coupled with the desire to
understand what is behind the data

The Weather Company

Source: https://accounting.broad.msu.edu/welcome/ms-business-analytics/events/

Data Science, Predictive Analytics, and Big Data in SCM

125

analysis therefore not only make valuable contributions to academic research, but also to the practitioner press.

TRAINING NEXT-GENERATION DATA SCIENTISTS

Educating scientists capable of mastering the challenges of predictive analytics is of utmost importance. This was highlighted in
a report by McKinsey & Company (Manyika et al. 2011, 3),
which estimated a shortage of 140,000 to 190,000 people with
deep analytical skills as well as 1.5 million managers and analysts to analyze big data and make decisions based on their ndings by 2018. This lack of trained talent thus represents a major
obstacle for realizing the full potential of SCM predictive analytics. In the following sections, we aim to contribute to efforts targeted at reducing this challenge. We commence with an expose
of desired skill sets for data scientists, as identied in our survey.
We then continue with a description of the design and implementation of a Master of Science (MS) degree program in predictive analytics at a major U.S. research university, as well as
efforts currently under way to bring the domain of predictive
analytics closer to our undergraduate population.
Skill sets for data scientists
While a data-scientic skill set and knowledge are crucial for
supply chain leaders (Waller and Fawcett 2013b), limited insight
exists on what these skill sets should entail. Waller and Fawcett
(2013b) conjectured about SCM data scientist skill sets that seem
to be more important. However, unable to conrm them, the
authors called for research to address which skill sets are
needed by SCM data scientists (Waller and Fawcett 2013b, 79).
We directly follow this call in this forward thinking article. Specically, we took the aspects that were deemed to be more
important for an SCM data scientists skill set summarized in
their table 1 (Waller and Fawcett 2013b, 79), and asked our
Figure 1: Growth in MS Analytics programs in the United
States based on initial year of program launch. MSA, Master of
Science in Analytics; MSBA, Master of Science in Business
Analytics; MSDS, Master of Science in Data Science.

respondents to indicate on a 7-point scale whether they feel the

particular skill is not needed at all (value = 1) or denitely
needed (value = 7). Table 5 presents the results.
The skills needed most, as suggested by our respondents,
come from the disciplines of forecasting (qualitative and quantitative), optimization, statistics (methods of estimation and sampling), and economics (determining opportunity cost).
Mathematical modeling and applied probability are also high on
the list. Skill sets still required, but deemed to be of lesser
importance, come from the disciplines of marketing, accounting,
and nance, which have been viewed as more general domains
and, on average, less quantitative, at least as compared to disciplines such as optimization and statistics. Contrasting the evaluation of the skill sets across our three usage groups and across the
functions yielded only nonsignicant differences.
We complement the insight derived by the survey with expert
interviews from companies promoting the use of predictive analytics. Table 6 summarizes the results of these interviews, suggesting additional skill sets required. Two key themes emanate
from these comments: data manipulation and communication/
interpersonal skills. With respect to data manipulation, data scientists must have the skills to extract transactional data from databases and data warehouses, scrape social media sites like
Facebook, and transform the data by integrating internal and
external data together into a single repository. Such skills require
database and data manipulation knowledge, both of structured
and unstructured data.
In addition, many interviewees emphasized the need for strong
communication and interpersonal skills. The implication is that
data scientists not only need to handle data very well, but they
also need to possess the capability to communicate the insights
derived in an effective way. The experts further noted that the
identication and development of appropriate skill sets is still
work in progress. For example, Richard Rodts, Manager of Global Academic Programs for Data Analytics at IBM, noted that
the position of the modern data analyst did not exist until the
very recent past. In addition, as Philip Lear, Manager of Trade
Analytics at Kelloggs, suggests, predictive analytics have
evolved from a somewhat obscure topic ve years ago, to something that everyone today wants.
Figure 2: Skill sets for data scientists.

80

60

MSDS
40
MSBA

Enterprise
Business
Processes &
Decision Making

MSA
20

0
2007

2008

2009

2010

2011

2012

2013

2014

Source: North Carolina State University (2014).

2015

Data
Management

Analytical and
Modeling
Tools

Data Scientists

126

T. Schoenherr and C. Speier-Pero

A MS in predictive analytics
Waller and Fawcett (2013a) encouraged the provision of a curriculum to train next-generation graduates to successfully tackle the
challenges of data science. We directly follow their call by providing a description on how a large U.S. university trains their
next-generation data scientists. In this section, we rst provide

some perspective on the growth of MS in Analytics programs,

and then focus on the development and initiation of the MS in
Predictive Analytics at this institution. The ensuing section then
describes the efforts that are currently being conducted at the
undergraduate level.
A number of universities have and/or are in the process of
implementing new curricula to serve the growing demand for

Table 7: MS programs in analytics

University
Arizona State
University
Carnegie Mellon
University*
DePaul University
Drexel University*
Fordham
University*
Louisiana State
University
Michgian State
University
New York
University
North Carolina
State University
Northwestern
University*
Northwestern
University
Rensselaer
Polytechnic
Institute*
Rutgers
University*
Southern
Methodist
University*
University of
Cincinnati*
University of
Tennessee*
University of
Connecticut*
University of
Maryland
University of
San Francisco
University
of Texas*

Time
Year
Credits commitment established

Program
length

Business
processes
Data
Analytical
and decision management and modeling Integration
making (%)
(%)
tools (%)
(%)

30

FT/O

2013

916 months

18

27

46

FT

2013

1221 months

36

23

25

16

52
45
30

O
PT/O
FT

2010
2012
2012

24 months
20 months
12 months

22
29

43
34
42

43
33
29

14
11

39

FT

2011

12 months

30

52

13

30

FT

2013

12 months

13

20

47

20

36

FT/PT

2013

1521 months

33

20

40

30

FT

2007

10 months

15

25

45

15

2011

20 months

10

10

70

10

FT

2012

15 months

20

27

33

20

30

FT

2013

12 months

16

34

34

16

43

FT/PT

2012

1221 months

30

20

40

10

33

FT

2013

1824 months

46

46

25

FT/PT

2011

1220 months

76

16

39

FT

2010

17 months

22

10

58

10

33

FT/PT

2012

12+ months

44

28

28

30

FT

2013

9 months

48

10

36

35

FT

2012

11 months

15

20

54

11

36

FT

2013

12 months

13

25

50

12

Notes: *The percentage mix reects required courses; elective course selections would change the actual percentage allocations. FT, Full time, on campus program; PT, Part time, on campus program; O, Online program.

Data Science, Predictive Analytics, and Big Data in SCM

data science capabilities. At one end of the spectrum this curriculum involves certicate and executive education programs to
expose participants to core concepts in data science. At the other
end of the spectrum are well-developed and robust curricula
leading to a masters or undergraduate degrees in data science.1
As illustrated in Figure 1 (North Carolina State University
2014), there has been tremendous growth in the number of MS
Analytics programs with a signicant uptick in the number of
programs in 2014 and 2015.
One of the most challenging curricular aspects of developing
sufcient applicable skills in data science is the breadth and
depth of diverse skill sets that are needed to be a highly capable
professional. Furthermore, individuals who may have an interest
in one dimension associated with data science are unlikely to
have developed skills or even been exposed to conceptual foundations and skills in another area. Thus, a data scientist must
develop deep conceptual understanding and applicable skills in
three areasenterprise business processes and decision making,
data management, and analytical and modeling tools (Figure 2).
Given the disparate nature of these areas, it is critical that the
development process includes experiential learning to provide the
necessary practice of integrating these areas and understanding
how to master the skills in specic domains.
We conducted an analysis of MS in Analytics programs that
were established in 2013 or prior, and which have received recognition as a top program (BISoftwareInsight 2014; MastersInDataScience 2014). We excluded programs that focused strictly
on concentrations in existing MBA programs to examine differences among the specic MS curricula. As can be seen in the
summary provided in Table 7, there is tremendous variation in
these programscredit hours vary substantially, but a majority
of the programs identied require between 30 and 39 credits,
with some programs requiring considerably more. Consistent
with the number of credit hours, a majority of these programs
support degree completion in one year or less. Those requiring
substantially more time are often designed as online or part-time
programs, and typically target working students who cannot
attend a full-time program.
Perhaps, the most interesting variation can be seen across the
three knowledge/skill domains previously described (enterprise
business processes and decision making, analytical and modeling
tools, data management), being indicative of the different programmatic foci of the offerings.2 Stressing our conviction about
the value of practical application of course content, we added an

127

integration category to reect the number of programs that

included formal hands-on projects enabling students to gain
critical expertise in linking these disparate areas of knowledge
together.
As can be seen in Table 7, almost all programs have some
split of content across these three core areas, with the general
emphasis being on analytical and modeling tools (44% on average), followed by business processes and decision making (24%
on average), and data management (23% on average). This stresses the importance of not only the prociency in tools and
approaches but also how to make use of the derived data and
develop actionable insight and recommendations. A particular
emphasis in one area can be explained by where the programs
department is housed or where the program originated. In addition, all but two programs have designed integrative coursework
into their required courses, highlighting the recognition of the
benecial learning outcomes of this approach.
To provide the foundation for the three skill set areas and put
them into perspective, we highlight the analytics certicate program developed within the Institute for Operations Research and
the Management Sciences (INFORMS) designed to enable analytics professionals (and their employers) to have condence that
a person will bring a core set of analytics skills to a project
team (Nestler et al. 2012, 26). The INFORMS certication
involves prociency in seven domains where some domains
receive greater emphasis than others as illustrated in Figure 3
(INFORMS 2014). The certication domains are consistent with
the primary foci of many existing university programs. To illustrate the overlap with the three skill set areas identied above,
we highlight the job tasks identied by INFORMS accordingly
(see Figure 4). This aggregation of domains and the ensuing
framework presented the foundation for the curriculum development for the MS degree program in predictive analytics at our
institution, which we further describe in the following.
The university designed and implemented an MS in Predictive
Analytics program that commenced with the Spring 2013 term.
The one-year program, split into three semesters (spring, summer, fall), is designed to enhance a professionals existing quantitative background with deep and focused skill development in
Figure 3: INFORMS certication curricular focus.
Business Problem
(Question) Framing
6%
9%

Analytics Problem
Framing

15%

The title associated with the degree varies by institution and

includes data science, business analytics, predictive analytics,
data analytics, etc.
2
Our assessment of the programs was based strictly on the
required courses (elective courses could be taken across the three
domains and, as such, could not be meaningfully included in the
total allocations). Further, the course description for each program was reviewed on all of the university websites. While it
was straightforward to identify the core focus for most courses,
there were a number of courses that straddled two of the knowledge domains. In these cases, the researchers assessed, based on
course title and content description, which of the domains
appeared to be the primary focus.

Data
16%

17%

Methodology/Approach
Selection
Model Building
Deployment

15%
22%

Life Cycle Management

Source: INFORMS (2014).

128

T. Schoenherr and C. Speier-Pero

the three principal areas identied above: enterprise business processes and decision making; data management; and analytical
and modeling tools.
Experiential learning, embedded in each of the three semesters
the students are enrolled, involves an integration of the three
foundational areas by performing a corporate analytics project in
partnership within an organization. Participating companies have
included Fortune 500 and medium-sized businesses in the insurance, automotive, nancial services, energy, and manufacturing
industries as well as governmental entities. While many of the
projects involve customer-facing applications creating predictive
models regarding customer churn and customer lifetime value,
there is increasing interest in nancial, cost management, supply
chain (e.g., predicting logistics and quality failures), and human
resource (e.g., employee churn) applications. The analytical tools
applied have ranged from traditional statistics to sophisticated
predictive modeling applications (e.g., SPSS Modeler) and programming in R to develop customized statistical and visual
analyses.

The current curriculum for the MS program is presented in

Figure 5, with Figure 6 illustrating its evolution. As noted in
Figure 5, the 30 credits of content associated with the MS program cut across all three of the core foundations, and are highly
experiential and integrative. We have found that developing
depth of knowledge in business processes and decision making
is critical, yet is most impactful when developed within the context of a specic analytics problem.
While the content of the program is highly consistent with the
job task skills presented in Figure 4, there are several observations to share as we begin our third year of delivering this program, which is scheduled to commence with the Spring 2015
term. First, we cannot overstate the importance of developing
communication skills and the use of visualization in presenting
analytic ndings to business managers. While we have only a
single course focused on developing and practicing this skill formally, students practice and enhance their presentation of ndings and actionable recommendations in the experiential learning
projects and in almost every course.

Figure 4: INFORMS job tasks associated with each domain.

Domain I
Business Problem (Question) Framing
T-1
Obtain or receive problem statement and usability requirements
T-2
Identify stakeholders
T-3
Determine if the problem is amenable to an analytics solution
T-4
Refine the problem statement and delineate constraints
T-5
Define an initial set of business benefits
T-6
Obtain stakeholder agreement on the problem statement
Domain II
Analytics Problem Framing
T-1
Reformulate the problem statement as an analytics problem
T-2
Develop a proposed set of drivers and relationships to outputs
T-3
State the set of assumptions related to the problem
T-4
Define key metrics of success
T-5
Obtain stakeholder agreement
Domain III Data
T-1
Identify and prioritize data needs and sources
T-2
Acquire data
T-3
Harmonize, rescale, clean, and share data
T-4
Identify relationships in the data
T-5
Document and report findings (e.g., insights, results, business performance)
T-6
Refine the business and analytics problem statements
Domain IV Methodology (Approach) Selection
T-1
Identify available problem solving approaches (methods)
T-2
Select software tools
T-3
Test approaches (methods)
T-4
Select approaches (methods)
Domain V
Model Building
T-1
Identify model structures
T-2
Run and evaluate the models
T-3
Calibrate models and data
T-4
Integrate the models
T-5
Document and communicate findings (including assumptions, limitations, and constraints)
Domain VI Deployment
T-1
Perform business validation of the model
T-2
Deliver report with findings; or
T-3
Create model, usability and system requirements for production
T-4
Deliver production model/system
T-5
Support deployment
Domain VII Model Lifecycle Management
T-1
Document initial structure
T-2
Track model quality
T-3
Re-calibrate and maintain the model
T-4
Support training activities
T-5
Evaluate the business benefit of the model over time
Enterprise Business Processes and Decision Making

Source: Adapted from INFORMS (2014).

Data Management

Analytical and Modeling Tools

Data Science, Predictive Analytics, and Big Data in SCM

129

Second, the use of co-curricular experiences (as highlighted in

Figure 6) has been a critical component of the programs success. As part of the program orientation, students are introduced
to a company project sponsor (e.g., General Motors, Steelcase)
as well as members from the IBM analytics support team who,
in concert with university faculty, guide the students in performing their initial analytics project concurrently with their rst
courses. The IBM team provides students hands-on experience
managing an analytics project, and formal general training and
targeted guidance in using SPSS Modeler. This being thrown
into the re experience has resulted in students quickly identifying their knowledge gaps and knowing where they will have to
differentially focus to develop the necessary depth across the
three foundational areas. At the same time, IBM and the corporate sponsors have provided stellar support and guidance to the
student teams so that teams are able to make actionable and
meaningful recommendations to the corporate partner.
Third, as noted in Figure 6, the curriculum continues to evolve
as we better understand the capabilities of our incoming students
and the needs of the marketplace. For example, while we require
that students have a formal understanding of statistics to get
admitted to the program, we discovered that additional coverage
of statistics was needed to get students to have the level of prociency in model development and data mining that is required to

be a successful data scientist. In addition, we have enhanced the

emphasis on marketing analytics by deepening student skill
sets in web and social media analytics. To accommodate these
changes, we have incorporated the critical content of project
management, and legal and ethical issues into existing courses or
skills developed while conducting the experiential learning projects.
All graduates of our rst cohort are employed full time with
an average salary of $75,000. While our second cohort has not
yet graduated, we have seen an increase in the average salary to
$80,000. Students received offers representing a variety of companies including General Motors, Dominos Pizza, Capital One,
and Ernst & Young. Students also received offers from companies that have sponsored projects, creating even greater synergy
between the university and the sponsors. Overall, the mix of
opportunities includes both corporate and consulting rms, and
the job titles include Business Intelligence Developer, Data Analyst, Solutions Architect, and Consultant.
Graduates highlight the importance of integrating their newly
learned diverse skills in the form of experiential projects as they
transition from their university experience to their employers.
They note that depth in one skill and not others is insufcient
for success. Further, the graduates all highlight the importance of
being able to effectively communicate ndings and insightsto

Figure 5: Current curriculum and content of the MS program.

Introduction to Business Analytics (3)

Applied Statistics Methods (3)

Data Mining (3)

How digitized business processes and data analytics are essential

to the performance and competitive advantage of a modern
corporation.

Application of regression models including

simple and multiple regression, model
diagnostics, model selection, one- and twoway analysis of variance, mixed effects
models, randomized block designs, and
logistic regression.

Techniques and algorithms for knowledge discovery in

databases, from data preprocessing and transformation to
model validation and postprocessing.

Intro Stats (2)

Web Analytics (2)

Applying Analytics to Solve Business Problems (3)

Application of statistical concepts including random variables,

distributions, parameter estimation, hypothesis testing, analysis of
variance and time-series analysis. Develop modeling
understanding of when to use what analytical capability.

The collection and analysis of information

from the web, including predicting future
behavior, search engine optimization, landing
page optimization, and mobile marketing and
analytics.

Application of data mining and analytical modeling

techniques to solve corporate business problems (e.g.,
customer churn, customer loyalty, market segmentation)
using data sets from within and across companies.

Computational Techniques for Large-Scale Data Analysis (3)

Practicum (3) OR Internship

Capstone Project (3)

Emerging issues in big data (e.g., collection, warehousing,

preprocessing, and querying; mining, cluster analysis, association
analytics; MapReduce, Hadoop; out-of-core, online, samplingbased, and approximate learning algorithms; model evaluation
and applications, etc.).

Corporate analytics project or internship

designed to integrate strategic business
understanding with analytical and modeling
skills. Manage project engagement with
organization.

Corporate practicum in the development and delivery of

predictive data analysis for strategic decision making in
organizations. Application of the principles and tools of
analytics to real-world problems in R&D, marketing, supply
chain, accounting, finance, and human resources
management. Development and presentation of analytical
insights and recommendations.

Different approaches for strategic data management and business

analytics including relational databases, SQL, data warehouse
concepts. Business process understanding using SAP.
Application of business intelligence tools including Cognos.

Communication Strategies for Analytics (1)

Social Network Analytics (2)

Development of managerial level business communication skills.

Communication strategy development in oral and written form.

Analyzing social media to support organizational decision

making, monitoring social media, measuring social media
return on investment.

Intro to Marketing/Social Network Analytics (2)

Develop understanding of marketing business processes that
increasingly rely on analytics including customer acquisition,
marketing segmentation, and understanding lifetime value.
Use analytical tools to develop models to support these business
processes.

The colors correspond to the prior figure highlighting business processes and decision making (green), data
management (tan), and analytical and modeling tools (blue). The yellow highlights integrative content across
all three areas. The numbers in parentheses represent the number of credit hours.

130

executives and business process owners in rmsusing appropriate language (e.g., presenting impact ndings using visual
display to executives and highly technical models to business
process owners).
More on curriculum
In addition to the launch of the MS in Predictive Analytics program, the university is committed to developing analytics awareness and understanding across our very large undergraduate
population. Given the number of credits associated with our business degree (across all majors, including SCM), there are insufcient credits available for a student to develop the depth of
knowledge needed in data management and analytics, and modeling tools. However, we are using the business core courses to
develop awareness and strategic understanding of the importance
of using predictive analytics when making many business decisions, and we are providing hands-on touch points for every student using integrative experiences, as outlined in the following.
The university has initiated the process of integrating a common thematic modulecognitive computingwithin each course
making up the core business requirement that every business
major must take (SCM, marketing, nance, human resources,
information systems, statistics, and a strategy capstone). Speci-

T. Schoenherr and C. Speier-Pero

cally, students are exposed to cognitive computing using IBMs

Watson technology, which adapts and learns based on user and
inputs (IBM 2012; MSU Broad College 2014). As such, each
course leverages the universitys partnership with IBM Corporation and others to provide insights into how cognitive computing
and, more broadly, predictive analytics ts within that functional
area, culminating with a case study focused on cognitive computing in the strategy capstone. In addition, each major will look to
develop functionally specic analytics content to develop student
understanding of how analytics will augment and enhance business decision making in their area of study. For example, the
core SCM course includes a discussion of how Watson can be
used to facilitate and enhance supply chain decisions.
A further specic implementation of predictive analytics
involves the integration of Cognos Insight (a management-oriented business intelligence tool) in the core information systems
course. Cognos Insight enables managers to peruse large
amounts of data with an easy to manipulate graphical user interface that relies on visualization of the data to identify interesting
insights and facilitate drilling deeper into the data to more fully
understand those insights. Every student conducts multiple analyses with the tool, and to meet industry needs, is required to communicate those ndings using actionable, business-oriented
analysis and recommendations.

Figure 6: Curriculum evolution over three iterations.

Notes: Blue arrows represent changes across the rst three years of the program. Shaded areas represent course modications (yellow) and new course developments (green).

Data Science, Predictive Analytics, and Big Data in SCM

CONCLUSION
This article brought insight into the rapidly evolving domain of
SCM predictive analytics and represents, to our knowledge, one
of the rst academic, large-scale surveys on the topic. With the
data collected, we were able to provide a current assessment of
the extent to which SCM predictive analytics are used in industry, together with the underlying motivations. We also identied
primary benets and major obstacles to SCM predictive analytics. In doing so, we offered additional insight into various usage
groups and their characteristics, explicating the current state of
data science, predictive analytics, and big data in SCM. Further,
we provided recommendations on how to train our next-generation data scientists. Insight in this latter part was generated by
the analysis of our survey data and expert interviews, combined
with our experiences in developing and implementing one of the
rst MS degree programs in predictive analytics, offering insight
into the future potential of data science, predictive analytics, and
big data in SCM.
Overall, it was our intent to provide a timely assessment of
the eld and motivate additional research and pedagogical
developments in this domain. As was illustrated, the eld of
SCM predictive analytics provides a promising avenue for transforming the management of supply chains, and offers an exciting array of research opportunities. For this purpose, and based
on our insight derived from the survey, expert interviews, and
the development of the MS program, we offer the following
possible avenues for future investigation. From a strategic perspective, there is a need for understanding the specic types of
supply chain questions that rms are addressing with analytics
and the measured value of the insights derived from analytics
activities (e.g., return on investment). Similarly, more formal
investigation into the barriers impeding the adoption and infusion of predictive analytics into the organization would be valuable. In addition, linked to the adoption of business analytics is
the organizational structure that is implemented to promote and
support enterprise analytics activities. Associated questions in
need of further investigation include the following: What are
the differing impacts of having a centralized, distributed, or
hybrid structure for successfully promoting analytics use within
the enterprise, and how might that structure change over time?
What corporate governance structures need to be in place to
enable and facilitate SCM predictive analytics? In addition, legal
and ethical issues in the use of predictive analytics, especially
as it pertains to consumer data, need to be investigated. Promising avenues for predictive analytics exist also in its application
to real-time risk management and dynamic resource optimizations. It is our hope that this article provides motivation and a
starting point to stimulate further research in SCM predictive
analytics, and to further infuse curricula with predictive analytics components.

REFERENCES
Akkermans, H.A., and Van Wassenhove, L.N. 2013. Searching
for the Grey Swans: The Next 50 Years of Production
Research. International Journal of Production Research 51
(2324):674655.

131

Balboni, F., Finch, G., Rodenbeck Reese, C., and Shockley, R.

2013. Analytics: A Blueprint for Value. IBM Institute for
Business Value, Executive Report, Somers, NY.
BISoftwareInsight. 2014. Top Big Data & Analytics Masters
Programs. http://www.bisoftwareinsight.com/resource-center/
big-data-analytics-programs/
Cecere, L. 2012a. Big Data Go Big or Go Home? Supply
Chain Insights (July 1, 2012). http://supplychaininsights.com/
wp-content/uploads/2012/07/Big_Data_Report_16_JULY_
2012.pdf
Cecere, L. 2012b. Voice of the Supply Chain Leader. Supply
Chain Insights (May 28). http://supplychaininsights.
com/voice-of-the-supply-chain-leader-2014/
Chae, B., Olson, D., and Sheu, C. 2014. The Impact of Supply
Chain Analytics on Operational Performance: A ResourceBased View. International Journal of Production Research
52(16):4695710.
Dwoskin, E. 2014. Big Datas High-Priests of Algorithms:
Data Scientists Meld Statistics and Software for Find
Lucrative High-Tech Jobs. The Wall Street Journal (August
8).
http://www.wsj.com/articles/academic-researchers-ndlucrative-work-as-big-data-scientists-1407543088
Hazen, B.T., Boone, C.A., Ezell, J.D., and Jones-Farmer, L.A.
2014. Data Quality for Data Science, Predictive Analytics,
and Big Data in Supply Chain Management: An Introduction
to the Problem and Suggestions for Research and
Applications. International Journal of Production Economics
154:7280.
IBM. 2012. Michigan State University to Introduce Students to
Analytics and Big Data Using IBM Watson. IBM News Release.
http://ibm.mediaroom.com/index.php?s=43&item=1453#
INFORMS. 2014. Candidate Handbook. https://www.informs.
org/Certication-Continuing-Ed/Analytics-Certication/Candid
ate-Handbook
Jaspersoft. 2014. Jaspersoft Big Data Survey Shows Rise in
Commitment to Projects and Decline in Confusion.
Jaspersoft Press Release. http://www.jaspersoft.com/press/
jaspersoft-big-data-survey-shows-rise-in-commitment-to-projec
ts-and-decline-in-confusion
LaValle, S., Lesser, E., Shockley, R., Hopkins, M.S., and
Kruschwitz, N. 2013. Big Data, Analytics and the Path From
Insights to Value. MIT Sloan Management Review 52(2):2132.
Lycett, M. 2013. Datacation: Making Sense of (Big) Data in a
Complex World. European Journal of Information Systems
22(4):38186.
Manyika, J., Chui, M., Brown, B., Bughin, J., Dobbs, R.,
Roxburgh, C., and Byers, A.H. 2011. Big Data: The Next
Frontier for Innovation, Competition, and Productivity.
McKinsey Global Institute. http://www.mckinsey.com/insights/
business_technology/big_data_the_next_frontier_for_innovation
MastersInDataScience. 2014. 23 Great Schools With Masters
Programs in Data Science. http://www.mastersindatascience.org/
schools/23-great-schools-with-masters-programs-in-data-science/
McAfee, A., Brynjolfsson, E., Davenport, T.H., Patil, D.J., and
Barton, D. 2012. Big Data: The Management Revolution.
Harvard Business Review 90(10):6167.
MSU Broad College. 2014. Broad Undergraduates Move Into
Our Cognitive Computing Future With IBMs Watson. MSU
Broad College Press Release. http://broad.msu.edu/2014/01/

132

22/broad-undergraduates-move-cognitive-computing-future-ib
ms-watson/
Nestler, S., Levis, J., and Klimack, B. 2012. Certied Analytics
Professional. Analytics, Sept/Oct, 2629.
North Carolina State University. 2014. Degree Programs in
Analytics and Data Science. http://analytics.ncsu.edu/?page_id
=4184
Russom, P. 2011. Big Data Analytics. TDWI Best Practices
Report, Fourth Quarter. http://www.sas.com/content/dam/SAS/
en_us/doc/research2/big-data-analytics-105425.pdf
Waller, M.A., and Fawcett, S.E. 2013a. Click Here for a Data
Scientist: Big Data, Predictive Analytics, and Theory
Development in the Era of a Maker Movement Supply
Chain. Journal of Business Logistics 34(4):24952.
Waller, M.A., and Fawcett, S.E. 2013b. Data Science,
Predictive Analytics, and Big Data: A Revolution That Will
Transform Supply Chain Design and Management. Journal
of Business Logistics 34(2):7784.

SHORT BIOGRAPHIES
Tobias Schoenherr (PhD Indiana University) is an Associate
Professor in the Broad College of Business at Michigan State
University. His research focuses on strategic supply management,
including strategic sourcing, the use of technology in procure-

T. Schoenherr and C. Speier-Pero

ment, and leveraging the supply base. He has published more

than 50 journal articles in outlets such as Management Science,
Journal of Operations Management, Journal of Business Logistics, Production and Operations Management, Decision Sciences,
and Journal of Supply Chain Management. He is an Associate
Editor for the Journal of Operations Management and Decision
Sciences, and serves on several Editorial Review Boards, including the Journal of Business Logistics and IEEE Transactions on
Engineering Management.
Cheri Speier-Pero (PhD Indiana University) is a Professor of
Information Systems at Michigan State University and currently
serves as the faculty director of the MS in Business Analytics
program. Her research interests include effective decision making
in technology supported work environments, individual acceptance and use of technology, and the effective use of information
technology to support supply chain relationships. Her work has
appeared in journals such as MIS Quarterly, Decision Sciences,
Organizational Behavior and Human Decision Processes, the
Journal of Marketing and the Journal of Operations Management, among others. She serves as Co-Editor of the Decision Sciences Journal. Dr. Speier-Pero was awarded the MSU
University-wide Teacher Scholar award in 2001 recognizing her
excellence in teaching and research.

Copyright of Journal of Business Logistics is the property of Wiley-Blackwell and its content
may not be copied or emailed to multiple sites or posted to a listserv without the copyright
holder's express written permission. However, users may print, download, or email articles for
individual use.