Business Analytics for Business Analysts in Manufacturing

Coleen Wilder

College of Business Administration

Valparaiso University

Valparaiso, Indiana 46383

Ceyhun Ozgur

College of Business Administration

Valparaiso University

Valparaiso, Indiana 46383

Coleen Wilder is an assistant professor of information and decision sciences in the College of Business at Valparaiso University. She earned a BS in Mathematics Education from Indiana University and an MBA with concentrations in Finance and Operations Management from the University of Chicago. Her PhD is in Management Science from the Illinois Institute of Technology. Dr. Wilder has 18 years of experience in the steel industry and five years in the real estate industry.

Ceyhun Ozguris a professor of information and decision sciences in the College of Business at Valparaiso University. He earned a BS in Industrial Management and a MS in Management from the University of Akron and a PhD in business (Operation Management/Operations Research) from Kent State University. He published a textbook by McGraw-Hill entitled Introduction to Management Science with Spreadsheets 1st edition with William J. Stevenson. Among others, Dr. Ozgur has published in Operations Management Research, Decision Sciences Journal of Innovative Education,Quality Management,Production Planning & Control, INTERFACES and OMEGA.

Business Analytics for Business Analysts in Manufacturing

ABSTRACT

Many of the skills that define analytics are not new. Nonetheless, it has become a new source of competitive advantage for many corporations. Today’s workforce, therefore, must be cognizant of its power and value to effectively perform their jobs. In this paper, we differentiate the role of a business analyst by definingthe appropriate skill level and breadth of knowledge required for them to be successful. Business analystsfill the gapbetween the experts (data scientists) and the day-to-day users. Finally, the section on Manufacturing Analytics provides real-world applications of Analytics for companies in a production setting.The ideas presented herein argue in favor of a dedicated program for business analysts.

Keywords: Business Analytics, Analytics in Manufacturing, data scientist, business analysts, big data, Business Intelligence, Data Driven Companies.

Business Analytics: the use of statistics and other operations research techniques, such as simulation, decision trees and other operations research techniques.

Analytics in Manufacturing: the use of statistical techniques to solve real world manufacturing problems.

Data Scientist: the skill set must include “a solid foundation in math, statistics, probability, and computer science” with the ability to write code at the forefront.

Business Analysts: a person need enough conceptual knowledge and quantitative skills to be able to frame and interpret analyses of a business problem involving big data in an effective way.

Big Data:the current phrase used to describe the changes in the accumulation of data over the past decade; the distinguishing factors of which are volume (2.5 Exabyte’s per day), velocity (speed at which data is created), and variety (images, texts, videos, etc.).

Business Intelligence: the process of gathering and transforming raw data into actionable insights yielding better decisions.

Data Driven Companies: companies that have real world problems involving big data and business analytics.

INTRODUCTION

Business Analytics is something most people have heard about but fewer know or can agree on the definition. Some argue it is nothing new, simply another word for Business Intelligence. Others argue Business Intelligence and Business Analytics are two different disciplines, each with their own set of skills and software (Gnatovich, 2006). The purpose of this paper is not to debate these issues; the two terms will therefore be used interchangeably. The focus herein is to position Business Analytics, by any name, in the undergraduate curriculum in a manner that best serves students. This task will begin with a discussion on the value of analytics to today’s businesses and will follow with suggestions on how to incorporate it into the curriculum. The expected challenges to implementing these ideas will be summarized in a separate section ending with ideas for future research.

BUSINESS VALUE

Business Intelligence/Business Analytics (BI/BA) is the process of gathering and transforming raw data into actionable insights yielding better decisions. Transforming data into insights is not a new discovery. Analysis of the 1854 Cholera Epidemic in London is one of many earlyexamples. Edward Tufte describes at length in his book, Visual Explanations, how John Snow used data and graphics to convince local authorities that the source of the epidemic was a water pump on Broad Street. Later in 1958, an IBM Journal article is credited as the first documented reference of the term Business Intelligence; it did not become popular, however, until the 1980swith the advent of Decision Support Systems(Kalakota, 2011). So, why are many higher education institutions just now creating Analytics majors, minors, and institutes?

Throughout history, businesses have adopted innovative management programs in order to remain competitive. In the early 1800s it was standardized parts. In the late 1800s, it was the era of scientific management followed by mass production. In the 1980s, businesses found their competitive advantage in lean production initiatives (Heizer & Render, 2011). Arecent Harvard Business Review article identifies analytics as the next source of competitive advantage for companies(Barton & Court, 2012).

Big Data has become an enabler for analytics. Big Data is the current phrase used to describe the changes in the accumulation of data over the past decade; the distinguishing factorsof which are volume (2.5 exabytes per day), velocity (speed at which data is created), and variety (images, texts, videos, etc.)(McAfee & Brynijolfsson, 2012). Big Data has opened the flood gates for data analysis to achieve heights not possible in the recent past.

Business schools are responsible for preparing students to succeed in current and future business environments. According to asurvey of CIOs, analytics and business intelligence was ranked as the number one technology priority; a position it has occupied in three of the last five years. Seventy percent of the CIOs rated mobile technologies as the most disruptive force with which they will be confronted in the next ten years, followed by big data and analytics each at fifty-five percent(Gartner Inc., 2013).

Analytics is a ubiquitous term in modern media. It is not difficult to find a story about a company using analytics to gain competitive advantage. Analytics is redefining companies. Overstock.com’s CEO once referred to his company as “a business intelligence company” not an online retailer (Watson, 2013). Several best-selling data-driven books have also attracted attention with catchy titles such as Freakanomics: a Rogue Economist Explores the Hidden Side of Everything and Super Crunchers: Why Thinking-By-Numbers is the New Way to be Smart. Hollywood has contributed as well with a block-buster movie “Moneyball” in which baseball players are evaluated using analytics; the film was nominated for six-academy awards. Google has created an Analytics service to help businesses monitor the effectiveness of their websites. And, Google’s Chief Economist, Hal Varian, said in a recent interview with James Manyika, “I keep saying the sexy job in the next ten years will be statisticians” (McKinsey & Co., 2009).

The value derived from analytics runs the gamut from cost savings to increased revenues. Nucleus Research found that analytics returns $10.66 for every dollar invested(Nucleus Research, 2011). Examples in marketing and managing customer relationships dominate the literature. Capital One, for example, used analytics to grow its customer base and increase the likelihood that customers will pay their bills. They conducted over 30,000 experiments a year using different incentives to find the best strategy(Davenport, 2006). Sears Holdings used data clusters to reduce the time it takes to generate new sales promotions from eight weeks to one week. The new promotions were even better than previous ones because they were more personalized and targeted to the individual consumer(McAfee & Brynijolfsson, 2012). Netflix, an early adopter of analytics, launched a million-dollar contest to anyone able to improve its movie recommendation performance by 10%; two of the top teams combined efforts to eventually win the prize with a 10.06% improvement (MacMillan, 2009).

Analytics opportunities are not exclusive to marketing departments. Talent analytics is one of several new terms used to describe the application of analytics to Human Resources. Companies are using analytics to help them improve everything from attracting new talent to making staffing decisions and evaluating performance (Davenport, Harris, & Shapiro, 2010). Location analytics is also gaining momentum. Companies are integrating geographic information systems (GIS) with other data sources to gain new insights about their business. Bankers are using location analytics to look at households and how they compare to their neighbors (Ferguson, 2012). Video analytics is yet another spin-off from the analytics family. AutoMotion Management, a North Carolina car wash chain, is using its video and analytics to both manage queues in real-time and to collect data for historical reporting (Zalud, 2013).

The skills needed to pursue the vast array of opportunities are almost as varied as the opportunities; the best place to begin perhaps is to look at the different levels of expertise that are required in an organization. Watson (2013) describes three skill levels corresponding to three different career paths as follows:

  • Data scientist. A data scientist’s skill set must include “a solid foundation in math, statistics, probability, and computer science” with the ability to write code at the forefront (Davenport & Patil, 2012, p. 74).
  • Business analyst. Business analysts “simply need enough conceptual knowledge and quantitative skills to be able to frame and interpret analyses in an effective way.(Manyika, et al., 2011, p. 105)”
  • Business users. Business users need to understand how data is stored, how to access it, and how to analyze it at a basic level.

McKinsey Global Institute estimates that by 2018 the talent shortage for those described as data scientists is estimated to be 140,000 to 190,000 and nearly 1.5 million for those described as business analysts (Manyika, et al., 2011).

ANALYTICS IN THE CURRICULUM

It is understood that business students are best served when they meet the expectations of industry; in other words, they have skills that are in demand. The first of the three skill levels, to be discussed, are those required of a data scientist. The title of data scientist was created by D.J. Patil and Jeff Hammerbacher in 2008; they were each working at LinkedIn and Facebook at the time (Davenport & Patil, 2012). Data scientists typically hold advanced degrees in a quantitative discipline. For this reason, companies typically use them to do the more challenging analyses and/or larger capital projects. These professionals know how to use advanced statistical methodologies and are adept at constructing complex models. Ideally, such a degree would be part of a dedicated institute such as North Carolina State’s Institute for Advanced Analytics. Most schools do not have the resources to pursue this path and need to choose from an existing platform; as long as the program is housed in one of the STEM (Science, Technology, Engineering, and Mathematics) disciplines, it should possess the necessary rigor required for the role of a data scientist.

The next skill level, to be discussed, is for the role of a business user. Their analytical responsibilities are primarily centered on descriptive statistics. They tend to look backwards at what has already happened. They need to use a company’s resources (data marts, warehouses) to access data and produce simple reports. Their primary knowledge base is in a business discipline that defines their work; for example, marketing, sales, or accounting. In addition to their core discipline, these professionals need courses in data management and basic statistics.

The final group and focus of this paper are business analysts. The McKinsey Global Institute referred to this group as “data-savvy managers.” They are often in positions that allow them to identify and exploit opportunities. In order to fulfill their responsibilities, they need to have a solid foundation in business complemented by analytical studies. Academia needs to take an integrated approach tothe curriculum of these professionals covering topics in data management, statistics, and communication, along with a business functional area such as marketing, finance and operations. This multi-discipline skill set is desirable in a data scientist as well; the difference between the two is in the weighting - data scientists’ studies are weighted heavier on the quantitative side whereas business analysts are weighted more on the business side.

It is imperative that the quantitative skills are reinforced in the business functional areas with projects and case studies. It is also highly recommended that courses for data scientists are not combined with those for business analysts and business users. It is tempting to teach general concepts to reach a larger audience thereby lowering costs but research has shown this method to be ineffective (Moore, 2001). Business students need business examples to facilitate their learning; it is unrealistic, in most cases, to present a generalized exponential growth model for bacteria and expect students to transfer the application to continuous compounding of interest.

Business analysts do not need to be experts in the various analytical tools. As “data-savvy managers,” they need to have confidence in the processes used by data scientists in order to identify opportunities and to fully exploit the results. Statistics departments are not usually structured to teach business functional content. Recruiters are often frustrated with where to find quantitative business students (Wixom, B. et.al 2011). An obvious choice to mitigate these frustrations is to position the quantitative courses (data management and statistics) within the business school. In a study of the top fifty U.S. undergraduate business schools, it was found that 68% of the schools taught the first statistics course in the business school and 81% did the same for the second course (Haskin & Krehbiel, 2011). Using these institutions as sources of “best practices,”it makes sense to position the quantitate courses within business schools. Further support may be found by looking at the preferences of recruiters. When practitioners were asked from where they hired business intelligence skills, 449 replied business schools, 158 computer sciences, 111 mathematics, and 109 engineering(Wixom B. , et al., 2011).

One of the authors worked in an operations research (OR) department for a large manufacturing company in a capacity which today would be called a data scientist. Personal stories follow to illustrate the importance of developing data-savvy managers.

  • Avaluable profitability model was developed using linear programming. Upper-level familiar with the methodology used to produce them. No manager is going to risk their career endorsing something they do not understand.
  • A new chemical lab was in the planning stage of construction. The chemical testing machines were seven-figure expenditures. The team leader did not want to purchase more machines than absolutely necessary to deliver a given service level. The team leader had an analytical background and recognized the value of using data to drive decisions. He also realized the stakes involved and escalated the problem to the OR team to resolve.

The first example was not usedby the company because the decision makers were not data-savvy managers; opportunities and profits were compromised. In the second example, the manager was data-savvy, and used the information to make an informed decision.

Analytics/Optimization examples in manufacturing

  • Soap Manufacturing Company was having problems with their capacity limitations. We established priority classes and wrote an optimization program to optimize the schedule given the priority classes. The priority classes automatically updated themselves. You can see the specifics in Production Planning Journal year 1997 (Brown & Ozgur, 1997; Ozgur, 1998).
  • In scheduling sequence-dependent set-ups for like items, we first used a cluster analysis method to group like items. In a cable assembly system, there were 93 cables produced on an automatic assembly line. We came-up with three groups/clusters of cables and once the cables are classified into 3 clusters of similar cables another program was used to sequence each cluster, once each cluster was scheduled another program was devised to combine the three clusters in an optimal way.Of course, each time subsets of items were selected, posing a different problem each time. (OzgurBai, 2010; Ozgur Brown, 1995).
  • In a problem faced by an Aerospace parts manufacturing company, there were defective units produced. It was the responsibility of management to determine out of 150 units of a single part which items were defective. We established thatusing discriminant analysis method after using regression analysis method. This was a manufacturing company that produced parts and supplied an aerospace industry company. We repeated this procedure for 110 parts produced by the Aerospace parts manufacturing company. This was a consulting assignment.
  • An appliance manufacturing company used SAS to help with quality control. Quality Control software was primarily used for predictive analysis of product defects. This was a consulting assignment.
  • A Fastener Manufacturing Company is interested in the optimal torque settings of their fasteners. They don't want the fasteners torque settings be too tight or too loose. The optimization program is designed to come up with optimal torque settings for their fasteners (Meek & Ozgur, 199; Meek & Ozgur, 1989).
  • A chemical components supplier to asteel manufacturing who supplies an automobile manufacturing company. The automobile manufacturing company passes back their quality standards back to steel manufacturing company who in turn passes back its standards back to chemical components manufacturing company. One of the authors was hired as a consultant for the chemical components company about improving their process capability. The company was using statistical software and trying to improve capability by increasing and above 1.333.

CONCLUSION