Innovative Education: What every business student needs to know about analytics

Outsourcing “analytics” markedly changes the corporate world that business school graduates will be entering.

By Peter C. Bell

Bell 1

Designing an effective college analytics course is no easy task, especially since they have to be constantly revamped.

Designing courses is one of a college professor’s most important tasks. Designing a course requires an understanding of the role of your course in your program or school. For the core (required) OR/MS/analytics (from here on “analytics”) course(s) in a business school the role is usually to cover what every business student needs to know about analytics. What is this body of skills and knowledge?

A few years ago, I would sympathize with my colleagues teaching information systems (IS) because technology was changing so rapidly that they had to revamp their courses every year while the world of analytics was pretty stationary. Since we teach using cases, every year my IS colleagues had to write new cases or find new cases that presented the IS topics of the day, and this took a lot of time and effort. Rapidly changing technology will be no surprise to readers of OR/MS Today, but it may be a bit more of a surprise if I suggest that the world of analytics is going through a similar period of rapid change, and we need to constantly revamp our analytics courses to keep them relevant to today’s business students; we can no longer deliver the same courses that we were taught as students.

Business students used to see required analytics courses as a chore to be completed and then forgotten as soon as possible, but most of today’s business students arrive with an appreciation of the value of analytics, and many thirst to learn as much as possible during their business program. This thirst for “analytics” usually has to be satisfied by one, or at most two, required courses and an elective or two, and we also need to recognize that most students entering business programs have limited quantitative skills.

In the short time available, we have little hope of teaching most students to actually do today’s advanced business analytics. There may even be a danger that if we try to teach them to actually do their own analytics, we may end up providing them with superficial skills that could get them or their firms into deep trouble. What, then, does every business student need to know about analytics? The world of analytics practice may provide help with this issue.

The Edelman Prize competition, which has been run every year since 1982, provides a view of the evolution of the practice of analytics that can be used as a guide to the teaching of analytics. A look at the applications that have been Edelman finalists provides a framework for the design of a practical, “real” analytics course that appears to match the expectations of arriving business students.

The early Edelman finalists were predominately “traditional” applications of OR/MS. Manufacturing planning/scheduling, vehicle routing/scheduling, inventory management, utility production/scheduling applications were common, and there was a recurring theme that most of the early works were done by in-house analytics groups, perhaps with the help of an academic to do some of the heavy lifting. The topic and methodological content of these Edelman finalists closely matched the coverage of the typical introductory analytics courses of the time, but the nature of the Edelman finalists has evolved.

Today, many more applications are revenue-side as opposed to cost-side, there are more applications in services as opposed to manufacturing, and the way that the analytics is done has changed – many Edelman finalists firms have very limited internal analytics capabilities and outsourced the analytics. This trend to outsource analytics is part of a more general trend in the way that firms do business.

Outsourcing as a Way of Doing Business

Once upon a time there were lots of call centers in North America. As telecom costs declined, the labor cost came to dominate the profit equation, and call center operators migrated first to the Caribbean, then to places like India, China and the Philippines in search of cheaper labor. Software coding and legal and accounting services have followed a similar migration. Many North American firms now buy significant portions of these activities offshore.

Analytics came late to the migration offshore, and the motivation was different. Nationals from India and Eastern Europe taking advanced analytics degrees at impressive U.S. universities saw the opportunity to take advantage of the strong quantitative education offered back home. Major multinational corporations, including American Express, General Electric and Dell, set up captive analytics groups in India, China and eastern Europe (Hungary) to take advantage of quantitatively well-trained labor forces and the tax code. Many U.S. multinationals, for example, have profits trapped offshore for tax reasons and these funds can be used to pay for analytics performed offshore but “implemented” domestically.

Offshore providers perform analytics for many major firms including Wal-Mart, Allstate, Goldman Saks, HSBC, Citibank and Royal Bank of Scotland, consequently a firm that identifies a possible application has easy access to “buy” the needed analytical skills.

The ability to outsource “analytics” from highly credible and successful suppliers markedly changes the corporate world that the business school graduate will be entering. Successful graduates will most likely be buyers of analytics rather than doers of analytics. How, then, should we teach “buyers of analytics?” Here is my list of five fundamentals that every business student needs to understand today:

  1. Understand what analytics is.
  2. Have some basic skills to actually do some limited analytics.
  3. Be well enough informed about analytics to be able to identify potential value-enhancing projects and provide sufficient direction to a team working on an analytics application.
  4. Understand the role of analytics as part of a competitive strategy.
  5. Be aware of the major issues involved in outsourcing analytics work.

Explaining Analytics

While it is much easier to explain “analytics” than operations research or management science, “analytics” is still a complex topic with many dimensions. The INFORMS framework of descriptive, predictive and prescriptive analytics is a useful start to explaining analytics, but for many it is too granular. The SAS 8-levels of analytics framework seems to work better since levels 1-4 (1 = standard reports, 2 = Ad hoc repots, 3 = query drilldown, 4 = alerts) are familiar to most, and this framework also highlights the big step from 5 (the statistical analysis of historic data) to 6 (forecasting) and 7 (predictive modeling) (from historical data or “managing looking in the rear view mirror” to forecasting or “managing looking forward”), which a great many firms and managers find very difficult.

It is also helpful that optimization is level 8 so your coverage of optimization can include the fact that you are covering the pinnacle of analytics. This framework can be the basis for a lively discussion that often concludes that SAS has optimization in the wrong place; many firms have more difficulty with risk analysis and coping with uncertainty that they do with simultaneity in their decisions.

Basic Analytics Skill Set

Business students will be expected to be able to actually do some analytics when they start work. They will most likely do their analytics in Excel, and so the analytics course needs to provide a decent training in Excel-analytics. Excel-analytics is particularly strong for cash-flow modeling, small-scale event simulation (example: pricing options), descriptive statistics (including regression, although Excel Data Analysis is not the best for multiple regression) and small optimizations. Exposure to some Excel functions that are really useful in analytics but little known by most Excel users [for example =VLOOKUP(), =RAND(), =SUMPRODUCT and some of the financial functions =NPV(), +IRR()], as well as an understanding of data analysis tools and Solver may seem very basic to the analytics professional, but in the real world these seem to be gold.

Content beyond this level probably requires additional software. We used to use additional separate statistical, optimization and simulation packages, but the students became so confused with what was what that we decided that if it can’t be taught using Excel, we were not going to teach it. Since then we have developed an extensive set of 100 to 150 cases illustrating real business analytics that can all be “solved” in Excel. In some instances these cases illustrate cut-down problems, but in the majority of cases they are full-scale which leads me to the conclusion that Excel-analytics can be used by reasonably intelligent people to address real-world problems.

Knowing Enough Analytics

Identifying potential applications that will create value for the firm requires an understanding of both analytical methods and key application areas. Decision and risk analysis, predictive analytics including descriptive statistics and regression, simulation and the optimization of simultaneous decision problems would seem to be the minimum required analytical methods, but others could be included if enough time was available and depending on the content of other courses in the program. Application areas that should provide the new business school graduate with many ideas for new applications of analytics include pricing and revenue management, supply chain design and operation, options and insurance, media planning, “stress testing” and contract bidding and on-line procurement.

Most business graduates will not go to work in operations so they need exposure to potential applications in marketing, finance, human resources, strategy and IS as well as operations. Through teaching with cases, we try to select cases from all the functions to ensure that our students see the value of the analytics approach and methods across all functions of the firm.

Analytics as Part of a Business Strategy

Business students are very focused on strategy and so are very interested in the role of analytics in strategy and creating advantage. Interesting Edelman examples include low-cost quick hits [example: Merit Brass (finalist 1992)], one-shot big data/big analytics applications [example: Industrial and Commercial Bank of China with IBM 2011], multi-year mega problem applications that created advantage sustained over a long period [example: American Airlines 1990] and firms that have used analytics everywhere [example: FedEx (Mason et al., “Absolutely, Positively Operations Research: The Federal Express Story,” Interfaces, 1996)].

Examples such as these can be the basis of a good discussion on the value of analytics as a creator of advantage and how the firm might use analytics over time to sustain an advantage. Few textbooks include this topic although “Analytics for Managers” (Bell and Zaric, 2012) has a whole chapter on analytics and competitive advantage.

Issues Involved in Outsourcing Analytics

The new topic of interest to business students (particularly executives) is how to do analytics. There are many strategically important applications that were done by in-house analytics groups (example: Proctor and Gamble, 1996; American Airlines, 1990; Chevron, 2012), and in the early days it was common to include an “out-sourced” academic in the project team (i.e., Merit Brass, 1992; DRGs 1990). These kinds of applications require a significant and risky investment in building an advanced analytics capability in-house.

There is currently a shortage of analytics workers [Fortune (May 10, 2013) reported that “online help-wanted ads for data analysis mavens have shot up 46 percent since April 2011 and 246 percent since April 2009, to more than 31,000 openings now, according to job-market trackers.”], making hiring an analytics team difficult, but the larger problem is finding someone who can lead a new analytics function, establish an analytics presence and build an effective team. These individuals whose knowledge and skill set integrates business functional knowledge with a strong analytics capability are very scarce; consequently it will probably be a mistake for a new business school graduate to recommend to their CEO that the firm invest in starting an analytics group.

Many firms now outsource some or all of their analytics, but the outsourcing horizon is complex and not easy to master. INFORMS conferences regularly have an exhibitors’ area where various vendors selling “analytics” sell their products, but these are overwhelming products to help internal analytics professionals do analytics inside their firms. For example, simulation, data analysis software and optimization engines are being sold to firms for in-house use. I have not seen major vendors of “analytics” (such as GenPact, Cognizant, Mu Sigma) with booths at the INFORMS conference (although IBM is a frequent exhibitor and also a major vendor), but this may not be surprising since these firms may be seen as competitors of INFORMS members.

Firms such as these will do your analytics for you, but they need an internal champion who understands the many issues in managing outsourced projects. Even firms with strong internal analytics groups outsource some of their analytics, so an interesting issue is why they do this: Is it because of a cost advantage, or is there more to it? A second interesting classroom discussion issue is whether a firm can create a sustainable advantage using an outsourced, perhaps offshore, analytics provider.

Edelman finalists Dell (2012), with analytics provided by Dell Global Analytics, and Industrial and Commercial Bank of China (2011), with analytics provided by IBM Research, provide two examples where firms appear to have done this. The ownership of intellectual property (IP) is usually raised as an issue: How do you identify and maintain control of IP that you paid for and that is critical to your competitive strategy? This is particularly important because the offshore analytics business has gone through many mergers and acquisitions; you may well find that the firm that you hired to do your critical analytical work is now owned by another provider who works for your major competitor(s). [For more on these issues, see the case, “Competing with Analytics by taking Analytics Offshore,” Fogarty and Bell, Ivey Publishing (B13E008).]

The Opportunity

The analytics space used to be well defined:  You saw a problem, did the analysis, implemented the solution and then moved on. In some notable and important cases, the problem was so large that “doing the analysis” took years, employed teams of in-house analysts and provided a competitive advantage for years (examples: crew scheduling and revenue management at American Airlines).

The world is now more complex; if you (an analytics professional) see an opportunity where analytics might be useful, you can still do the work yourself or you could assemble an in-house team, but you can also form a captive offshore team to take advantage of skills available offshore, or call an outsource analytics provider, perhaps in India, China or eastern Europe and have them do the work for you.

Who will resolve decisions such as these intelligently? This looks like a perfect role for the business school graduate who has a good understanding of both analytics and business. These are our current students and making them aware of these choices is our opportunity to have a major impact of the success of our students and the firms where they are employed.

Peter C. Bell (pbell@ivey.ca) is a professor of management science at the Richard Ivey School of Business at Western University. He served as chair of the 2012 Franz Edelman Prize Competition.

Reference

  1. “Analytics for Managers” by Peter C. Bell and Gregory S. Zaric (Routledge, N.Y., 2012) covers the materials mentioned in this article in more depth and included in the core (BA, MBA and EMBA) courses taught at the Ivey School of Business and includes recommendations for cases that match the topics covered.