Leading the Spreadsheet Revolution
Making management science the best course
in the business school
By Steven G. Powell
A quiet revolution has occurred over the last decade in how management science is taught to business students. In the mid-'80s the typical management science course was strong on the concepts and algorithms of our field, but made relatively little use of computing. Specialized software, such as LINDO or STORM, were the most commonly used computing tools. Since that time, the spreadsheet revolution, which swept through the business world in the '80s, has come to management science.
The stage was set for this revolution with the development of linear and nonlinear optimizers for the spreadsheet (by Sam Savage, Dan Fylstra and Leon Lasdon), and Monte Carlo simulation add-ins (such as @Risk and Crystal Ball). Then in the early part of this decade, a series of textbooks appeared which presented management science using spreadsheets. I believe Don Plane's text was the first, followed by texts by Ragsdale, Winston and Albright, Savage, Camm and Evans, Hesse and others. Although no systematic survey is available as to how many management science courses currently use spreadsheets, the evidence from the textbook market and other indirect evidence suggests that spreadsheets are becoming the software platform of choice for teaching management science to business students.
Advantages of spreadsheets
What are the advantages of spreadsheets for teaching management science? The one overwhelming advantage of spreadsheets is that they are everywhere in business. Every business computer has a spreadsheet; many businesses rely heavily on spreadsheets for day-to-day operations; most managers recognize that spreadsheet skills are desirable in new hires; and (perhaps most important) every business student knows that he or she will use spreadsheets on the job.
Students once resisted learning specialized programs for management science because they believed these programs were not used in their firms. But not only is student resistance to spreadsheets low, many students actually want to improve their spreadsheet skills. If they learn a little management science along the way, so much the better. The high level of interest in spreadsheets throughout business gives all of us who teach management science a unique opportunity.
Spreadsheets have other advantages beyond wide acceptance. Business people recognize spreadsheets as important, but most think of the spreadsheet as simply a powerful calculator. That is to say, they see the spreadsheet as allowing business users to calculate useful numbers from existing spreadsheet models by specifying appropriate inputs (e.g., today's prime rate is X) and reading the appropriate outputs (the price of this security is Y).
But the spreadsheet can be much more than just a calculator; it can be in fact a platform for powerful end-user modeling [see "From Intelligent Consumer to Active Modeler, Two MBA Success Stories," Interfaces, Vol. 27, No. 3, May-June 1997, pp. 88-98; and "End-User Modeling," OR/MS Today, Vol. 24, No. 4]. By this I mean that the properly educated business analyst can use the spreadsheet to build simple but flexible and powerful models for making important business decisions.
This possibility simply never existed before the advent of the spreadsheet and the development of complementary management science tools for optimization and simulation. In fact, spreadsheets allow the average business analyst to use management science in many ways he or she could not before; for example, to bring cost-effective analysis to medium and small businesses, or to deliver a working model on a laptop to an executive.
So the spreadsheet revolution in business offers the teacher of management science two opportunities. One is to teach business students to be better consumers of management science analysis by using spreadsheets to make the concepts more concrete. An even more exciting and challenging opportunity is to teach these students (for the first time) to be active modelers. That is, to harness the fundamental concepts of modeling, such as abstraction, parameterization and sensitivity analysis, and to call on the sophisticated tools of management science (optimization, simulation and so on) when needed, to improve their own decision making and that of their organizations.
With these opportunities, of course, comes the challenge of adapting our teaching methods to the new technology and the new pedagogical possibilities. As a profession we know a great deal about teaching management science in the traditional ways. We have a lot to learn, however, about teaching with spreadsheets, largely because they have so radically lowered the costs and increased the power of analysis for mathematically-unsophisticated analysts. Furthermore, the transition from the traditional to the new teaching approach is difficult for many teachers. In order to help those new to teaching with spreadsheets make this transition, and to encourage experienced teachers with spreadsheets to share their ideas, we have organized a workshop on teaching management science with spreadsheets to be held in the summer of 1998 (see sidebar).
Disadvantages of spreadsheets
Spreadsheets do, of course, have disadvantages for some management science applications. There are many sophisticated management science models that are not well-suited to the spreadsheet platform. The spreadsheet calculation process can be slow for very large models, and specialized languages (such as GAMS, AMPL and others) allow one to describe a model in much more parsimonious terms than does a spreadsheet. But business students typically will not build large-scale or sophisticated models, and most teachers have found that the spreadsheet is entirely adequate for models that these students will want to build.
Spreadsheets do have some weaknesses, however, even for teaching basic concepts of modeling and management science. One weakness is that the spreadsheet grid itself has no intrinsic structure, so novice users need to be taught how to structure their modeling and spreadsheet building lest they create poorly organized models that are difficult to debug and use.
A second weakness is that spreadsheets do not require the separation of data and calculations, nor do they make it easy to see the formulas behind the numbers. So users need to be trained to separate data from formulas and to document their work so that the underlying logic is clear.
A third weakness is that, depending on how a spreadsheet is built, it can be difficult to add or delete dimensions. For example, adding products to a product-mix model, or years to a dynamic model, can require completely rebuilding the spreadsheet if it was not built from the start to accommodate these types of changes. So, again, users need to be taught to anticipate how a spreadsheet model will be used and changed over its lifetime.
Another issue that arises in teaching with spreadsheets is how to train the student to make the transition to more powerful software. Some teachers have rejected spreadsheets on the grounds that more powerful software is within the reach of their students. For example, one business program uses LINGO for optimization, Extend for simulation and Expert Choice for decision analysis. There is no doubt that all software is gradually becoming easier to use, and students are gradually becoming more skilled in using computers. As this occurs, perhaps we will find less reliance on spreadsheets in teaching management science and use of more specialized and powerful tools. This is only one of many open issues in the teaching community at this time.
What can we learn?
As an observer of the changing environment for teaching management science, I am constantly impressed with how much our profession still has to learn about teaching management science, and specifically about teaching with spreadsheets. I have learned many principles and tricks from some of the most innovative teachers, such as Don Plane and Sam Savage, that I would not have acquired on my own in several decades of practice. But as with all innovation, every discovery opens up new questions.
For example, I am a strong proponent of the idea that spreadsheets allow us now to teach modeling, but we are just beginning to work out the fundamental principles of modeling (see "Six Key Modeling Heuristics," Interfaces, Vol. 25, No, 4, July-August, 1995, pp. 114-125). Of course, if we teach more modeling in a time-limited course that must imply we teach less management science (in the traditional sense). A lively debate is in order about what the appropriate content of a management science course should be now that we have a technology that allows our students to be active modelers. And as I mentioned earlier, a related debate is needed as to the proper role of specialized software.
Another important question is how to integrate the teaching of management science with the teaching of statistics, now that both can be taught with spreadsheets. Most business programs devote more time to statistics than to management science. Rarely are the two courses integrated in terms of software, modeling approach or applications. My belief is that it is time for the management science community to rethink the artificial division between management science and statistics, and to push for these topics to be effectively integrated in the curriculum.
Another area where innovation is needed is in exploiting the graphics capabilities of the spreadsheet to improve both our teaching and end-user modeling itself. Just watch Sam Savage use simulation to show graphically how an exponential smoothing model performs over time on randomly generated data to get a sense of what is possible!
This is just a sample of some of the issues that will be debated at next summer's workshop.
What's needed now?
We find ourselves in a transitional period, with some fraction of the teaching community continuing to teach in the old ways with specialized software and traditional textbooks; another group introducing and refining the use of spreadsheets in their courses; and perhaps a third group which has gone beyond spreadsheets. As with all innovation, the rate of change is slow, and communication among all elements of the teaching fraternity is disorganized and faulty. Nevertheless, there is currently an unprecedented level of interest in teaching within the management science profession: witness the creation of an INFORMS task force on business school education (see their report at http://www.informs.org/informed/magnanti.html), the creation of INFORMED (the INFORMS Forum on Education), frequent sessions on teaching at our meetings, the creation of a Teacher's Forum column in Interfaces, and this "Issues in Education" column.
In addition to these efforts, we need a regular forum in
which to share and debate teaching innovations around
spreadsheets and other topics as our teaching evolves. Next
summer's workshop is a first step in this direction. If this
first workshop is a success, we plan to repeat it in 1999
and 2000 at different locations, emphasizing different
topics. Our long-term dream is to institutionalize a
permanent Summer workshop on Teaching Management Science. If
we are successful, this summer camp will become required for
all new teachers of management science, and experienced
teachers will want to return every few years to refresh
OR/MS Today copyright © 1997 by the Institute for Operations Research and the Management Sciences. All rights reserved.
Lionheart Publishing, Inc.
506 Roswell Street, Suite 220, Marietta, GA 30060, USA
Phone: 770-431-0867 | Fax: 770-432-6969
Web Site © Copyright 1997, 1998 by Lionheart Publishing, Inc. All rights reserved.