OR/MS Today - February 2000

February 2000

Changing the Face of Industry
Communication is the key that unlocks the boardroom door

By E. Andrew Boyd

The strength of the OR community is its ability to model complicated real-world problems. We speak the abstractions of mathematical programming and stochastic processes, of variables and equations. We can be relentless in our pursuit of the fastest algorithm, and take great pleasure in arguing the finer details of problem formulations with our OR colleagues. In doing so, we bring tremendous value that would not be brought by other disciplines. Where we frequently fail is in our ability to communicate with individuals outside the discipline of OR. While we have made great strides in raising the visibility of OR, we have yet to make the type of broad cultural impact that we are capable of making, and that will be necessary for business to achieve greater levels of profitability in the new millennium.

Communication with Executive Management

The discipline of OR is far removed from most boardrooms. Executive management is largely focused on strategic issues — product placement, capital spending, acquisitions, alliances, and profit and loss statements. These are typically the domains of the marketing, finance and legal groups within organizations. Tactical and operational issues characteristically show up on profit-and-loss statements, and the improved application of OR techniques is not always the first response that comes to mind for a CFO looking to improve the bottom line. Though OR is sometimes applied on an ad hoc basis to address strategic issues, it is rare to see an OR group with direct access to executive management, and much of the strategic OR modeling is performed by highly paid consulting firms with button-down credentials.

The introduction of business school courses in OR has had an important impact on the practice of the discipline. By introducing the concepts of modeling and interpreting the results of models, future executives know what OR is capable of even if they never practice it as part of their day-to-day activities. Those business school graduates who take an interest in OR can and do rise to prominent levels within organizations. Still, most of the strong advocates of OR within the business community are not at the executive management level. Until this level of management truly embraces OR, industries will make improvements in specific areas, but cultural change will not occur.

Altering this state of affairs requires the ability to communicate with executive management in their language. The requisite level of communication goes beyond being sensitive to the modeling issues that arise in real-world problems. Most CEOs do not think in terms of constraints and variables in a model. Most do not even think in terms of models — nor should they be required to. CEOs quite rightly think in terms of making decisions that will have a positive impact on the financial goals of the organization they are responsible for. It is a language entirely different from the language of OR, yet the semantics are simple: prove to me that your proposal brings real value to my organization.

Bob Crandall, former CEO of AMR, the parent company for American Airlines, understood the value of OR even though he probably never once formulated an OR model (see "The Patron Saint of Operations Research," ORMS Today, June 1998, pp. 28-29). In doing so, he changed not only the way American Airlines did business, but helped to influence the direction of the entire airline industry. By speaking Crandall's language, former TIMS president Tom Cook was able to build the AA Decision Technologies group to some 600 OR and related professionals before merging AADT into the newly formed SABRE group. While it would be a disservice to the airline industry to place too much credit with one carrier, there is not an OR professional alive today who has not benefited from the visibility of OR at American.

One of the most effective ways of bringing about long-term cultural change is to institutionalize the change in the form of a successful decision-support system. To gain high-level support, the purpose of such systems must be expressible in the language of executive management and not the language of OR. Supply chain management is one of the great OR success stories even though its growth in industry is still in its early stages. Companies such as i2 Technologies, Manugistics, IBM, AspenTech and others have laid out a business proposition that speaks to executive management. Revenue management is another area of very rapid growth driven by overwhelming success in the airline industry and the explosive expansion of the Internet. It is no accident that the name "operations research" rarely if ever appears at the highest levels of communication.

Communication with Software Developers

While speaking the language of executive management is essential to bringing about cultural change, equally important is the ability to speak the language of software developers.

OR-centrism all too often leads us to conclude that mathematical models constitute the main value of a decision-support system or, even worse, that the models themselves are the most complicated aspect of the system. For many of the applications of OR today this is, in fact, the case; data are gathered, a model is applied, and the results are used to generate reports. This limited archetype of the use of OR is not capable of bringing about true cultural change.

Next generation decision-support systems that genuinely impact the way business is done will be extremely dynamic and highly complex from a software design standpoint. Many different users will interact with such systems on an asynchronous basis, quite probably via the Internet from around the world. Terabytes of data hosted on geographically disperse platforms will be involved. "Enterprise-wide operating system" will be a more appropriately descriptive appellation than "decision- support system."

Professional software development is a skill that many people outside the field fail to truly appreciate. Writing and debugging a mathematical algorithm is far less difficult than designing, developing, deploying and maintaining a complicated system with many users. While it is not necessary for OR practitioners to be familiar with all aspects of software development, the inability to speak the language of software developers severely restricts the impact of OR. Large volumes of data, fixed computational time windows, and asynchronous system access all pose constraints that are integrally related to the mathematical model. It becomes impossible to separate the mathematical model from the abstraction layers and data models defining how the mathematical model interacts within the broader decision-support system. Defining abstraction layers and data models for software systems is typically the role of highly skilled software designers, but few of these designers come with the requisite skills to understand what makes sense in the context of the underlying mathematical model. If OR practitioners can't bridge the gap, the mathematical model may well fail to provide the value it was initially intended to.

A potentially worse state of affairs arises when OR practitioners are not involved in the core design of a system at all, but are brought in by software developers to "solve an OR problem that has arisen." The overall structure of the system frequently will not support the preferred mathematical model, necessitating costly design changes or compromises to the models that dilute the value of the system. This can only be avoided if OR practitioners step outside of their historical role as mathematical modelers and take on the added responsibility of speaking the language of software developers.

Conclusions

In the last decade OR has made tremendous progress in reshaping itself into a vital applied discipline. OR is now used more widely to solve business problems than ever before. Yet, in spite of the many successes, it remains for OR to make the broad cultural impact on organizations that it is capable of. Many industries that practice OR do so only within specific spheres of influence, and there remain a vast number of industries that recognize the need for the tools of OR, though they have never even heard of the discipline.

For OR to recognize its full potential requires that we learn to communicate in the language of other disciplines. Communication is essential to any successful undertaking, but it is a special challenge for a discipline that is fundamentally mathematical. We need to begin by understanding for ourselves that OR is not only about models and algorithms, but also a fundamental way of viewing the world around us. We can then convey the importance of OR to non-OR practitioners, and create software systems that deliver the value inherent in OR without requiring a degree in OR to use them. By setting our sights higher, we can help to focus business not just on solving specific operational problems that arise in the course of doing business, but to think of driving business forward from an OR perspective — with or without calling it OR.

E. Andrew Boyd received his Ph.D. in operations research from MIT in 1987. He is vice president of R&D and Senior Research Scientist at PROS Revenue Management, a software products firm serving more than 70 clients worldwide. He can be reached at aboyd@prosweb.com.

Table of Contents

OR/MS Today Home Page

Lionheart Publishing, Inc.
506 Roswell Street, Suite 220, Marietta, GA 30060, USA
Phone: 770-431-0867 | Fax: 770-432-6969
E-mail: lpi@lionhrtpub.com
URL: http://www.lionhrtpub.com