Software for managing rules has re-emerged in the market of late following a "near death experience" [1]. Originally known as expert systems and later knowledge-based systems, the term "business rules engines" has gained favor in recent years as vendors seek to capitalize on market opportunities.

Business rules engines are certainly a good approach to automating rules of compliance such as billing procedures, but are they an appropriate framework for decision-making? Consider a company selling travel packages over the Web. A visitor to the company's Web site might click on "warm weather deals" and be taken to a page that lists Caribbean destinations. Rules of this nature go beyond compliance to the very foundation of the sales process. Should Caribbean or desert destinations be displayed? Both? What options? By analyzing the continuous flow of information on how customers purchase, a company can adjust its actions to better achieve sales goals. In a nutshell, this is the e-sales promise of analytics/operations research.

Analytics and rules offer very different conceptual approaches to e-sales. While they are not altogether contradictory perspectives, the practical reality is that they evolved from different roots on different paths. As operations researchers, we appreciate the power of analytics as embodied in models, algorithms and data analysis. But rules constitute a view of the world many key decision-makers are more comfortable with; a world in which decisions are made and then executed. Attractive as business rules may appear, many hidden problems are often overlooked when adopting a business rules framework. Among the problems:

• Adopting a business rules framework postpones the difficult task of making decisions. Business rules engines provide a means by which to implement rules. It is easy to be enticed by a business rules engine's ability to implement rules only to find that when it is deployed the question of what rules to implement still remains. Without supporting analytics, a business rules engine represents a means by which to make poor decisions far more quickly and efficiently. Off-line analytics can be performed to help write the rules, but this circumvents the real-time potential of the e-commerce model.

• Performance measurement is difficult in a business rules framework. Businesses are adamant about justifying projects based on their return-on-investment. Since the value proposition of business rules engines is implementing rules, their return-on-investment is implicit — streamlining processes or providing a mechanism to apply rules more efficiently (again, sidestepping the value of the rules themselves). Systems based on analytics are designed from the outset to monitor revenue improvement against baseline metrics.

• Business rules proliferate. With a powerful tool in hand, it is easy for business users to create new rules, causing the list of rules to grow ever larger and more complex. Eventually it becomes difficult to determine if the rules are doing what they are intended to do. When instances are discovered where rules are not leading to expected results, the response is typically to add more rules.

• Business rules can impede analytical inquiry. Analytical approaches to e-sales derive their value from inferring future customer behavior. These inferences are more dependable when offers are made in a way that lends itself to analysis, specifically, limiting the type of offers in order to develop statistically significant conclusions. Business rules engines generate a web of offers that often make it difficult to infer what caused a customer to behave in a particular way: if A then B unless C except when D, in which case perform F until G or H.

• Suboptimal legacy practices are easily institutionalized. An early selling point of expert systems was the ability to automate the decision processes of human experts. If these processes can be improved upon by analytics, then business rules engines only serve to institutionalize suboptimal legacy practices.

The growth of the Internet is providing a rare opportunity to fundamentally rethink how we sell. By understanding what customers want and responding accordingly, higher sales and customer satisfaction can be achieved. While analytics and rules both have their place, decision processes with demonstrable value require an analytical foundation. Realizing the full potential of the Internet as a sales channel requires the dynamic application of analytical tools.

As systems for driving e-sales evolve, it is important that we remain sensitive to decision-makers' comfort with rules. But it is also important that we provide the intellectual leadership necessary for analytics to flourish at this critical period in the evolution of commerce.

References

1. Frye, C., "Business Rules are Back," Application Development Trends, July 2002. See also http://www.adtmag.com/article.asp?id=6492.

Table of Contents

OR/MS Today Home Page


OR/MS Today copyright © 2004 by the Institute for Operations Research and the Management Sciences. All rights reserved.


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


Web Site © Copyright 2004 by Lionheart Publishing, Inc. All rights reserved.