OR/MS Today - October 2006|
The Shift to Knowledge-based Decision-Support Systems
By Brian Lewis
Recent articles in OR/MS Today have reminded me of the adage, "If you only have a hammer, every problem looks like a nail." The dominant decision support system (DSS) tools on the market today focus on manipulating data for simple information displays, and businesses are consequently forced to make decisions using this basic analysis, rather than by leveraging the real knowledge of their organizations. I argue that businesses should shift their attention away from these data-based DSS tools to knowledge-based DSS tools that allow them to capture knowledge in an active form and use it to improve decision-making throughout their enterprises.
The Data, Information, Knowledge and Wisdom (DIKW) hierarchy provides a conceptual framework through which to consider the common software tools that businesses use for enterprise decision-making. Each level of the hierarchy builds upon the previous level to provide increasing value to a decision-maker. Data is the most basic level of the hierarchy and consists of nothing more than raw observations and measurements, for example, transactional sales history. Data warehousing software and executive information systems (EIS) that store data are ubiquitous; however, the stored data provides no value on its own.
The Information level adds context to data by defining relationships, for example, the average sales per year describes one relationship between sales and time. Online analytical processing (OLAP) and traditional business intelligence (BI) software products slice and dice data and provide dashboard-style interfaces to report how a company is doing on key performance indicators (KPIs). The theory behind these systems is that you cannot improve what you do not measure.
One problem with these systems is that they provide historical views of a company's performance, not the forward views decision-makers need to make proactive, strategic decisions. It's like driving your car, but only using the rear-view mirror. As long as the road is long, featureless and straight, you can do it by watching how the road drifts behind you and making minor corrections to stay on course. But what happens when there is a fork in the road ahead (an opportunity)? Well, you won't see it in your rear-view mirror until it has already passed. And, if there is a sharp curve, you crash. While OLAP and traditional BI products are becoming more and more common, are they really the best technology investment given their focus on historical information?
I believe that they are not and that the next logical evolution of DSS tools will focus on the more valuable Knowledge level instead. Knowledge answers the how questions by applying scientific rigor to business problems and management estimates, expectations and insights. For example, how to calculate an optimal production schedule or how to determine the cash requirements to ensure you can cover all planned activities over the next year. New knowledge-based DSS tools will allow organizations to capture their own knowledge in an active format, such as quantitative computer models, as well to organize, distribute, integrate and ultimately use these models to make better decisions. These next generation DSS tools are now entering the market and integrate flexible, quantitative modeling capabilities and wiki-style, interactive knowledge portals.
There are software tools already on the market that claim to be knowledge management systems. However, most are really just information management systems. These products are typically Intranet sites with a sophisticated text-based search feature. The idea is that if employees document how they do things and place this information in static Web documents, a portal front end can be used to make this knowledge available to others. However, these systems don't really contain knowledge they contain information about knowledge. Someone still has to read the static document in order to understand how to make a decision and then go off and follow the newly learned process in order to determine a solution. With true knowledge management systems, knowledge is represented in an active format such as quantitative computer models, and decision-makers need only run the models to produce solutions.
The best knowledge-based DSS tools will capture the expertise and institutional knowledge of employees, who are an organization's most valuable resource. Analysts know how the market can affect the demand of a new product. Managers know how to calculate a company's return on equity. Engineers know how to estimate the cost of materials for an airplane wing panel given specific dimensions and machining processes. As businesses grow and expand globally, knowledge-based DSS software must be able to capture the knowledge that is spread throughout the organization in a decentralized way. Subject matter experts must therefore be provided with an interface through which they can translate their knowledge into quantitative computer models and then contribute their models to a central library.
While individual knowledge has value, complex problems often require a collaborative modeling approach where the problem is systematically broken down into smaller, more manageable pieces based on process, function, responsibility, expertise, etc. Component-based modeling is a systems modeling approach in which the pieces of the complex problem are modeled as separate component models, which are then integrated together into a parent model that represents the original problem. To ensure the efficient integration of these components, quality standards should be enforced as models are created. Dynamic links between models should ensure that the most up-to-date versions of models are used when the parent models are run, essentially enabling real-time decision-making.
Knowledge-based DSS software that uses component-based modeling concepts is ideal for collaborative work environments and provides four important advantages over traditional modeling practices when properly implemented:
Knowledge-based DSS software with component-based technology is already in use today. For example, Rolls Royce, in conjunction with the University of Southampton and the University of West of England, uses knowledge-based DSS software from Vanguard Software as part of its process to optimize the design of future aerospace products. They recognize that the majority of the manufacturing cost of a product is determined during the conceptual design phase. By collaboratively building a component-based, virtual factory model, design engineers at Rolls Royce can estimate the manufacturing costs of their designs and therefore design not only for performance, but also for cost.
Finally, at the top of the DIKW hierarchy is the Wisdom level, which involves an understanding of why and when to use knowledge for example, understanding that improving inventory management will increase profits. Wisdom is gained through experience, through networking and through a solid understanding of the lower levels of the DIKW hierarchy. Since the state-of-the-art DSS tools are just now beginning to focus on the knowledge level, it is difficult to foresee how the software products of the future might address the Wisdom level. But, anything is possible.
The DIKW hierarchy provides a framework through which to consider the enterprise software tools that businesses use to make decisions. While most tools on the market today focus on the data and information levels, new state-of-the-art DSS tools are focusing on the knowledge level and will improve an organization's collaborative, knowledge-based decision-making ability. While the OLAP and traditional BI hammers have their place, I believe that businesses will soon find a new favorite tool in knowledge-based DSS software.
OR/MS Today copyright © 2006 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
Web Site © Copyright 2006 by Lionheart Publishing, Inc. All rights reserved.