Using The Theory Of Constraints To Improve The Forecasting Process: A New Approach To An Old Challenge
Source: 2003 APICS International Conference Proceedings.
Copyright APICS - The Educational Society for Resource Management. Used by Permission.
Forecasting has come to the forefront of the corporate agenda. Improving the accuracy and timeliness of internal forecasting and planning increases revenue predictability and operating margins — key to survival in a softening economy — and is often a prerequisite to participation in inter-enterprise collaborative planning networks. Furthermore, a growing number of businesses recognize that not all resources can be managed on a "Just-in-Time" (JIT) basis — downstream enterprise resources planning (ERP), supply chain management (SCM), and other operational system efficiencies still depend on the quality of demand signals fed to them.
Even those companies that have invested heavily in SCM, demand planning, and forecasting software continue to struggle with accuracy and speed. The results are missed revenue targets, excessive and inadequate inventory levels, and reduced customer satisfaction, as illustrated by a recent corporate forecasting controversy. Despite a $40 million investment in supply chain software, a major shoe manufacturing was faced with bloated inventories in some product lines and shortages in others. The company partially attributed its inventory problems to inaccurate demand forecasts generated by its new demand planning application. Quarterly earnings came in at 33 percent below initial estimates. And why?
Effective forecasting and planning requires more than merely selecting the optimal set of statistical forecasting techniques. It also involves a cross-departmental process that leverages enterprise-wide data, qualitative knowledge, and performance feedback — at its essence, effective forecasting requires a systems approach. This is the first nexus between forecasting and the theory of constraints (TOC).
The second nexus between forecasting and TOC is in its application. Fundamentally, TOC-based manufacturing processes are based on the pull approach (drum-buffer-rope). In TOC, the identification of constraints is fundamental. Identifying constraints, however, requires effective and timely identification of demand signals.
In the broadest sense, a manufacturing organization that follows TOC principles should also consider alternative ways to impact demand to better match existing constraints. This practice, known as demand management and exemplified by the sales and operations process (S&OP), is the logical evolution of an enterprise-wide consensus forecast.
ABOUT THE AUTHOR
Glen Margolis has over 15 years of operations, manufacturing, and systems integration experience, is a frequent speaker and author on the topic of forecasting and supply chain management, and is the founder of Steelwedge, Inc., the leading provider of enterprise forecast optimization software.