As organizations confront the limits of forecasting, they finally realize the folly in a blind pursuit of unachievable levels of forecast accuracy. The best accuracy we can ever hope to achieve is limited by the nature -- the forecastability -- of what we are trying to forecast. Anything better than that limit (e.g. correctly calling a coin flip 60% of the time) is not consistently achievable. We need to focus our attention on reaching that limit of accuracy, however high or low it may be, and achieving it as quickly, simply, and cheaply as possible.
It is in this context of rude awakening that Forecast Value Added analysis continues to gain traction, and is approaching the mainstream as a forecasting performance metric. This was evidenced at two recent professional conferences.
At the INFORMS Conference on Business Analytics and Operations Research in April, the forecasting track drew a large and engaged audience for all five sessions. Andy Fisher and Mert Sanver presented on "Large-Scale Statistical Forecasting for Cisco's High-Tech Demand Patterns," including mention of Cisco's use of FVA analysis and reporting. (More information on Cisco's use of FVA can be found in Gartner's "Case Study: Cisco Improves Demand Forecast Accuracy" by analyst Bill Hostmann.)
At the IBF conference in early May, FVA was a featured topic in the Demand Management Forum, was mentioned in the keynote address by Jack Harwell of RadioShack, and was touched on in additional sessions from NEI and Newell Rubbermaid.
Jack was an "early adopter" who began using FVA analysis several years ago, comparing their forecasting process to having no process at all (just using a naive forecast), and to determine the impact of process changes (are they making the forecast better?).
NEI is a leading provider of purpose-built server platforms, appliances, and support, delivering products and services for software developers to deploy their applications as a complete, turnkey solution. Ferose Lambay and Rob Sheriff of NEI are pushing FVA into new areas like forecast compliance measurement, and re-affirmed the key message of not wasting effort on the "unforecastable" in the industry.
Sean Schubert and Ryan Rickard delivered a full hour on their FVA tracking and reporting at Newell Rubbermaid, and kept most of a large audience well into the lunch break with post-presentation Q&A. Sean will be speaking again in June at the IBF/APICS conference on Sales & Operations Planning. Ryan will be speaking (along with me and Mert Sanver of Cisco) at a 90-minute session on Forecast Value Added Analysis at the INFORMS conference in Charlotte, NC, November 13-16.
If you are ready to try FVA analysis at your organization, a good place to start is the whitepaper "Forecast Value Added Analysis: Step-by-Step," available for free download.
