The Institute of Business Forecasting's FVA blog series continued on March 2, with my interview of Steve Morlidge of CatchBull. Steve's research (and his articles in Foresight) have been a frequent subject of BFD blog posts over the last couple of years (e.g. The "Avoidability of Forecast Error (4 parts), Q&A with Steve Morlidge of CatchBull (4 parts), Forecast Quality in the Supply Chain (2 parts)). I asked Steve about the application of FVA analysis with his clients.
Steve notes that his goal is to measure forecasting performance compared to the alternative of simple replenishment based on prior period actual demand (i.e., compared to using the naïve random walk (aka "no change") model). He advocates use of the Relative Absolute Error (RAE) which is the ratio of the absolute error in the forecast to the error in the naïve forecast.
His findings, which are in some sense astounding, are that most businesses struggle to beat performance of the naïve forecast by more than 10-15%, and that 40-50% of low-level forecasts are actually worse than the naïve (so negative FVA!). His studies, reported in more detail in a series of Foresight articles, illustrate the generally poor state of execution of the business forecasting process.
The IBF interview includes two innovative reports Steve has developed to help visualize the state of forecasting at his clients. He argues that a value added perspective to performance measurement works better to expose the problem to management, and engage their efforts to make process improvements.
Read the full Steve Morlidge interview, and others in the FVA series, at www.demand-planning.com.
April Interview: Erin Marchant from Moen
Coming soon on the IBF blog site is my interview with Erin Marchant, Senior Analyst in Demand Management at Moen. Erin and I crossed paths last fall at the APICS conference, and her group is now in the early stages of an FVA analysis project.
In the interview Erin provides valuable guidance on engaging your business unit partners, and avoiding an "us versus them" situation. As she points out, FVA is a measure of the forecasting process, and is not intended as a means to embarrass individual process participants. [Note, however, that I myself am not above embarrassing participants who bias the forecast and contaminate the process for their own personal agendas.]