While fancy new forecasting models will always be of interest to researchers, there is plenty of really interesting and practical new work being led by forecasting practitioners.

Last month Steve Morlidge (who spent 30 years at Unilever, now with CatchBull), shared his promising new approach on the “Avoidability of Forecast Error” in the SAS/Foresight webinar series. (Follow the link for the on-demand recording.) Steve’s presentation was based on his article in the Summer 2013 issue of Foresight, and was discussed in a series of The BFD blog posts August 1 - 7.

The Fall issue of Foresight has another valuable contribution, this time from Martin Joseph and Alec Finney (who each had long careers at AstraZeneca, and are now running Rivershill Consultancy). Martin and Alec discuss the application of statistical process control (SPC) methods in their article, “Using Process Behaviour Charts to Improve Forecasting and Decision Making.”

I first met Martin ten years ago, when we both served on the Advisory Board of the Institute of Business Forecasting. Martin turned me on to what has become a favorite book:  Donald Wheeler’s Understanding Variation. Not about business forecasting at all, but rather a scathing critique of management reporting and the (mis-) interpretation of data, Understanding Variation got us thinking about the application of SPC to forecasting and planning.

In their Foresight article, Martin and Alec show how Process Behavior Charts can be created, and how to use PBCs to identify signals of a forecasting problem needing to be addressed. There is a lot of practical guidance on how to create charts (e.g., when there is trend in the data, as often observed when forecasting sales). They also show how to identify signals in the data – so that the forecaster will know when to act (and when to avoid overreacting).

They end the article with three “vignettes” illustrating their approach in familiar business situations, which you will likely recognize.

2014-Q1 SAS / Foresight Webinar

To get more of Martin and Alec, join us February 20, 11am ET, when they deliver the next installment of the SAS / Foresight webinar series, “The Forecasting Mantra.” Based on their Winter 2009 article in Foresight, Martin and Alec outline a diagnostic tool to enable the design of new forecasting / planning processes, and also assess the quality and efficiency of current processes. (I'll post a registration link when it becomes available.)

Prior webinars in the series are still available for on-demand review:

• Paul Goodwin, “Why Should I Trust Your Forecasts?”
• Mike Gilliland, “FVA: A Reality Check on Forecasting Practices”
• Mark Moon, “Achieving S&OP’s Strategic Promise”
• Steve Morlidge, “Avoidability of Forecast Error”