Generating the most accurate forecast often involves creating a statistical forecast and then judgmentally changing the statistical forecast to account for market conditions which aren’t fully reflected in the historical data. Some common examples include: promotions; pipeline fills; large one-time orders; and product introductions or phase outs.
Forecast Pro TRAC accommodates judgmental changes to the statistical forecasts it generates via the Override Forecasts screen. This spreadsheet-style display, as shown in the screenshot below, contains five rows labeled Statistical, Override 1, Override 2, Override 3 and Forecast.
Override Forecast Screen Rows
The Statistical row contains Forecast Pro’s statistical forecast (based on history). The Forecast row contains the values Forecast Pro reports as the “Forecast.” If no values appear in Override 1 through Override 3, then Forecast is equivalent to Statistical; if, however, the user places a value into any cell in Override 1 it replaces the Statistical as the value which appears in the Forecast row (and therefore the “Forecast” reported out of the system). A value placed in any cell in Override 2 replaces both Override 1 and Statistical as the “Forecast.” Furthermore, any value placed in Override 3 replaces any previous Overrides in row 1 or 2.
In a nutshell, you can add or delete Override rows making it possible to have as many as ten judgmental changes or as few as one judgmental change to any given forecast, with the changes being treated as overrides. To customize the labels on the Override rows, simply right mouse click on the label directly in the grid and choose “Edit button label”.
Forecast Pro TRAC includes a number of flexible features for incorporating judgment into your forecasts. You can:
- Treat judgmental changes to the forecast as either incremental adjustments to the baseline statistical forecast or as overrides.
- Have up to 10 adjustment/override rows, each with its own customized row label.
- Create, retain and reapply adjustments/overrides as either hard numbers or formulas.
- Generate reports which detail all forecast adjustments/overrides.
Incremental Adjustments – Base Forecast Plus
Companies often seek to develop a statistical forecast which is a projection of underlying or “base” demand. In these cases the statistical modeling process is likely to incorporate some combination of outlier detection/correction and/or event modeling. As a result, the statistical forecast doesn’t necessarily include volume associated with things such as planned future promotions, expected one time orders, and anticipated competitor actions. The volume tied to these discrete events is added to the statistical forecast in the form of incremental adjustments for each event.
Settings – Defining Your Approach
To specify incremental forecast adjustments and the number of adjustment rows, choose Settings>Options on the main menu to invoke the Options dialogue box, and then click on the Overrides tab. The default is to treat judgmental changes to the forecast as overrides and to have three override rows. To change the setting to treat judgmental changes as incremental adjustments click the “Incremental” button in the “Adjustment mode” setting as shown in the screenshot below. The “Number of adjustment rows” field defines how many rows appear–5 in this example. Clicking “OK” after you’ve changed any settings will apply the settings to the current Project. Clicking “Set as Default” will change your settings for all Projects.
Reporting Judgmental Changes
Regardless of whether you choose an override or incremental adjustment approach, or whether you choose to save and reapply overrides, it is important to keep track of when, where and why changes are made. Forecast Pro TRAC allows you to make comments on all judgmental changes. In addition to making comments, Forecast Pro TRAC’s “Hot List” makes it easy to find and review all judgmental changes, regardless of how big your hierarchy is and how many adjustments/overrides have been made. Reports such as the one shown below make it easy to review and report on adjustments/overrides.
If you would like to see how Forecast Pro TRAC can help manage judgmental overrides and address other forecasting challenges, schedule a personalized Web-based demo with one our specialists.
We also invite you to attend the one-hour live Webinar “How to Effectively Combine Judgment with Statistical Forecasts” on July 17 presented by Paul Goodwin, Professor Emeritus, University of Bath. He will present a series of practical recommendations to ensure that you are making forecast adjustments for the right reasons and show you how to avoid pitfalls. The session will include a demonstration in Forecast Pro TRAC using real-world examples.