Q&A from Users

In particular product-market scenarios there could be situations whereby, depending on difference between what I bought the product at and what I could sell it at, I could actually make profit for the organization but with higher forecast error.

Is there a merit in calculating forecast accuracy for dollarized sales and decomposing it into Unit error and Price error?

There is every merit in tracking Forecast Accuracy in any organization. In fact, we will go on to say that the single most important metric in the entire organization is Sales Forecast accuracy. Unattended forecast error can be the root of most organization ills. Although it can manifest in many other forms, the diagnostic process has to be diligent to observe the root of the forecast error.

In this specific situation, what you really need to observe is the Unit forecast error which will abstract away from profit margins. So for performance measurement reasons, you should break this into two variance measures, namely the unit forecast error and the price variance or the profit variance.

You may want to reward the sales person who actually met his sales targets in units (or exceeded slightly) and exceeded the profit variance significantly. So the reward system will motivate unit sales within a smaller tolerance but profit margins with a larger tolerance.

Remember that a substantial positive unit variance is not necessarily a good thing. If the Sales forecast is selling 100K gallons of fuel while the actual sales performance was 200K gallons, the supply chain may be unprepared to meet this sales reality and may have to go through unnecessary expediting costs.

In the case of price variance, there is more tolerance but an exceedingly high price variance needs to be examined as well. This may earn a bad reputation for the firm as opportunistic. This may also illustrate a bad price forecast.