How to Measure The Accuracy of a Sales Forecast

If you are a CEO or a VP of Sales, it is extremely important for you to know how to measure the accuracy of your sales and demand forecasts. The forecast processes most often used by companies today, as a fact, results in higher inventories, longer customer order lead times, and good customer delivery performance. However, not all salespeople are successful in measuring the accuracy of a sales forecast. In this article we will discuss how to assess the accuracy of your forecast, how to compare forecasts, calculate and show mean absolute error and how to calculate MAPE. Your company will be forecasting today either using purely manual processes, MS Excel or dedicating forecasting tools. Calculating the absolute error The mean absolute error (MAE) has strong capabilities for assessing forecast accuracy in a context of inventory optimization and it is very simple to calculate and use. The absolute error is the absolute difference between forecasted and actual value in number of items. Intuitively we can think of the absolute error as the number of items the forecast is off from what actually happens and absolute means that the formula disregards weather the forecast is too high or too low. All that counts is by how many items the forecast is off the actual value. Negative algebraic signs are therefore not regarded. As mentioned above mean absolute error is expressed in number of items, that is why when calculating the MAE and taking the absolute error of each row we should remove the negative sign (if there is) and if it is positive leave it, as it is. Calculating MAPE MAPE stands for mean absolute percentage error. We arrive at the MAPE by dividing the absolute error by the forecasted value. Intuitively MAPE as a parentage error provides us with the measurement of the forecast error relative to the actual value. Let’s discuss also why MAPE is not suited to compare forecasts. MAPE is in most cases not suited to compare sales forecasting and demand forecasting. The main issue is the sensitivity to sparse time series. Sparse series are items selling in very low quantities. Most retailers have large amount of such products. What is a good forecast? Statements such as "this was a good forecast" or "this was a bad forecast" are heard quit frequently. A quantity assessment of the accuracy of the forecast must be set in the context. For example 5% error forecasting national electricity consumption 24 hours ahead is very poor, but 80% error for a product launch is extremely excellent. Factors affecting the acccuracy of the demand forecast are many such as the volatility of demand, data aggregation level, amount of data, forecast horizon, sparse data, availability of event data and many more. To conclude, the easiest way to set the context is to have your forecast to be more accurate than your status quo. The key to effective planning is having a solid foundation to build on. By effectively planning for future demand you will be able to add significant value to your company’s profitability and bottom line.