There are many systems that forecast future demand of products. Most of these systems utilize a “time series forecast model”. That is, they utilize some average of past sales or usage. For non-seasonal products, they will calculate an average of what you sold over the last several months. For seasonal items, the average is determined using what was sold in the upcoming months, last year. Some of these systems will even identify increasing or decreasing trends in usage. But is what you sold in the past always an accurate indication of what you will sell in the future?
- Did you experience unusually high usage due to one-time only projects or promotions? Calculating future forecasts based on temporarily high sales will cause you to over stock products in the future.
- Are you planning promotions to increase sales? Or, are there planned projects that will require unusually large quantities of certain items?
- Were sales temporarily lower than normal due to vendor delivery problems, inclement weather or some other condition that will end soon? Using unusually low usage to calculate future forecasts will probably result in stockouts when conditions return to normal.
Forecasting future demand solely on past sales is like trying to drive down the highway during rush hour after someone has applied paint to your car’s windshield. All you can do is look at your mirrors as you try to drive forward. You have a very good chance of ending up in an accident!
This is why “best practice” forecasting systems supplement sales history with additional information in the following ways:
- A listing of significant differences between the forecast and actual sales in the just completed month or week. Buyers and salespeople can determine if these differences are caused by unusual activity or represent the start of a new trend. With this information, buyers should adjust past usage to reflect what sales would have been under “normal” circumstances. If large differences between the forecast and usage cannot be explained, you probably need to look for better forecasting software.
- The ability to manually enter estimates of predicted increased sales of specific products on specific dates in the future due to planned promotions or projects. The accuracy of these estimates should be evaluated at the end of the event to help improve future estimates.
- The capability to enter trend percentages reflecting anticipated increasing or decreasing sales in certain product lines due to changing tastes or market conditions. Again, these predictions should be compared to the resulting actual sales.
Predicting future demand of products is difficult. And your results will not be 100% accurate. But implementing a “best practice” forecast will result in better estimates of what will sell in the future.