Our friends at APICS recently pointed out some thoughtful considerations for those companies who find that their forecasting performance for inventory, sales or production are not as accurate or stable as they would like. In the Mar/Apr 2014 issue of APICS Magazine, research director Jonathan Thatcher writes about the key issues companies must consider. We’ll reprise a few of his key suggestions today.
First, notes Thatcher, “Don’t schedule production based on the greater of the forecast or sales orders. Instead, make sure orders consume the forecast as they come in. Ideally, your ERP system should display figures for forecast, customer orders and requirements summary.”
So for example, when the forecast calls for 100 units and customer orders equal 25 for a given period, this leaves a remainder of 75 forecast units. Production does not care whether the units are ordered or forecast – they’re just “units” to them. Through Sales & Operation Planning then, your team can consider the ideal forecast to project, based on sales trends as well as how well production is meeting demand while avoiding adding unnecessary inventory.
A second issue Thatcher notes is what’s called the MRP demand time fence (DTF). Set the DTF equal to production lead time, and make sure your MRP system shows forecasts to zero within the demand time fence. If it takes one week to manufacture a unit, then the DTF would be one week. As the article posits: “Forecasts made inside the DTF should be ignored as it is too late to produce them, and those forecasts will overstate demand.”
And of course, don’t forget to flag extraordinary or non-repeating (“odd”) orders. Don’t make these outliers part of history on which regular forecasting is based.
Finally, make sure that your weekly forecasting period matches the periods used by sales. Sales rarely occur evenly week over week. Aggregate numbers, Thatcher points out, “are your friends” insofar as “data based on long histories is easier to forecast than daily or weekly data” with less variability.
By definition, APICS notes, “no forecast is completely correct. But we can get a little closer to perfectin with a stronger forecasting practice.”
For more information on this topic, try starting at the APICS Magazine site here. (Note: there is generally a lag-time between the appearance of an article in print form and its appearance at their site. The article excerpted above came from the “Ask APICS” section of the Mar/Apr issue.)