How Often To Forecast

HOW OFTEN TO FORECAST By Larry Lapide (This is an ongoing column in The Journal, which is intended to give a brief view on a potential topic of interest to practitioners of business forecasting. Suggestions on topics that you would like to see covered should be sent via email to llapide@ mit.edu.) S S omeone recently asked me if it is better to generate forecasts more frequently. I gave my usual answer, which is to say “yes,” because I believe more frequency allows one’s forecasts to be fresher and reflect better any changes in demand sooner. However, the more I think about it, the more I believe it would have been better if I said, “It depends.” Updating forecasts too often can be as fruitless as forecasting too infrequently. One should probably take a Goldilocks approach—that is, do not forecast too often, yet not too infrequently either. I recently heard an interesting paradox from someone talking to some students about operational forecasting. He was saying how interesting it is that at many companies demand (i.e., order) data is captured on a daily basis and then aggregated into months to develop a monthly forecast. This monthly forecast is then used ...