Staffing Demand Forecasting

An important ability for managers is forecasting staffing requirements. A variety of forecasting methods are available – weighted average, exponential smoothing, trend/linear regression, etc. A special complication to forecasting is the effect of seasonality on demand. To deal with seasonal effects in forecasting a forecast for the entire period (e.g., a week, month, year) is developed using an appropriate forecasting technique (exponential smoothing, regression, etc.). Then this forecast is adjusted for seasonal effects in each period (e.g., day, month, quarter). The most common technique for seasonal adjustment is the multiplicative seasonal method. It adjusts a given forecast by multiplying the forecast by a seasonal factor.

The steps in the multiplicative seasonal method are:

Step 1: calculate the average historical demand for each season (e.g., day) for the period (e.g., month) by dividing total demand for the month by the number of periods in the month. For example, if on the last four Tuesdays demand for employees on the day shift has been 9, 13, 8, 10, average demand for a week is 11 = (9 + 13 + 8 + 10)/4.

Step 2: calculate the average demand over all periods by dividing the total average historical demand by the number of seasons. If the average total average demand for a month is 196 and there are seven seasons (days of the week), the average daily demand is 98/7 = 14.

Step 3: calculate the seasonal factor for each season by diving a period?s actual historical demand (Step 1) by the average demand over all periods (Step 2). For example, if the average historical Tuesday demand over the last four weeks is 11, the seasonal factor for Tuesday is 11/14 = 0.786. Similarly, If Friday demand were 18, its seasonal factor would be 1.286 or 28.6% greater than on an average day.

Step 4: determine the forecast for a given period in the future by multiplying the average seasonal factor by the forecast demand in that future period (using your chosen technique).