Accurate forecasting is critical for managing what in most call centers is a fluid environment. The goal is to achieve labor cost savings by balancing staffing needs against call volume expectations. Although many call centers use workforce management software in creating forecasts, manual forecasting is an option for a small call center environment. Common forecasting techniques include time-series, averaging, point-estimate and intra-day methods.
According to Six Sigma, a lean business philosophy, a time-series call volume forecasting technique is as appropriate for service desks and small call centers as it is for large businesses. Time-series forecasting bases call volume predictions on historical data, most often from the previous three years. The process involves plotting historical data in a graph that displays call volumes for each year on the vertical, or y-,axis, and the time measurement, such as months or weeks, on the horizontal, or x-,axis. Plotting historical data reveals past call-volume patterns, which you can then use to make future predictions.
Averaging Forecasting Techniques
Forecasting using averaging techniques includes simple mathematical averaging, moving averages and weighted averaging, which, according to the Society of Workforce Planning Professionals, is the most accurate. With weighted averaging, data that is more recent has more weight than older data. For example, if historical call volumes for a specific day over the past three years reveal the center received 2,400, 2,500 and 2,600 calls, the simple average is 2,500 calls. However, if you use weighted averaging and give 2,600 an 80 percent weight, and assign both 2,400 and 2,500 calls a 10 percent weight, the forecast is (2600_.80) + (2500_10) + (2400*10) = 2,570.
Point-estimate forecasting is the simplest forecasting method. However, according to the Society of Workforce Planning Professionals, it has shortcomings that most often affect its accuracy. It assumes that future call volumes will exactly match what happened in the past, regardless of whether the days, weeks or months included in the historical data were typical or atypical. Since the point-estimate technique doesn’t account for events or trends that affected historical data, what actually occurs on any given day the can be dramatically different from the forecast prediction.
Dealing with daily call volumes that vary significantly from daily projections is a challenge that most call centers face. Intra-day forecasting helps you deal with daily fluctuations that often require scheduling adjustments. This technique compares the current day’s forecast to actual call volume and agent scheduling requirements, aggregated into 15-minute to 30-minute periods. It then allows managers to create what-if scenarios based on service-level objectives, and if necessary, alter the forecast to suit changing conditions. Scheduling adjustments might include sending agents home early, assigning offline tasks or asking volunteers to work overtime.
- George Doyle/Stockbyte/Getty Images