How Accurate is your Call Forecast?

Call center staffing and scheduling will be largely determined by forecasting of the call volume. Thus, when a forecast is errant, it can cause serious repercussions in customer service. However, even in the best call centers there will never be 100% accuracy in forecasting. The number of variables from day to day, and week-to-week, as well as unexpected scheduling changes, can all affect how a workday varies from projections. When this happens it is important to drill down to find the reasons for the variations, and factor them in to future forecasts.Measuring the level of accuracy in your call center forecast requires more than just calculating workload percentages. Take a typical week where the Monday forecast was 12% under actual call volume, Tuesday was 8% under, and the remaining three weekdays were all 8% over. When those numbers are run the result would be an overall weekly forecast variance of 4%. Sounds pretty good – but it doesn’t recognize how customer service may have suffered on Monday and Tuesday by an insufficiently staffed call center. Or even more, how Monday morning between 9am and 11:30am there was even a bigger The lesson here is to be aware how instances of overstaffing and understaffing can cancel each other out, resulting in a forecasting picture that looks more favorable than it is. Forecasting can be rendered more accurate through the use of a simple standard deviation approach, and by examining intra-day forecast accuracy as well as just how close the daily or weekly numbers compared to the forecast.Of course, the ability to forecast schedules is dependent on the ability to forecast call volume. The challenge here is the number of factors that can impact this statistic, from online marketing to economic conditions to social networking. Analyze call forecasting data to uncover trends and over time these forecasts should zero in more accurately numbers. Look at the following:

  • Forecast in 15, 30 or 60 minute increments
  • Look at daily, weekly, monthly or seasonal pattern
  • Look for “special days” (holidays, sales promotion, payday, end of month, etc.)
  • Look for external factors (weather, events, etc.)
  • Plan for “internal” events such as marketing and social media campaigns, newsletters, company news, product launches, etc.

Watch this short video to see how call forecasting tools and simulation can help. However, even with these tools it is important to continuously “learn” from your past forecasting – what assumptions resulted in better forecasts, and what assumptions did not result in a good forecast?   

Close