WFM Survey Results

This article details the results of the most recent SWPP quarterly survey on critical workforce planning topics. In this survey, which focused on forecasting processes, approximately 125 call center professionals representing a wide variety of industries participated and provided insight into this topic.

Participant Profile

The largest percentage (44%) of the participants was from large call center operations with over 500 seats, followed by 15% with 100-200 agents. However, 19% of the survey participants represent centers with 100 or fewer agents. All types of call center operations were represented in the study, with the biggest percentage representing financial, insurance, telecommunications, retail/catalog, and outsourcers.

Use of Commercial WFM Software

1

The survey participants were asked if they use a commercially available WFM system to create their forecasts. The vast majority (85%) said that they do use commercial software. Only 4% indicated that they use a homegrown system while 10% do not use a commercial product at all to forecast. This suggests that the market penetration of WFM software is high but that there are still some centers (perhaps mostly those in the smaller size) that do not have such tools. It is also true that some centers that have a WFM tool do not use them to forecast, focusing primarily on the scheduling and intraday performance management aspects of the tools.

Frequency of the Forecasts

2

The survey respondents were asked how often the WFM team creates a new forecast. Just over half of the respondents (51%) responded that new forecasts are created weekly while 18% indicate it is done monthly. Only 16% indicated that it is done daily with 6% in the “other” category that may be more frequently than daily or “as needed.” Some centers forecast only quarterly or less frequently (a total of 8%). The frequency of the need to forecast is generally driven by the volatility of the workload and the opportunity to adjust staffing to meet new  needs. For many, the seasons of the year, billing cycles, marketing events, and other drivers can shift the workload and staffing requirements substantially. Forecasting the changing need can provide a valuable guide to adjustments needed in the staffing that can be planned for rather than reacted to within each day or week.

Forecast vs. Actual Call Volumes

When asked to describe the variance between the forecast and the actual call volume at various intervals, the respondents generally provided answers in these ranges:

  • Monthly volume forecast variance is between 1% and 15% of actual.
  • Weekly volume forecast variance is between 2% and 25% of actual
  • Daily volume forecast variance is between 5% and 25% of actual
  • Half-hourly volume forecast variance is between 5% and 50% of actual
  • Don’t know – approximately 20% of respondents

There are some outliers in the responses with much wider variations reported by some. It is fair to say that the smaller the interval, the more difficult it is to have an accurate forecast and that is certainly reflected in this data. Some call types are relatively stable and others quite volatile with many unpredictable drivers. That makes it impossible to define an “industry  standard” goal for forecasting accuracy as it may vary considerably from one call type to another in the same center. Of course, matching up the staffing to this inaccuracy is the real challenge for operations to have any hope of achieving a consistent service level goal.

Call Volume Drivers

3

When asked what drives changes in the call volume, the top three picks were weather, marketing efforts, and system outages. Billing cycles and product releases, along with mailings, were the next most common answers. While some of these are certainly more predictable than others, building a solid history of the most common drivers in your center can greatly improve
the accuracy of the forecast over time.

When asked to describe the variance between the forecast and the actual handle time at various intervals, the respondents generally provided answers in these ranges:

  • Monthly AHT forecast variance is between 3% and 10% of actual
  • Weekly AHT forecast varianceis between 3% and 15% of actual
  • Daily AHT forecast variance is between 5% and 35% of actual
  • Half-hourly AHT forecast variance is between 5% and 30% of actual
  • Don’t know – approximately 40% of respondents

Like call volume accuracy, the responses include some outliers with some variance reported at 50% or higher, especially in the shorter intervals. It is important to notice that  approximately 40% of the respondents did not know what variation they experience in AHT forecast to actual compared to about 20% on the call volume accuracy. AHT is an “equal  partner” in the workload calculation so is of equal importance in getting to the right staffing requirement.

AHT Drivers

4

It is interesting to note that most respondents noted system outages as the largest driver of changes in the average handle time. This was followed by weather, marketing efforts, and product releases that more closely aligns with the call volume drivers.

Opportunities for More Consistently Accurate Forecast

Survey participants were to name one thing that would help the WFM team to achieve a more consistently accurate forecast. Here are some of the more common responses:

  • Better communication/links to other departments including Marketing and Operations.
  • Better information from client stakeholders.
  • More advance notice on known impacts to both volume and AHT.
  • Take away the non-WFM tasks such as payroll to enable more concentrated effort on forecast accuracy.
  • Increased ability to do “what if” scenarios and tracking of events that do occur.
  • Fewer skills.
  • Better and more complete data analysis to keep a record of abnormal intervals and days.
  • Better models for changes such as catastrophes.

There were a number of mentions of items on the technology wish list but several noted shortcomings in their current WFM tools. These may be real limitations but may be lack of knowledge of the possibilities that could be addressed with more direct training from the vendors on their products. If your team members have not had training directly from the vendor of your WFM products, you may be missing a number of useful capabilities. Be sure to put that in your next budget.

Conclusion

Achieving an accurate forecast of the workload that must be completed by the staff is the basic foundation of the WFM process. It is critical to creating a schedule that closely aligns with the demand. If the forecast is not accurate and schedules do not put staff in place when needed, the intraday operation can be chaotic and customers may experience significantly variable
service. Focus on producing a consistently accurate forecast is one of the most important functions within the WFM team. While there is no preset goal that every center or call type
should work toward, it is best to measure the current level of accuracy at each interval and work to improve it incrementally over time.