Early in my career, I was presented with a sophisticated analytic deck with a lot of pictures that honestly I didn’t understand by a well-dressed twenty-something-year-old consultant. He completed a customer service improvement engagement analysis, including developing a new staffing model that provided a significant improvement in contact center costs.
This consultant put together an annual staff plan using a curious method. He looked at the average volume week, developed staffing requirements for each hour of that week, added up the requirement, added in some shrink, and called the resulting staffing number the agent count for the entire year. He assumed that the resulting staff plan could be managed through overtime during any peak weeks. What he also did was to ludicrously oversimplify a very complex problem.
His analysis was obviously wrong. Our very seasonal operational plans cannot be reduced to an average week. Volumes, handle times, attrition, shrink, outbound contact rates and many other impactful metrics are all seasonal. Staffing a contact center requires serious rigor to make sure that, with all of the seasonality, each specific week is well planned for. Proper analysis needs to ensure that week-over-week we have the right number of agents, not too many or too few.
The consultant made a mistake in classification. He classified capacity planning as a short-term problem and used a short-term analytic technique to solve a fundamental long-term problem. Or maybe he didn’t understand contact centers.
The methods you choose for contact center analyses, in its simplest form, must match the problem you are trying to solve. If you want to know how many agents you exactly need next week, you can use a technique similar to that of the consultant’s—look at next week’s forecast and roll up the requirements to get an FTE count, add a little shrinkage, and you’ll know the number of agents needed. That works OK. But that one week cannot be extrapolated to any other week; too many important things change weekly.
Similarly, if you want to know how many agents to work overtime for later in the afternoon, looking at a forecast of those specific hours will give you a clear picture of the number of agents needed. But you cannot extrapolate those hours farther than, well, those hours.
I’ve seen this sort of analysis several times at conferences, usually coming from industry experts. Don’t buy into it. This sort of analytic myopia is the sort of simplification that can lead to terrible resourcing decisions!