A quarterly publication of Society of Workforce Planning Professionals

A Deep-Dive into WFM: The Forecast Algorithm

Polite Stranger: So, what do you do?
Me: I’m a Workforce Manager at a Contact Center.
Polite Stranger: What does that mean?
Me: I make sure we schedule enough people throughout the day so that when you call us, you don’t have to wait a long time for someone to answer.
If there’s a follow-up question, I will go on to explain that we accomplish this by carefully generating accurate forecasts with creative schedules, and then we track all of that in real-time and report back on our progress. There’s a lot that goes into a functioning workforce management (WFM) plan.

But that all gets boiled down into a single question: “Do we have enough people to make our service goal next Tuesday at 3 pm?”

When you can definitively answer that question, then you know WFM. There’s usually a laundry list of obstacles standing in your way to the answer. Things like missing data, shrinkage assumptions, stale forecasts, schedule rules, handle time questions, attrition, utilization (or is it occupancy?), adherence-compliance-conformance, then understanding the difference between adherence-compliance-conformance, and the list goes on…

These are extra things we build into our plan to arrive at the correct answer. If next Tuesday’s handle time forecast isn’t up to date with recent changes, I may have the right number of calls predicted, but the workload will be wrong, which means the utilization rate, which is a derivative of workload, will also be wrong. Then the shrinkage factor will be wrong, which means any schedule optimization efforts might do us more harm than good.

And before we ask an agent to do something that will potentially disrupt their life, like changing their lunchtime or asking them to stay late, we should at least qualify that what we are asking them to do will truly solve our problem.

When You Understand the Group, You Can Anticipate What the Needs Are, Which Pain Points Hurt the Most, and How You Can Help

Once we trust how we look next Tuesday at 3 pm, we can also see how we look next week and if we’re positioned to meet service goals for the rest of the month and the rest of the year. Knowing this answer shows when we need to hire more people or when we already have too many people, so we can allow attrition to occur naturally without immediately re-hiring.

And it all begins with building a strong forecast.

The forecast is critical because it tells us what kind of work we’re doing, how much work we have, how long it will take to complete, how these service goals are defined, and ultimately, how many people we need.

This flowchart is a diagram that represents each step in the forecasting process.

There are some situations where a group has such a stable and predictable customer base that my re-forecasting process only needs to happen a few times a year. Other groups are so volatile that they require a midday refresh on both volumes and handle time forecasts. Step #3, Clean the Data, is usually the point in the forecasting process where I can identify how frequently that group will need to be treated.

Polite Stranger: Excuse me, I think I’m getting a call,and I really should take it. (Turns around and pretends to answer their phone)

I know this was a lot, and it’s always easier to show someone how to do this rather than just telling them the steps. If you’re interested in learning more about how to build a contact center forecast in Excel, let me know, and I will add your name to our waiting list for our virtual, two-day WFM Cra

SWPP Member Tiffany LaReau is a Certified Workforce Planning Professional (CWPP) at Human Numbers, a consulting firm that provides contracted forecasting and scheduling services. For more information, visit http://www.humannumbers.com.sh Course on Forecasting. But for now, good luck with your own forecasting efforts!

cross