WFO Insider: Forecasting Shrinkage

Shrinkage in the call center is inevitable.  Things will come up.  Whether planned or unplanned, agents will be late for work, call in sick, or have to leave early from time to time.  So, how do you plan for that? 

The concept of forecasting shrinkage is somewhat new to me. I've always been told shrinkage is very erratic and difficult to forecast. Most call centers I‘ve spoken to just use a flat line approach or go with the most recently observed shrinkage percentage. Our call center uses the most recently observed shrinkage percentage to plan for and forecast future shrinkage. It seems to work for the most part, and is much better than a flat line approach, but I've wondered if we could be forecasting more accurately. 

I recently had the opportunity to hear Penny Reynolds, of The Call Center School, speak at SWPP in Nashville and she shared some great insights on forecasting.  Not only did she share advice on forecasting for agent shrinkage, she discussed call center forecasting methods in general.  One takeaway: perhaps the most important step in forecasting shrinkage (or any other metric, for that matter) is gathering the proper data. If you’re not keeping an accurate log of agent time, how can you even begin to forecast for the future?

Generating historical data allows you to look for trends and plan and prepare for those similar times in the future.  For example, which days do more agents tend to call in sick?  Which month is the most popular for vacation time?

As Penny suggested, many companies are now trying to retain interval level historical data for 12-18 month periods. That’s not something we’ve tried in the past, but we are now considering it for the future.

How does your company forecast agent shrinkage?  What have you found works best? Feel free to drop me a line or share your thoughts in the comments.