Published: February 13, 2013 | Comments
A big headache that I have come across in forecasting for a multi-channel center is the absence of any meaningful history for the non-phone contacts and back office activity. It becomes very difficult to produce a forecast when there is no report to show a pattern of what has happened in the past. Unfortunately, social media contacts seem to be the worst offenders. There are some 3rd party tools out there, like Google Alerts or HootSuite, but they tend to give high-level volume results.
Measuring handle times is often the roughest part, because sometimes when you simply ask the person who is handling the work "how long does it take?" they lean towards sandbagging you and tend to inflate the real length of the work. A process I have used to help with this is a Time Trial Study. In this method, a stopwatch is used between an agent and an observer to literally time the activities that are not measured otherwise.
In my ACCE session Workforce Management for Multi-Channel Centers I’ll be going over exactly how this works and the tools you’ll need to perform your own Time Trial Study.
First, I’ll tell you about the supplies you’ll need to carry this out (for both an analog and a high-tech environment) and how to spend the first week getting prepared and organized. It’s important to involve the agents who are being observed in the planning steps and to formally schedule the time that they’re going to be timed. When they are brought in as a partner on the project it helps with the defensiveness that sometimes follows from an exercise like this.
I’ll go over a step-by-step example of how to perform the data collection, how much data to grab, and give you tips for getting the best results. Doing a project like this on a buddy system helps immensely! The results are less biased and you can capture twice the sample-size on a short-term schedule. The entire Time Trial should be over in four weeks: one week to prep, the next two weeks to execute, and the final week to wrap-up and report.
I will also give a substitution method to deal with complex work types that have no real averages (like back office and IT work), and crazy pitfalls to avoid, all from personal experience. For example, during one time trial study, the agent’s system ran so slowly it was impossible to capture good information; but when the taskbar was unhidden, and the multiple YouTube windows that were paused in the background were closed, the problem was solved. Time Trial Studies have a tendency to shed light on oddball and suspicious behaviors that go on in the call center that may be otherwise unnoticed.I will share how to analyze the results and turn that into meaningful metrics that you can then apply to your forecasts and schedules. One of the most interesting results from an exercise like this is seeing how often your agents are being interrupted by other people in the call center, either from instant messages, interoffice phone calls, and even walk-ups. When this is captured as a new shrinkage category in the form of lost utilization or increased occupancy you may be surprised at how much more accurate your staffing models can become.
The rest of my session is also informative. I have plans to share exactly how to overlay a forecast with social media characteristics, and how to integrate that into scheduling. I’m also expecting a drop-in visit from Todd Hixson, an expert on using social media himself. Social Media is here to stay, even though it is really still in its infant stage. So for now we may have to sometimes rely on creative ways to capture the information we need to keep our centers in a planning culture. I’m hoping that the segment of my session on Time Trial Studies is helpful to anyone out there who may be struggling with the same issues.
Discover how well your workforce is doing with the Workforce Management Scorecard