 |
||
|
Measuring Productivity in the Before I joined the call center world I headed up the Budgets department for my organization. In that role I had learned something about how to create financial plans for, among other functions, the call center. The methodology for calculating staffing needs, reduced to its simplest measure, was total calls predicted divided by calls per agent. The calculation was much like that for other functions that handled customer service letters, or collection cases, or whatever the widget was. Then, when I foolishly accepted an assignment in the call center, I learned what every call center professional knows. Agents have no control over the number of calls they have to take. I puzzled over why and where the link between high level budgeting and individual productivity targets broke down – which over time led to an experiment in measuring true, rather than relative, productivity. What follows is a short(ish) description of my attempt to re-establish that link, and to improve agent morale in the process. The Issues Measuring productivity in the call center is relatively simple: cost per call (delivered at an acceptable service level and meeting certain quality criteria – which could include FCR, accuracy of information provided, and so on). Taking this measure from the macro (department) to the micro (agent) level represents a real challenge. The accepted wisdom is that measuring agent productivity in terms of calls processed is not possible. Because agents have no control over the supply of calls, they cannot be held responsible for meeting specific calls processed targets. Instead, agent performance goals revolve around measuring relative, rather than absolute, performance using Average Handle Time and Adherence. Neither measure is really satisfactory. An agent may have the lowest AHT in the department, but may take far fewer calls than other agents who have a higher AHT. It is possible for an agent to have very high adherence while spending relatively little time available to take calls because of excused absence, off the phone time (special projects, coaching, performance feedback) and so on. In many call centers these exceptions are adjusted in the adherence calculations, so that gross and net adherence may be quite far apart. It proves that agents do have some control over minimizing the number of calls they take, and in fact can become quite adept in manipulating the system. So there is clearly a behavioral component that determines true agent productivity! Besides the AHT and adherence criteria, call center agents are often forced into a straightjacket of additional performance criteria – attendance, tardies, quality, and wrap time as a subset of AHT, and possibly other factors (First Call Resolution for example). It is hard to focus on too many goals, and many agents experience it as overly controlling. After all, most people who work in call centers have “grown up” responsibilities. They pay mortgages, raise families, do volunteer work and contribute to the well being of their communities – skills that allow them to set their own course in a responsible way. Why should we not build on these skills in the workplace? A Different Way Using that concept, I developed a different process in my call center and created a simple performance matrix:
That was it – no AHT, no wrap time, no attendance, no tardies (with some limits, which will be explained). Setting call targets had to overcome the problem where agents have no control over how many calls they take. To create targets that are equitable for all agents the following factors had to be considered and included in the calculations:
As a result, each agent would have potentially a unique target that was specific to his or her scheduled work time – so that the “unfairness” issue was taken off the table. A couple of examples illustrate the point:
These are obviously raw numbers, and they require some adjustment. First, lunches and breaks need to be taken out of the time available to work, so the number of calls would be reduced to allow for that. Second, an allowance is given for off-the-phone time. This would include a standard amount for coaching / feedback / team meetings, and general “wander time.” A flat percentage (which may be different for each call center, this is very much a cultural consideration) was applied to the calls target and reduced to allow for this unavailable time. Third, a standard allowance was given for absenteeism time. So an “x” number of sick days across the year was allowed against productivity targets, and again the calls target was adjusted by that factor. Vacation time and holidays were also adjusted to reduce the calls target. On the other hand, if someone signed up for overtime, calls to cover that period would be added to the target. Failing to work the overtime would then affect the agent’s ability to meet their goal. These allowances gave the agents some flexibility in how they wanted to meet their call targets. Some agents are naturally quicker on AHT than the standard, which would give them some extra “wander time.” Other agents are fine on talk time but have trouble with(or like to extend) wrap time, perhaps to relax between difficult calls. Others again are a bit slower on the AHT front, but can make it up by using less “wander time.” Agents could also influence their ability to meet call targets by controlling their sick time. Although taking sick time was no longer a performance measure in itself, using it beyond “x” days would affect their productivity and would make it more difficult to meet their call target. On the other hand, people who did not use all their sick time allowance would be able to take more calls, and beat their target. Of course, the question arose of “what will an agent do who has met her target two days before month end.” Would she stop working as she would be entitled to do? To deal with that issue a bonus plan was developed that paid out for calls taken over and above the monthly target (as long as strict quality criteria were met, and short calls statistics – a potential disconnect indicator – did not exceed averages). The result was that agents wanted to take more, rather than fewer calls. To further encourage this, a fourth allowance was created. Agents were allowed to set their own breaks, and they were given 15 minutes flexibility around their scheduled start time (except agents who opened up or closed a call group in the morning or evening). This had the effect that agents started looking for high and low occupancy periods caused by under- and over staffing. If their scheduled break was during a period where the staffing matrix showed understaffing, agents would remain on the phone until the queues got cleared, and then took their breaks. So it served their interests to be more productive – it helped them meet their call targets, or increase the calls handled numbers so that they would qualify for pay outs. Allowing some flexibility around start times had a similar effect – agents would target slow and busy periods. More than anything, though, they appreciated not being charged with tardies; if they got stuck in traffic and ran a few minutes late it was no longer a big stress issue. It created a huge amount of good will, corrective action related to tardiness was reduced to near zero, and early and late starts balanced each other out almost perfectly. A Culture Change To make this process work it was important to bring all the stakeholders on board:
The Agents caught on very quickly. The concept was explained in focus groups, and invariably one or more of the agents in the audience would pick up on the fact that using high occupancy would be to their advantage, so it was an easy sell. There was some unease about not getting credit for being out sick beyond the allowance, but they appreciated the offset of being able to cash in on not using all of their accrued sick time. Having just one metric to focus on turned out to be a tremendous benefit. Most wanted to see their call statistics every day to make sure they stayed on or were better than target. They really enjoyed being able to manipulate all the underlying variables that make up the call target to suit their strengths and weaknesses. Stress levels dropped significantly. It was very important that agents understood the calculations of their targets. Some tried to cheat by disconnecting calls; this can easily be controlled by checking for excessive short calls – which is standard anyway as it is a shortcut to reducing AHT. On the positive side, it no longer made sense for agents to avoid calls by going off available for a few seconds as calls are offered, then making themselves available again as the ACD worked its way down the available agents. The Team Leaders and Managers were initially very concerned about losing control of their agents. However, the enthusiasm of the agents alleviated much of that very quickly. Team Leader and Manager goals were aligned with those of the agents, and they were also eligible for bonuses depending on the team performances, so that the culture was reinforced across the call center. Workforce management was also wary in the early days about loss of control. Overtime, letting the floor self-regulate to some extent became a plus and created a much more relaxed atmosphere in the call center. There was far less data entry work to make exceptions and adjustments to performance statistics, so it saved a good deal of time in the group. Human Resources provided an important ingredient in the mix, to ensure that the format did not go against regulations. Also, to do an incentive plan requires HR approval. The process requires an automated application that calculates targets based on schedules, collects data to compare actuals against targets, allows making adjustments to targets and generates daily updates for agents and management to track performance. It also needs the ability to calculate bonuses at month end. I developed this custom application in-house. Results This different approach to managing productivity paid off in several dimensions:
Dick Lexmond has worked for and with a number of corporations, managing and consulting for a variety of call centers. He is founder and partner of Scintella Solutions, LLC, a company that develops call center oriented software applications. He can be reached at dlexmond@scintellasolutions.com. Copyright ©2012 Society of Workforce Planning Professionals. All rights reserved. |
||
