Original Publication: Customer Management Insight - September 2008
View Online
There’s no magic formula for determining which KPIs are important or at what levels they should be set, says Jay Minnucci, president of consulting firm Service Agility, Inc. Rather, the goal should be establishing KPIs that provide a snapshot of performance that can help call centers stay on track and identify areas for improvement (see Checklist for Setting KPIs).
“Call centers have a ton of KPIs,” Minnucci says. “A lot of my work focuses on helping contact centers learn how to use them. There’s a whole set of KPIs that can tell you how an agent is performing, but each one of those shouldn’t necessarily have an objective or an action attached to it.”
First Do No Harm
What executives, managers, supervisors and even consultants should keep in mind, says Minnucci, is that KPIs shouldn’t put agents into conflict.
“If you set a transfer percentage KPI, you may keep some agents who should transfer calls [because they are inexperienced or ill-prepared] from doing so, “ he says.
This can set off a chain reaction: Your agents will do whatever it takes to meet the set transfer percentage standard; your call resolution is probably going to suffer for that, which means your customer satisfaction will fall; agents will be confused and probably angry when their supervisors take them to task for poor call resolution when, actually, agents can’t control that KPI (especially when they’re focused on keeping call transfers low).
The same goes for call handle time, says Minnucci: “You tell an agent to handle calls in 300 seconds or less, but you tell them they must meet the quality rate and give them certain things to do to meet quality. That creates conflict for the agent, and it doesn’t really contribute to better service for the customer.” And coaching sessions devolve into attempts to drum standards into agents’ heads. But no amount of this kind of coaching and no incentive — no matter how desirable — will help.
A Varied Approach to KPIs
To understand where contact centers go wrong with KPIs, you have to start with their starting points, says Minnucci.
Some call centers don’t even have KPIs. While they may be avoiding some of the pitfalls, they are losing out on the benefit that KPIs can actually provide.
Other centers have well-defined KPIs — but usually way too many, and they create a lot of conflict. In these call centers, data is overvalued and often misunderstood or misused.
Call recording is a good example of a tool that is allowed to take over. “If you pull out three or four of an agent’s monthly calls for quality monitoring, that’s not enough for me to determine if your center’s performance is consistent,” says Minnucci. “You shouldn’t put an agent on progressive discipline for a one-off.”
Another area where call centers can go astray with KPIs is customization. “When you start slicing and dicing columns of information and then adding them together with others, you get a really muddy picture of what’s really going on,” Minnucci says.
The point is this: KPIs are a gauge, not a control; they can tell you how things are running, but not how to run things. (See the sidebar on the L.A. County Child Support Services Call Center.)
Checklist for Setting KPIs
Jay Minnucci, president of Service Agility, Inc., has developed a checklist that he uses when he works with call centers to determine Key Performance Indicators.
1 Is the data accurate? That also includes finding out if the data is understandable, especially customized sets that have been sliced, diced and combined. (If the answer to this question is “No,” then don’t bother with steps 2–5 for this data.)
2 Is the data consistent? First-call resolution, for example, with all of its influencers, is not going to be a consistent piece of data.
3 Is the data statistically valid? Random or specious data sampling (say, pulling out three or four calls fore each agent per month for call monitoring) is not statistically valid.
4 Is this KPI in the person’s full control? Agents can’t always control FCR, but they can control adherence, making it a more fair agent KPI that promotes the right behavior (see number 5).
5 Does the current or proposed KPI promote the right behavior? KPIs such as call transfer percentages can be valuable if they’re used to work with an agent for retraining, but training to standards can promote behavior that sacrifices quality.
How KPI Reform Helped Turn around Performance at L.A. County’s Child Support Services Call Center
L.A. County’s Child Support Services Call Center’s 250-agent operation responds to some one million live calls and another three million automated calls each year. Dean Degruccio, former division chief at the center, says that during his tenure wait times were excessively long, agent morale was bad, and there was a general lack of expertise in managing call centers. So he called in Service Agility’s Jay Minnucci, then an executive in the consulting division of the International Customer Management Institute (ICMI).
In this mini interview, Degruccio describes the turnaround at the call center.
Q: What KPIs did your call center focus on before that might not have been as important as you thought?
A: Talk time. We changed from looking at particular call stats to looking at averages. No particular call stat was important; averages, though, became paramount.
Q: How were you able to strike a balance between KPIs such as call handle time and call resolution?
A: We changed the focus so that call center case workers were instructed to never worry about wait times; instead, they were to do whatever it took (within their power) to resolve the caller’s questions. If calls took a long time, that was okay. We didn’t want callers to be rushed off the phone (because if they were, they just called back, piling up the wait times). The mantra was “quality.” We learned that by investing a few minutes on the call, we received far fewer complaints, much higher call satisfaction, and far fewer call-backs. So, paradoxically, by spending more time on the phone, we reduced call wait times.
Q: What effect did loosening your grip on KPIs have on agents?
A: We began to use KPIs only as tools to discover what was really going on, not as an end unto itself. If a case worker, for example, had a very high average talk time, we didn’t criticize them for that. Instead, we tried to figure out why the talk time was long (whether it was inexperience, inability to focus or just goofing off); then we addressed that reason. We shielded agents from KPIs over which they had no control.
