Rethinking Contact Center Metrics: The Problem with Averages
By Susan Hash
While the contact center environment has evolved a great deal over the past decades,
some metrics have not. Many centers focus on metrics that have been in place for
years without considering whether they still make sense for the current business
environment. Or, in some cases, the metrics may make sense for the operation, but
they're being measured the wrong way.
If you're relying on average metrics, then you're not getting the full picture of
your performance—and you're not collecting actionable data. Yet, that is what
most centers do. Service level, for instance, is typically measured as an average
number at the end of the day, which doesn't mean very much because most centers
have peaks and valleys of calls. "During some hours, you're overstaffed, so the
service level looks great; during other hours, there aren't enough agents on the
phones and service levels are low," says Penny Reynolds, founding partner of The
Call Center School. "If you're only looking at the average at the end of the day,
it's like having your head in the freezer and your feet in the oven. Your average
body temperature might be OK, but it doesn't tell you how uncomfortable the extremes
are—or how you can adjust scheduling and staffing levels to make it better."
Average handle time (AHT) is another example of a traditional metric that's commonly
misused today. Why? "Early call centers were primarily telephone operators—everyone
was doing the same thing and it took about the same length of time, therefore, looking
at the average handle time made sense," explains James Abbott, president of Abbott
Associates, and author of The Executive Guide to Call Center Metrics. But
because modern contact centers handle a vast array of things, "the metric should
be 'handle time,'" he says. "'Average' monitors the center point. We need to add
another metric to show how consistent the talk time is. That's where standard deviation
sigmas come into play."
In today's centers, variability is the largest driver of cost, Abbott adds, "the
bigger the variation, the more costly the operation." For example, if a contact
center handles 100 calls per day on average, and there is zero variation, forecasting
and scheduling would be a straightforward task. But let's say a center handles 100
calls on average. The first day, it receives 50 calls. The next day, it's 150, then
25, then 175 calls. "It's still an average of 100, but managing that center is a
whole different ball game," he says.
Publishing reports based on averages often shows the center performing at a higher
level than it actually is, which creates a disconnect between what your customers
say about your performance and the information that you present to your CEO.
Obviously, productivity metrics are important to run an efficient operation and
should not be scrapped entirely, but it may be time to rethink your metrics—in
fact, Reynolds recommends doing so at least once a year. "Start with a clean slate
and ask yourself what's most important, what does it look like when done well, and
how will you measure it? Sometimes starting from scratch is really what's needed."