Chris Crosby

You only have to spend about 30 seconds in any call center on the planet to hear the term Service Level thrown out as a KPI (Key Performance Indicator) for measuring “customer experience” and “accessibility”.

I’m going to tell you a little secret about Service Level that the rest of the industry is too afraid to acknowledge: not only is it a misleading and meaningless metric, it’s causing you to hemorrhage cash by overstaffing your call center. It is the single largest cause of wasted labor dollars in the industry.

Now, I don’t know the actual origin of Service Level, however my theory is that it’s roots go back to a long time ago in a galaxy far, far away when Ma Bell accounted for the majority of contact centers. Not only were these early call center pioneers not quite sure of what they needed to be measuring or how, but the first ACDs on the market couldn’t report on much of anything.

Service Level and ASA were probably logical choices because calculations such as Erlang C had existed since the early 1900’s and could easily be adapted to accommodate Call Center Staffing Models.

So one day the Jedi Council convened and Service Level, Average Speed of Answer (ASA), and Abandon Rate (ABA) were born.

In an effort to maintain peace and balance in the galaxy, our friends at the FCC and Public Utility Commissions (PUCs) legislated these core metrics into permanent call center history.

Lets take a step back for a moment and define what Service Level is; traditionally there have been two primary calculations:

(Calls Answered within Target/Calls Answered)

(Calls Answered within Target/Calls Offered)

Although I won’t go on record as saying that the first is “the industry standard”, I will say that in my experience I’ve seen the first one used far more often. This means that abandoned calls typically do not count against your Service Level. (What ever Service Provider dreamt that baby up was freggin brilliant…). 

OK, let’s say for example that your service level for yesterday was 80%, with a target answer time of 20 seconds. That tells us that 80% of the calls you answered, were done so within 20 seconds. Nice job, right? Well maybe, but to truly answer this question we need to take a look at what Service Level DOESN’T tell us. 

In this hypothetical scenario, we’ll assume you were offered 10,000 calls. Of those, you Answered 9,000. Of the 9,000 that you answered, 80% of those were in less than 20 seconds, or 7,200 calls. Based on this data we know that out of 10,000 callers 7,200 of them listened to less than 20 seconds of elevator music before getting to an agent. It’s probably safe to say that those 7,200 folks are relatively satisfied with that part of their experience.

Accounting for your Service Level Calls looks something like this:

But what about the other 2,800 customers? What was their experience like?

Well, let’s start with what we know:

1,000 of them abandoned in queue at some point. We don’t know if that was at 3 seconds, or 30 minutes. Or even if they called back and eventually got through.

The other 1,800 were answered, but at some point beyond the desired 20 seconds. What happens here can be a gruesome story that often goes untold. Based on Service Level alone, there is no way to ascertain if these callers were in queue for 21 seconds, or two hours…  

So in other words, there are 2800 customers, or roughly 1/3 of your callers, who’s experience you can’t measure or account for.

That excruciating sucking sound you hear right now is your ACD creating a giant data vacuum, void of the vital information you need to get a true picture of your call centers’ performance.

The other dangerous downside to Service Level is over staffing. One bad interval could ruin your results for the day. Often times when Service Level is contractual or regulated you could spend an entire day just trying to raise your % to recover. Or in some cases, you may already be overstaffed trying to insulate yourself from that spike in volume.

By it’s vary definition Service Level is a historical metric. Meaning that it tracks what has already happened. So why is your current staff on the phones compensating for what happened two hours ago? 

Call Center Magazine recently referred to Service Level as “a measure of accessibility” and insinuated that the general consensus is Service Level should no longer be used as a stand alone metric, but should be combined with other things like first call resolution to get a clearer picture of the customer experience.

This is actually disappointing to me as I don’t see any real value in a measuring something that is misleading, regardless of what you couple it with. I say we get rid of Service Level all together and start looking at what really matters to customers. Trust me, not only will your customers be happier, so will your stockholders.

-Chris Crosby

Trackback URI | Comments RSS