I am going to age myself a bit and talk about an old movie. Google a video: “Spinal Tap – This One Goes To 11.” In this funny “rockumentary,” the not-so-smart lead guitarist is showing a person his amplifier, and how impressive it is, because the volume dial “goes to 11.” His amplifier plays “one louder” than the normal amplifiers, and he’s proud of it.

Now, of course, everybody knows this is silly. When setting up a metric, you can make the scale anything you want it to be. But what is important to know is that your metric choice might accidentally create some strange behaviors if you choose the wrong ones.

I recently was speaking to a contact center consultant who was discussing RFPs that they had seen. Several have some very peculiar service standards. For example, one was 98% within 60 seconds, and it was not unusual to have 90/20s or 95/30s in an RFP.

I may ruffle some feathers here, but in my mind, these are all examples of “it goes to 11 metrics.” Let me tell you why.

Goals are malleable and can be stated many ways.

Our service metrics can be expressed a number of ways. We can staff to an average speed of answer, an abandon rate (my favorite), and a myriad of service level metrics. If choosing service level as our goal, we pick a threshold (say, 30 seconds) and a goal (say 80% in 30 seconds).

Implicitly, there is a distribution of service associated with attaining any goal. For example, even though we may say that we want to hit an 80% in 30 seconds, we could – at the same time – also measure which percentage of our customers wait less than 20 seconds, or wait less than 60 seconds, or wait less than 5 seconds. When we hit that 80/30 exactly, those measures would tell us something interesting: we would be getting a view of the overall distribution of service we provide when we hit our chosen 80/20 goal.

I am making up these numbers, but let’s say in our example call center, that when we hit an 80/30, the distribution includes:

• 80 percent of our customers wait less than 30 seconds
• 85 percent of our customers wait less than 45 seconds
• 62 percent of our customers wait less than 20 seconds
• 28 percent of our customers wait less than 5 seconds
• 88 percent of our customers wait less than 60 seconds
• We also hit a 4.2 percent abandonment rate and a 65 percent agent occupancy

This just describes the natural distribution of wait times when we achieve an 80/30 and the natural occupancy and abandons that flow from our goal. But here is the rub — if we chose, instead, any of these other measures as our goal, the rest of the distribution will be achieved as well.

To achieve an 80/30, we could also use an 88/60 or a 4.2% abandon rate as our goal. And we will service our customers exactly the same way. Our goals can be expressed differently, but our actual distribution of service will be pretty close to equivalent no matter which of these we pick. Our true customer experience will be the same.

But why is this Spinal Tap math? It isn’t.

That wasn’t the Spinal Tap part. But this is: The higher a percentage of goal you choose, the harder it is to naturally hit. And this has to do with simple math.

If I choose a 95% goal (it doesn’t matter what the time threshold is), it is nay impossible to hit that goal after a bad couple of intervals. (I looked up nay—it means “verily,” and is fun to use). The reason it is hard to recover from a few bad periods is simply because there is not enough space between the 95% and the 100% to average your overall score back up! Holding volumes and AHTs, etc. constant, if during your first interval you hit an 80/30, it would require three periods of 100% to average back to a 95%. But it is awfully expensive to hit a 100% service level.

In contact centers, our economies of scale curves tend to asymptote near 100 percent. Figure 1, below shows the service you would expect by adding agents. Note that as you get past 85ish%, your curve starts to creep toward 100%. To get a 90% service level, you would need 46 agents. To get to 100% you would need an additional 10 to 15 agents, or to increase your headcount by 20-32%.

Figure 1. Agents plotted against service expected, holding volume and handle time constant

Figure 2 below shows what happens to your costs as you attempt to hit 100% (note these 2 graphs are not from the same model). As staffing levels zoom, so do costs.

But please know that if you are below your goal, it is not a wise practice to try to catch up on service level so you hit your goal by the end of the week. But goals are funny things, huh? If your center is measured by hitting a high percentage goal, it might mean that you go to great lengths, like adding a bunch of not-needed agents to bring your numbers up.

Figure 2. Service achieved versus cost per contact, varying staff, and keeping volume and AHT constant

To be honest, it would just be smarter to start with a goal that isn’t hard to average to. Say instead we choose 70 as our percentage, and whatever makes sense as our time threshold. By choosing a 90% or a 95% or a 98% service standard, we are telling management that our operation “goes to 11.”

Ric Kosiba is a charter member of SWPP. He can be reached at kosibas@comcast.net or (410) 562-1217. Please know that he is *very* interested in learning about your business problems and challenges (and what you think of these articles). You can find his calendar and can schedule time with him at realnumbers.com. Follow him on LinkedIn! (www.linkedin.com/in/ric-kosiba/)