Who is Answering your Calls?
By Ric Kosiba, Ph.D., Interactive Intelligence’s Decisions Group
One of the tricks to building accurate computer models of an operation is to make sure that the data that feeds the model is clean, well defined, and understood. In the case of contact center planning software, it is important to know the call routing possibilities so we can model call overflows, agent’s secondary tasks, and the like.
Therefore, at the beginning of the engagement, we always ask our customers to map out the contact routing—we need to understand which contacts are going to which staff groups. This is a straightforward exercise, where we discuss with the planners and IT the various call flows, and the skilling of agents.
Once we know this, we start looking at the data. We plot a simple table: Staff Groups on one side, with Contact Types on the other. In the table we fill out the number of contacts that went to each staff group.
Now, we wouldn’t expect many surprises here, but we often get a few. They are of the sort that beg the question: “why are so many of my customer service calls ending up with our sales agents?” or “How did those calls end up over there?”
There are many reasons, the most basic being that the best contact center routing scheme can be defeated by a poor staff plan. More on this in a bit.
Capture and Purity Rates
I was schooled by two solid contact center pros from American Express, Tricia Payne and Wayne Parrish, early in the life of our product. When we were developing our first real multiskill simulations, Tricia and Wayne presented to me the concepts of capture rate and purity rate, and their importance to developing solid staff plans.
Both metrics are simple concepts that keep staffing engines and scheduling engines from going off the rails. They help us understand who is answering the calls:
Capture Rate: The percentage of contacts that go where you want them to go. In other words, what percentage of contacts in my table end up in the right box? How many customer service calls are answered by customer service reps?
Purity Rate: What percentage of an agent’s time is spent doing what I most want them to do? This is a similar concept, but from a different perspective: are we wasting our agent’s time?
A well-designed contact routing tree will serve to make sure that, except in overflow scenarios, capture rates and purity rates are high; the calls go where we want them to go and we are taking advantage of our agents’ skills.
Impact of Staffing Ratio on Routing Design
That is nowhere near the whole story. I believe that companies do not really internalize this basic concept: if staffing and scheduling is even a bit off, the best routing plans will unravel. To fine tune the argument: if the ratios of different staff groups aren’t perfect, then overflows will be triggered and capture rates will come down.
The staffing ratios of different groups dictate which contact routing paths are followed by the next set of contacts. If customer service is a tad overstaffed and sales is a tad understaffed for several intervals, it is conceivable the contact routing will allow sales contacts to flow to customer service agents, which will result in lost sales opportunities.
If, when developing staff plans, the planning analyst can draw out the tradeoffs associated with various staffing scenarios, the executives can determine the tradeoff themselves. If the planning model/spreadsheet will allow for the analyst to state, for example, “with 30 bilingual agents the Spanish line’s Capture Rate will be 97%, and the agent’s Purity Rate will be 45%, but with 25 bilingual agents, the Spanish line’s Capture Rate will be 82% and the agent’s Purity Rate will be 75%,” then the executive can draw out the trade-off and make the appropriate decision.
Transition Math Modeling
Like the scorpion on the back of the frog, we must now reveal my true nature: all of this is about having a good solid representation of your contact center operation with the contact center routing and the staff types all mapped out correctly, and the modeling process able to accurately predict both capture rate and purity rate. But that isn’t how most folks do this.
These types of models are hard—especially if your planning technology is a spreadsheet. Among the simplifications many analysts are forced to take in our spreadsheets is to build siloed, single skill models where call routing is handled through call allocations or through fudge factors (we also see this in some workforce management systems). The issue here is that these processes do not recognize the real-world problems of staffing in a multi-skill world.
Good models will allow you to add people to one group and see the resulting change to Capture Rates associated with other, non-primary types of contacts. Meaning, if I add more sales agents, the models will automatically know how many extra customer service calls will be routed to them. By being able to model the true routing structure, executives and analysts can provide resourcing scenarios that reflect the dangers of ignoring capture and purity rates.
You’ve heard me drone on about this before, but this is one of the benefits of true simulation modeling; it can help you understand as you are putting together your staff plans, who will answer your calls.
On a related topic: are you interested in learning more about contact center analytics?
We’ve put together a four-part educational series: Contact Center Forecasting Planning & Analytics from the Beginner to the Advanced. Go to https://swpp.org/sponsor-web-seminars/ for more information and register today!
Ric Kosiba, PhD is a charter member of SWPP and Vice President of Interactive Intelligence’s Decisions Group. He can be reached at Ric.Kosiba@InIn.com or (410) 224-9883.