Learning customer personalities to better manage call centers

Fast Company had an interesting article on the technology being pushed by firm called eLoyalty to improve call center operations (How a Personality Test Designed to Pick Astronauts is Taking the Pain Out of Customer Support, Dec 1). The roots of this approach date to a methodology developed by a clinical psychologist to categorize personality types.

The methodology, called the Process Communication Model, was created in the 1970s by a clinical psychologist named Taibi Kahler. He divided people into six main personality types, each of which has a different communication style and each of which has different stress triggers. If you know the personality type of the person you’re speaking with, Kahler explained, you can modify your own communication style to work more effectively with them, prevent misunderstandings, and avoid inadvertently pushing the other person’s buttons.

Apparently, this approach has been used by NASA to determine who’s got the right stuff to for space missions. But how does this apply to call centers?

In call centers, eLoyalty’s system uses the PCM framework to compile a personality profile of each caller from the moment they first contact the center. The system, which is automated, analyzes the caller’s language patterns and other behavioral cues to identify their personality type. (A team of 250 linguists, behavioral scientists, and statisticians have compiled a massive set of linguistic libraries and behavioral algorithms to parse callers’ every word and mode of expression.)

Each time the customer calls back, the system uses the existing profile to steer them to a customer service representative who’s the best match for their personality type, and it continues to analyze their subsequent conversations to deepen and enrich their profile.

eLoyalty, which has clients in the banking, health care, and insurance industries, among others, is the only organization in the call center industry licensed to use PCM. Typical call center quality assurance programs train reps in specific issues and rely on supervisors listening in to a small percentage of calls, and providing coaching to individual reps. The automated eLoyalty system not only allows a larger proportion of calls to be analyzed, but it moves coaching out of the realm of intuition and grounds it in evidence about how to communicate effectively.

So the key for eLoyalty is the capability to evaluate a high volume of calls automatically. This allows for both tracking a large number of customers as well as for better training and coaching of agents. Coaching sessions can now focus on supposedly quantified guidelines. The article reports that several eLoyalty clients have seen impressive gains in customer service measures.

I have, however, a few questions on this. The first is how effective the classification of customers can be if customers call relatively rarely. When you are working with astronauts you have months and probably hundreds of interactions to evaluate their personality. I don’t know the last time I called by home insurance provider. How quickly would they be able to pin me down when I call?  It may be that all the action is on a relatively limited number of callers who call frequently. But does a firm need to classify squeaky wheels in order to be able to manage their calls better?

My second question is how this interacts with call routing. If the best service requires matching specific personality types with specific agents, that adds a layer of complexity to how calls are sent out to agents. This could create a tension between good service and short waits. After all, “all other things equal, pooled groups are more efficient than specialized groups” is one the Six Immutable Laws in Incoming Call Centers. The challenge here is that waits are easy to measure while superior service is a little more nebulous. Further, even if eLoyalty can evaluate service quality it may not be able evaluate how customers trade off service and waiting.