Loyalty Learnings #18: Integrating Survey Data With Internal Metrics

Edited Transcript:

A really valuable tool for most companies is to integrate loyalty analysis with metrics that firms capture internally about the customer experience. 

Let me give you an example.  We had a client who had product changes that the customer did online.  So, I as a customer would get online to change the configuration of my product.  Self-service.  In the loyalty analysis that we did, we found a few things. 

One, there’s a pretty high level of vulnerability.  That’s consistent with the fact that this firm had a fairly high level of quits, fairly high level of defection.  Two, that vulnerability was driven by customer service.  If they gave low evaluations in customer service, they were likely to be vulnerable and to quit.  And specifically, the customer service attribute “Successfully Resolving My Issues.” 

We looked in their database, they didn’t have very much. But they did have a time stamp for every call by customer into customer service and that really gave us a ton of information.  Our modeling of that gave us the insight, one, if the customer called into customer service three or more times in a given thirty days, their probability of being vulnerable and defecting was 88%, or .88.  That’s great information. 

Second, we could model that.  We could then take that result, that model, and project it onto every customer in the database.  Even if we hadn’t surveyed them, we knew their vulnerability.  We stack-ranked the customer base by economic value and probability of vulnerability and defection. 

The company came back with an immediate tiered strategy of intervention.  Those with a high economic value and a higher quit probability got executive calls, those with more moderate economic value and high probability of quit got sales person calls, and those with lower levels of economic value got emails.  They all got the same message: “We understand you had a bad experience.  We want to intervene – that shouldn’t have happened.  That’s not who we are and you shouldn’t have it again.”  Great, let’s stem the tide of defection. 

The more important question became then why were they calling in?  Why do we see numerous calls into customer service creating vulnerability?  It was because the customer service response wasn’t addressing their issue.  They were going to the website, they were asked to perform a relatively complex transaction and they didn’t understand it.  It wasn’t well-written. 

They called customer service, customer service didn’t understand it either.  So they had to call back and call back until the customer threw up their hands and said “No more – I’m leaving!” 

So the second thing they did was train customer service to answer the questions better.  Which they did. 

And the third thing, well, let’s go back to root cause, let’s change the design of the product and change its description on the website so that it obviates the needs to call customer service in the first place.  So a three-tier strategy. 

One, let’s intervene before we lose any more customers and we know who they are.  Two, let’s change the scripting for customer service so we do a better job.  Three, let’s change our product and website so they don’t have to call customer service in the first place. 

By the way, this wonderful model we built was worthless within 6-7 months because the entire environment in which it was based changed.  That’s good news!  The value was that they understood how to take that information and activate it once they got it.  Once we could integrate this information, build the value for them, they could take it and run with it.