Q&A: The Art and Science of Predictive Analytics

Predictive analytics is gaining a reputation as a disruptive new dealership technology. But experienced sales professionals are already adept at determining when a given customer will be ready to buy or lease.

To learn how dealers are using predictive analytics to leverage big data and enhance those skills, Auto Dealer Today met with Marco Schnabl, who spent 10 years as a Mercedes-Benz salesperson in his native Germany and at Mercedes-Benz of Manhattan before co-founding automotiveMastermind with fellow sales veteran Johannes Gnauck in 2012.

Marco, what should the term “big data” mean to a dealer? What are the sources and why does it matter? The original definition is a compilation of data so large, complex, and vast that traditional data-processing software can’t handle it. If I’m a dealer, I’ve sold thousands of customers in the last 10 years. Therefore, in my DMS, I have multimillions of datasets.

Your DMS data alone wouldn’t necessarily be defined as big data. You could probably still handle it with a fast computer and Excel spreadsheet. However, once you say, “For all those records, I want the VIN number, the credit data, accident data, their social media and email,” you have stepped up to a different level, where traditional data-processing software won’t be useful.

When we started at Mercedes-Benz of Manhattan, the economic environment was quickly turning. We were in a full-blown crisis. People weren’t walking in the doors anymore. One idea we had was to calculate how many miles a customer would have on their car based on their service history. Then we could call and sell them an extended warranty. And we thought, if we can do all this by hand, imagine what we could do with even more information and better technology.

And the better technology is predictive analytics. Similar to big data, the definition of “predictive analytics” deserves clarification. In the dealership realm, a highly trained, highly experienced sales consultant does their own predictive analytics. If you come to my dealership and I ask how many kids you have, do you have a dog, how far do you drive each day, I’m predicting what kind of car you might like to buy.

CRM systems are for the most part data storage systems. The CRM won’t tell you which customer to contact, when, and with what message. We have had good technology that helped with outreach. But until now, we haven’t had technology that provides information to a sales consultant that anticipates individual consumer needs.

Does predictive analytics produce standard-format reports or are different dealers asking for different things? Great question, and the answer is yes and no. At its core, predictive analytics provides the answers to two very important questions: Who’s ready to buy, and why? The why is just as important as the who. Those are sales drivers. They can be utilized in the conversation.

Dealerships tend to operate very differently. A single-store owner may be interested in getting each of their salespeople to perform at their best. The owner of a big group may wonder how predictive analytics will benefit their centralized BDC. Our field team tries to truly understand what matters, and they differentiate the process integration based on the dealer’s needs. But the scoring and ranking will be similar.

Have the applications changed? In the beginning, we were very much focused on retention, but we never intended to stop there. Then we extended into other areas. How can we convert your service customers to sales customers? How do we conquest customers who are driving a competitor’s car?

This presents a new degree of difficulty. That’s what we’re doing now: Remove the “black box” of conquest and make it transparent. With a holistic market view, we can help the dealer identify the next best customer — whether they are a loyalty, true conquest, or service customer — communicate with them, and help close the deal. We want customers to truly understand what the offer is and communicate back what they like or don’t like about it.