Skip navigation
For service customers, $200 is “the sweet spot” for a vehicle repair order, but a $600 bill can turn off price-sensitive customers, says Experian.
For service customers, $200 is “the sweet spot” for a vehicle repair order, but a $600 bill can turn off price-sensitive customers, says Experian.

Car Dealerships Use Data to Determine Service Customer Worthiness

Predictive analytics play a growing role in deciding who to woo – and who not to.

LOS ANGELES – When it comes to marketing, certain dealership service customers aren’t worth the effort to get their business, says Brad Smith, vice president-product management at data tracker Experian.

Some service customers are highly loyal, some are highly likely to defect, and others are in the middle, he says. “You’ve got to be willing to let the highly-likely-to-defect go. There is no point on spending a lot of money on them.”

The middle group is worth going after with promotions and specials to keep as customers lest they bolt to a competitor, such as an independent shop or a national car-care chain.

“Make the right offer at the right time to them,” Smith advises dealers. “It doesn’t always have to be an incentive. The possible defectors are the ones to focus on.”

Predictive analytics are playing a growing role in deciding who to woo – and who not to. In the latter case, it is the highly loyal and highly disloyal. It’s a topic of discussions at Thought Leadership Summits’ automotive conferences on analytics and customer experiences here.

It’s a matter of using data to know what to offer, to whom and when, Eric Flow tells WardsAuto in describing how to get, keep and engage with service-department customers.

He is a board member of AmplifyRPM, a joint venture that aims to “bring science to customer engagement.” Eight dealership groups, including Holman Automotive and Serra Automotive, are part of it.        

Traditional customer information centers on service history. That includes the last service date, last known vehicle mileage and services rendered.

Today’s predictive analytics draw from extended customer data such as geographic distance from dealer, age, income, gender, family type, ethnicity, previous satisfaction and behavioral trends, says Flow, a member of a dealer family.

Drawn from various sources, the crunched data helps forecast consumer behavior.

AmplifyRPM’s model focuses on recovering lost or missing service customers, rather than seeking new ones. “It’s not spray and pray,” Flow says during a conference presentation.

“We can predict within 90% accuracy if they’ll come back,” he says “We need to market to those people in various ways, including direct mail. You might throw extra money (through specials and discounts) at some. But we look at the big middle, predict who will come back and start making offers to them.”

The model uses algorithms, features an unprecedented level of personalization and creates a holistic view of the customer.

Much of customer retention is a balancing act, says Experian’s Smith. He recalls an initiative of about six years ago when an automaker put incentives on parts. That drew in service customers, but some dealerships too aggressively upsold in the service lanes. “You started seeing defections,” Smith says.

Dealerships need to keep service customers close but not too close, he says. If customers with chronic car problems are returning to the dealership every 30 days or so, “that’s a bad thing,” he says.

Then there are the ones who don’t patronize the service department enough. “If they are only coming back every 360 days, that’s a bad thing,” Smith says. “You have to figure out how to engage with them to bring them in more often.”

For customers, $200 is “the sweet spot” for a repair order, he says, citing Experian data. A $600 repair can turn off price-sensitive customers.

TAGS: Fixed Ops
Hide comments


  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.