Automotive dealerships are no longer competing with other dealerships to win customers in the digital world. The real competition is coming from sites outside automotive such as Amazon, Netflix, Spotify, and Tinder, which have created new standards of personalization for shoppers. These destinations use machine learning to understand customers’ wants and needs and to provide smarter product recommendations. Customers expect automotive sites to match the experience they’re getting outside the industry.
And automotive has a lot of catching up to do.
Machine Learning Raises the Bar
Machine learning is a form of artificial intelligence in which computers train themselves to make smarter decisions. The self-learning comes from reading vast amounts of data (usually too vast for people to analyze quickly and accurately). With machine learning, a site goes beyond making superficial product recommendations based on your purchasing behavior and the purchasing habits of people like you. Sites learn from your likes and dislikes and from your lifestyle interests and behaviors to make personalized recommendations that you might not have thought of yourself.
Shoppers have had their expectations raised by personalization and will respond favorably when their expectations are met. Not surprisingly, 76 percent of consumers surveyed by Cars.com said they purchase products based on personalized recommendations either half or most of the time.[i]
But this kind of personalization is not the norm in automotive. Too often we’re leading the conversation with customers by talking about the product, not the person who is about to make the second-most expensive purchase of their life. For example, even though seven out of 10 consumers are undecided about make and model when they shop for a new car[ii], nearly all online car search experiences force people to select make or model as the initial step in their journey — instead of first learning about the shopper and offering intelligent suggestions based on those learnings.
No wonder automotive shoppers would rather go to the DMV, clean toilets, have an extended phone conversation with their mother-in-law, or get stuck on jury duty than shop for a car.[iii]
How Automotive Can Catch Up
To catch up, automotive brands need to learn from the leaders outside of the industry. For example:
- Spotify[iv] and Netflix[v] famously apply machine learning to recommend songs and movies based on customers’ preferences matched against the interests of other customers with similar tastes. Spotify sifts through listening data – both yours and the people you follow – to recommend playlists that create true music discovery rather than simply replicate what you’ve been listening to already.
- Dating site Tinder[vi] matches people with other people by first asking members to set up personal profiles and then suggesting matches to them. Tinder refines recommendations based on each person’s feedback.
Machine learning can make an automotive website offer smarter, more personal product recommendations to each shopper based on their browsing behavior and information that shoppers are willing to share about their personal lifestyles (e.g., whether they commute, love music, or live in an urban area). And with machine learning, a site can make recommendations that might not have been obvious to the shopper just like Spotify suggests an artist you might not have heard of but who is close enough to your tastes to interest you. It can also help determine the right salesperson at the dealership to connect you with by matching your persona and behaviors. Or a dealership’s search and retargeting investments can become more personalized as your machine-learning enabled CRM tools comb vast sets of user data.
What the industry considers personal today will likely be superseded by an even better experience as machine learning evolves. The time is now to start learning.
[i] Cars.com consumer survey, 2018.
[ii] Cars.com consumer metrics, 2018.
[iii] Cars.com consumer survey, 2018.
[iv] Forbes, “How Did Spotify Get So Good at Machine Learning?” February 20, 2017.
[v] Netflix Technology Blog, “Using Machine Learning to Improve Streaming Quality at Netflix,” March 22, 2018.
[vi] The Date Mix, “How Does Tinder Work: A Beginner’s Guide,” June 11, 2018.
Author: Matt DiBari
Matt DiBari is director of product management at Cars.com. His role includes managing product teams spanning areas ranging from machine learning to DevOps and big data to improve the visitor experience on Cars.com. He began his career in the trenches of retail, managing sales and service at Best Buy. He developed automotive expertise while assuming executive roles at CDK Global for more than 10 years. There, he was responsible for multiple products in automotive retail and accounting systems working with development teams co-located and remotely located across the globe with product revenues between $30 million and $100 million annually. He joined Cars.com in 2015 as a senior product manager. Matt has a bachelor of the arts in computer science, math, and philosophy from Augustana College and an MS-MIS and MBA from the University of Illinois at Chicago.