Dealership websites aren’t doing enough to “wow” today’s car shoppers.

Put simply, they’re “utilizing unsophisticated technologies built on an antiquated notion of how consumers buy cars.” That’s the blunt conclusion from a survey of 1,000 vehicle consumers and an analysis of 100 dealership websites conducted by Affinitiv, an automotive marketing technology company that uses artificial intelligence in some of its tools.

“There was a significant gap between what consumers expect and what’s being delivered on the dealer website,” said Doug Van Sach, Affinitiv’s vice president of strategy and analytics, who works on the company’s data science and machine learning efforts. “I was personally surprised how important the personalization expectation was around financing and finding vehicles within a price range, and the significant difficulty on dealer websites” in doing that.

Affinitiv wanted to better understand how e-commerce companies such as Amazon design a user experience that resonates with consumers, and why that hasn’t materialized at many auto dealerships.

One reason, Van Sach told Automotive News, is that dealerships use a number of website tools that don’t interact well with each other and therefore can’t learn as well as they might about shoppers’ preferences.

To improve, dealerships should evaluate vendors based on their level of personalization, he said, as well as allow machines to predict what shoppers might be interested in based on what they’ve looked at before.

Van Sach said Affinitiv’s recommendations around AI are not based on the products his company offers, which don’t include websites, but rather reflect an opportunity for the industry at large.

Van Sach talked with Staff Reporter Lindsay VanHulle about what surprised him about the gap between website expectations and reality, and how dealerships can improve. Here are edited excerpts.

Q:What surprised you in the survey?

A: The lost opportunity around understanding consumer preferences around vehicles and vehicle features, which you would expect to be the primary focus of the shopping experience on the dealer’s website.

Of the 47 percent [of websites] that recommended vehicles, the only thing they recommended was the last vehicle we looked at.

There was no learning about the customer through the shopping experience. I contrast that to the Amazon experience: They’re constantly learning about my preferences and making recommendations after every item that I look at. And probably the biggest surprise for me was there wasn’t any sort of accumulation of knowledge about the consumer on the website that was used to better enable the experience. And as someone who shops on a lot of different websites, it’s really an expectation I have, and an expectation a lot of consumers have.

How can dealerships improve?

The opportunity for dealers is to retake ownership of the consumer experience. I think they’ve ultimately outsourced the consumer experience to a lot of different vendors with the hope of creating all the right buying tools you need on the website. The issue is, because none of these tools talk to each other, there’s a lot of missed opportunities from a consumer-experience standpoint.

The opportunity for dealers is to reevaluate the vendors they’re working with and try to ultimately work with vendors that offer a more integrated experience.

How can artificial intelligence aid personalization?

AI ultimately allows you to make better decisions on massive amounts of information in a rapid manner that a human can’t typically make. So with AI, for example, the website or the tool within the website can constantly be accumulating this information about the customer — each page they’re looking at, everything they’re clicking on, everything they’re hovering over.

All of that information is helpful and predictive. And the opportunity with AI is to centralize that information and make smarter decisions based on all these different indicators customers are giving you.

The other important aspect of AI is it can’t be focused on learning about a customer based on a very limited interaction. So, for example, if you have a chat-based AI feature and that chat only has access to other chats customers have had and that’s where all the learning comes from, it’s a very small aspect of that customer’s profile, and that AI would be very limited.

To read the original article on Automotive News, click here.