On the brink: Tech execs see major transformation coming to auto retail

Asked to identify their biggest challenge, nearly 50% of the 500 dealers polled by Strategy Analytics listed "Identifying the best places to invest marketing spend” as their No. 1 pain point. Listed at No. 2 was “Sourcing and pricing vehicles to maximize profits.”

Solutions to both challenges could be buried in dealers' data.

But it’s more than just digging out consumers’ names and contact information and having an inventory management tool. It’s knowing how to wade through a sea of data stored in customer relationship management (CRM) and dealership management systems, extracting what's salient, overlaying it with local market data, and then using it to drive dealer strategies.

A daunting task, for sure. That might explain why two-thirds of the dealers polled by Strategy Analytics believe artificial intelligence/machine learning will be the most significant development in the next five years.

Brad Perry, founder and chief product officer for DealerSocket, says dealers using the company's CRM in conjunction with its data-mining tool are experiencing an early form of AI. The tool employs predictive algorithms capable of stitching together a customer's story. And based on the buyer's patterns, it tells dealers the likelihood a customer is ready to buy. 

Perry says these predictions will become even more precise in the not-so-distant future, which he believes will bring about a major transformation in the automotive retail industry.

"That's really where things are moving, because machine learning can take in such large quantities of data over such a large period of time that humans can’t possibly process," he says. "And as it's learning, it’s taking into account real-time conditions and changing its predictive algorithms and results.”

That vision also extends to the company’s inventory management solution. Based on those same shopping patterns and ownership habits, the tool will be able to tell dealers which vehicles to source, down to the trim, color, and even options. 

Jeff McCurry, DealerSocket’s vice president of product, says Inventory+ already makes stocking and pricing recommendations. However, he says users are only scratching the surface in terms of the actionable insights they can be glean from the tremendous amount of data the tool collects.

“It’s one thing having data. It’s another being able to develop algorithms around it to get predictive,” he says. “AI and ML can do things we just can't do physically with so many data points. We'll be able to start making projections and predictions of what the future might hold. Then, based upon economic changes or market changes, we can help a dealer know what to stock because we know what people are looking for, and we know the cars they make money on and the ones they don’t.”

Through an algorithm that looks at historical transactional and live market data, DealerSocket’s Inventory+ tool can identify a dealership's core vehicles — or vehicles a store does well with from both a profit and turn standpoint. The software company refers to this as the Ideal Inventory Model. 

Currently in the first phase of a new and evolving initiative, the company expects to make those inventory buy/sell recommendations more robust and more predictive. Key to that vision is the flow of data.

For years, dealers have had access to their own current and historical sales data along with tools to give insights about their local markets. But the information often resides on different platforms and is in neither a usable form nor readily available. It’s the reason for DealerSocket’s platform play. The software maker believes data living within a single digital ecosystem — or single suite of integrated products — is the key to advancing the predictiveness of its solutions.
  
Merritt Critcher, senior product manager for DealerSocket’s Inventory+ line, uses the used-car manager evaluating vehicles in the auction lane or appraising a customer’s trade as an example. Being able to triangulate online shopping insights, customer data, and how quickly a particular make, model, color, and trim level of a vehicle turns at his or her dealership is critical.

“They need to make decisions within seconds, as opposed to going back to spreadsheets,” he says. "We're bringing all of this information and tying it together in a meaningful way so they can make those decisions."

Recent research backs DealerSocket's current strategy. According to the Strategy Analytics report, nearly 50% of dealers listed "Teaching me how to make smarter decisions" as a top value proposition when it comes to tech-vendor relationships. The research firm notes that the finding means dealers want help understanding the data. 

In a separate study commissioned by Urban Science, 81% of the 200 dealers polled indicated a desire for a simpler way to review performance metrics. Seventy-eight percent of respondents said they wish they had someone to consult with on how best to utilize metrics. The study also revealed that dealers spend, on average, 3.7 hours pulling data from 4.5 different tools and dashboards.

Then there’s the consumer experience. The good news for dealers is seven out of 10 consumers polled in the Urban Science study said they would not buy a vehicle without the dealership involved. Eight-one percent of respondents said they trust the information they receive from a franchised dealership. Also revealed is that the average car buyer submits three different leads. 

Randy Berlin, the firm's global director of dealer services, says the latter finding points to consumers crosschecking pricing. That tells him that the notion that car buyers want to do more online remains true. But they still want the dealership involved, he says.

"Think about autonomous vehicles. I don't know that the majority of people, or even a fraction of people, are saying, ‘Yes, I'm all-in on that,'" he says. "But from autonomous cars, we've learned a lot. It's given us safety features like adaptive cruise control, automatic stopping, or lane-departure alerts.

"I think digital retailing is synonymous with that," he adds. "I can do more of that transaction online, and it's more than just submitting a credit application. So now I can get a long way to expedite the process, and then still go to the store to complete the transaction because I want to see the car."

That transition has been a major focus for DealerSocket. Perry notes that consumers are doing their research and completing parts of the car-buying process online with the expectation that the data will carry over to their in-dealership experience. It's not a transparent experience for them when that doesn't happen, he says.

"We have a website platform and a digital retailing platform that is 100% connected with our customer relationship management software in the store,” Perry notes. “Everything they do online is the same when they get into the store.

“We’ve closed that gap significantly from a technology perspective, and now we are working with dealers to make sure the process is the right experience for the customer.”

Perry believes dealers are especially receptive to machine learning because of the time and money it can save them. It can save time because the technology automatically tracks customer online search habits and actions, equity positions in their vehicles, and when they are most likely in the market for a vehicle. 

DealerSocket’s RevenueRadar data-mining tool does that today, but Perry says the ability to automate follow-up communications is what's coming. The system will know what message to relay and when — such as a lease pull-ahead program for someone nearing the end of a lease — then help them find the right vehicle. This will further save on labor costs because computers and machine learning are doing the work and not people.

“Today, a salesperson is making five to 10 phone calls over 30 days trying to reengage that customer,” Perry says. “With artificial intelligence, we can engage the customer without having a human involved, so the quantity of work that they do significantly improves. And what that does is improve profitability.”

The executive admits that, even though many dealers are receptive to the potential benefits of machine learning and artificial intelligence, they still hesitate to use it. Dealers are trying to figure out how it fits into their processes and the experience they want to deliver to their customers.

Perry sayshe expects, especially in the beginning, that dealers will want to know how they can trust the recommendations made by the tools. That's why presenting the data supporting those recommendations is critical.

Eventually, as dealers become more familiar and comfortable with the promise of AI and ML, they'll use it to reinforce or challenge the decisions they make. Other dealers will embrace those recommendations completely. But it will take time, Perry says.

“People, processes and technology. They all have to evolve together, and our industry is at the beginning of this,” he says. "We have some of the tools built and some of the technology being used, but we’ve got a long way to go.

“What we see today is going to be so much better, so much more accurate and so much more ingrained in processes over the next three to five years.”

For more information, visit DealerSocket.com or follow us on Twitter, LinkedIn, and Facebook