Eric Brown of LotLinx on how AI and machine learning can allow dealers to add the human touch to sales.
AutoSuccess: Before we begin, Eric, let’s define some terms. What is artificial intelligence and machine learning when it comes to automotive marketing?
Eric Brown: Very good question. Artificial intelligence, which we certainly hear a lot about these days, simply put is a machine or a computer mimicking human behavior. It’s using data and “if/then” statements to mimic human behavior. Machine learning is essentially one step further, in that is not only is it mimicking the human behavior, but it’s also learning through that process. So, if there are better outcomes — better paths to solutions — it’s learning those and then making decisions and evolving its behavior beyond merely mimicking something it’s been taught or observed.
AS: You’ve spoken about artificial intelligence being used with marketing, such as with vehicle-specific campaigns. How does this effect the conversion rate and why does this seem to be more effective?
EB: Artificial intelligence allows us, in a marketing context, to take much more complex problems and simplify them and drive toward better outcomes. The machine learning aspect of it is learning from the decisions we make literally every millisecond with a computer, and then evolving to better and better outcomes. So, if you think about sort of the traditional media landscape, it’s a bunch of channels. It’s a bunch of audiences. Of course, the question has always been, “How do I find the right audience and the right people within that audience that are interested in my product or service?”
The Internet came along, and all those machines that powered it starting throwing off a lot of data. As consumers carried those machines around in their pockets, they threw off even more data. So, all of that data now allows artificial intelligence — more specifically, machine learning — to analyze that data and, with a high degree specificity, determine which consumers in those audiences are most likely to respond to my ad message and buy my product. The analogy I often use is if you think about those audiences as a haystack and the individuals within that audience who have the most interest in my product or service as the needles, today we don’t have to buy the haystack anymore. We know where the needles are because we have the math, we have the machines, we have the technology to find those needles and drive out a tremendous amount of waste to get to those needles, to find those consumers who are going to buy the product.
AS: Eric, how can AI assistance, the machine learning and chat bots help bring in customers and keep them engaged? What does that process actually look like?
EB: I don’t know if there’s one process; there are lots of processes. Going back to that early definition about mimicking human behavior, we’ve always known as salespeople, as marketers, that the better we understand our customers, the better we engage them, the more compelling our message, the better our conversations align with their interests, the more likely those consumers are to engage our business and the folks within them. So, what AI does is allow us to use that at scale, and, again, with the specificity that as individual humans we just simply can’t keep up with. As we learn from those interactions with the consumers, the AI just gets more and more proficient at engaging consumers in compelling ways and getting.
We have drive-thru windows and microwaves and all kinds of convinces. DVRs fast forward because we want our information faster and faster. AI allows us to get to what that consumer needs much more quickly and much more precisely. It allows us to engage them in ways we couldn’t previously.
AS: Can you describe how this technology can aid dealerships with search engine marketing costs and effectiveness?
EB: Again, it’s going to be a sort of reoccurring theme, in that there are billions of searches done on those platforms. All of that, of course, creates data. Tons of intelligence that, quite frankly, is sort of thing “day traders” of SEM could never keep up with. So, by pulling all that data in, artificial intelligence can now sift through billions, if not trillions, of data elements — terabytes of data — and determine what keyword yields the best outcome at the least cost with the greatest amount of return on investment. What time of day? What vehicles and what markets? All of those granular, independent characteristics that drive performance can be found and harvested to make better decisions about how I spend the dealer’s SEM dollars, and how and whom do I engage with those dollars.
AS: In most cases, that’s the million-dollar question, Eric. In what way can this technology help dealers increase gross margin profits?
EB: Well, it’s pretty simple: reduce the cost of sale. Earlier I was talking about that haystack. You think about traditionally we’ve been buying media in a sort of passive environment in that, “OK, I’ve got my cars on my lot. I’m not going to move from over here to this third-party site and put them on that lot. Maybe I’ll buy a bunch of keywords and throw a lot of stuff against the wall, hoping something sticks.” Of course, it’s always the minority percentage that sticks. We’re spending a lot of money to find those needles. By being able to throw out all of that waste, we can dramatically reduce the cost to market that car. In fact, typically we’re seeing AI drive down that cost of selling by 80 to 90 percent over what traditionally NADA, for example, has reported average cost of sale.
Here’s another way of looking at it — if you think about it, every car only has one steering wheel. So, I only need one consumer to buy that car. Every consumer I have to market to after that one is a waste. The artificial intelligence inverts that and says, “OK, let me first look for the one. If I get that right, yay, I move on to selling the next car. If I don’t get it right, I’ll move on to number two, number three and number four until I get it right and I sell the car. I’m learning with each interaction with each of those consumers, again, opposed to the more-traditional model of, ” What’s my cost per thousand? I’m buying 10,000 people.” I know out of there, there’s got to be 10 or 15 people who want to buy my car. Well, that’s buying a lot of waste. Again, I’ve got one steering wheel in that car. Let’s focus on one person and move to the next. That, of course, naturally mitigates my waste in marketing that car.
AS: How can AI aid in maximizing incentive programs to help both dealerships and customers?
EB: Very much in the same way. Again, incentives can be very complex and so, by using artificial intelligence, we can discern which incentives are going to be the most compelling to which consumers. We can look at a dealer’s inventory and we can run artificial intelligence against that and determine which of these cars, based upon the demand we see in the marketplace, best align and how much investment should we make in marketing that car. If you think about it, incentives exist because the car is not seeing the demand it needs to sell at the rate it needs to. The whole point of a high incentive is to level-set the value proposition with the consumer. Well, what the AI does is it goes out and it says, “OK, I know those shoppers exist. They may not exist at the scale that I hope for, but I know where they are. I know which cars have those incentives.” It then pulls those two things together, and essentially builds a strategy and the related tactics for the dealer with the push of a button.
AS: Now, Eric, many, many years ago we were talking about the hesitation to use what was then the new technology called the Internet. What advice would you give to dealers or GMs who want to take advantage of this new technology but might be intimidated or simply not know where to start?
EB: I’m going to sound a bit like a nerd, but I would say artificial intelligence really is about freedom because it takes what are typically today very complex decisions that require a lot of data, a lot of research and a lot of analysis, and it allows us to make those decisions in seconds. Quite frankly, it makes them with a degree of accuracy much greater than we as humans could ever do because of that complexity. So, by freedom, I mean, it’s freeing up time. It’s empowering me as the dealer to take my expertise and free up time and money and resources so I can take my expertise as an individual, as a human, as a car dealer, and apply that where I can get the most return from my hours and time, and takes away a lot of the effort and the work and decisioning that, quite frankly, is often is too complex to get right.
AS: Finally, what one idea would you like to leave with us today?
EB: If you think about the auto industry, it is an industry built on selling machines that were created to replace horses, a biological entity. I know there’s a lot of fear and consternation out there about artificial intelligence and replacing humans and all those kinds of things, but we actually work in an industry today where that is exactly what happened. Those machines brought to our society and to our community as an ability to travel and to accomplish things we never could accomplish before. So, I see artificial intelligence and machine learning today as much the same thing. If you think about it, machine learning has been on the assembly line for a long time. It’s a natural evolution that this would roll into sales and marketing. It’s going to provide us with competency and capability and new opportunities that, as car dealers, as people who sell cars have never had before. So, I would encourage folks to embrace it and to seize upon the opportunities that it will bring to each and every one of us.
Eric Brown is the president of LotLinx.
Click here to view more solutions from Eric Brown and LotLinx.