Trade Plates TV: Founder of CarGurus Langley Steinert

Sure. So in my prior life, I co-founded TripAdvisor. And after we sold the company in 2006, I had a big fat non-competition agreement that disallowed me from going back into online travel. So I was looking for new categories to explore. I looked at automotive and real estate in the United States. I wanted to find other categories that had similar characteristics to the online travel space, namely a place where a consumer would be doing a lot of research on a high-purchase, expensive product like a car or a house. In the real estate market, there was a company in the US called Zillow, which had started. I don't know if people in the UK know much about it, but it's a pretty big company now, publicly traded. So I didn't want to go tangle with them. In the automotive space, at least in the United States, there seemed to be quite a void in terms of shopping platforms that provided transparency, providing consumers with information around how much a car is worth or what the dealer's reputation is. So we thought there was a real opportunity to provide this level of transparency to consumers in the United States around automotive shopping. Obviously, in my prior life at TripAdvisor, we prided ourselves in providing transparency around hotel quality with hotel reviews, and that seemed to have worked pretty well. So we launched the company in 2006. We've been in the business about nine years now. CarGurus has about 21 million unique visitors a month, and depending on how you measure traffic, we typically look at average daily shoppers. We have more average daily shoppers than any other platform in the United States.

Yeah, yeah. We're the largest online automotive shopping site in the United States now.

Very good. As I said, we have about 21 million unique visitors a month using the site. We have 16,000 paying dealers that work with us in terms of putting their inventory, or their stock I should say in the UK, on our platform. They've been really excited about both the volume of leads we can send them as well as the close rate, the percentage of leads that turn into sold cars. We're helping our dealer partners sell quite a bit of cars. We decided to expand this platform. We're now live in Canada, the UK, and we're looking at other European markets as well. We've been set up here since December.

Probably trending ahead of where we were in the US when we first started. We have a little over a million unique visitors a month now from basically zero in December. We do already have paying dealers on the platform. I believe we have about 200,000 to 300,000 units of stock on the platform now for consumers to use. Our traffic is growing leaps and bounds, so it's really exciting.

Big market. Two reasons. Number one, it's a big market. And we felt that it had a lot of the same characteristics that we saw in the US when we first started, in terms of there not being a platform that provided this level of transparency for consumers. With all due respect to Auto Trader here in the UK, great company that's obviously doing quite well, but we thought that if you look at Auto Trader or Motors.co.uk, they don't have this transparency around whether a car is a good deal or a fair deal. They don't have dealer reviews, they don't have days on market, number of price drops. These are all the elements that we provide to consumers that, frankly, consumers are really enjoying having in terms of providing that level of transparency when they're doing their automotive shopping.

I don't know. It's a good question. But we're happy that it hasn't come to be yet and that we're the ones bringing it. So we believe we're well equipped to provide that to consumers.

That's a good question. With TripAdvisor and CarGurus, we started with a business model that was the wrong business model. About a year or two into the start of both companies, TripAdvisor we had this crazy idea of having a search engine that would help people find information. It didn't have any concept of user reviews and it wasn't even a consumer site. We were actually selling it to people like Expedia so they could put it on their site. It was more of a B2B play. We quickly learned that was a horrible idea, so we had to change the whole direction of the company to be more focused on consumers. That's when the whole thing took off. Same thing with CarGurus. We actually started with just doing more focused on editorial content, providing user reviews and editorial reviews of cars. Frankly, that didn't really work. About a year and a half into the start of the company, we had to relaunch the whole product to focus more on shopping and providing this price transparency for consumers. It's worked out really well.

No, those were the two. The first iteration didn't work, but the second iteration has done quite well.

I have a whole litany of advice for people around starting companies. Most of one is be flexible with your business plan, because as I just mentioned, often your first iteration doesn't always work. My second piece of advice, which we did both in the case of TripAdvisor and CarGurus, is keep your cost really low, keep your burn rate really low, because chances are you're going to mess something up at the beginning. So it's best to have enough cash in the bank to be able to iterate on something new.

We can spend hours on that, depending how much mathematical specificity you want to get into. I'll talk a little about what we do in the United States, what we've been doing for nine years now, and how it's evolved into the UK. For every car that we upload onto our platform in the US, we upload about 5 million cars a night. Every single car gets analyzed at the make, model, trim, options, geographic market, vehicle condition level. We look at all the cars we've seen over the last, depending on the car. If it's a really popular car, we probably only have to look back 90 days. If it's a rare car, often we have to look back 12 months. We look at all the cars of a similar mileage, make, model, trim, options, and geographic market. We do a linear regression model to look at all the cars that are similar, we do a regression line, and then we calculate how many standard deviations away from the mean. In layman's terms, we look at all cars we've seen over the last 90 to 12 months that are similar to that, and we look at where those cars have been listed. If the car is priced above where all cars similar to it have been priced, we'll typically characterize that as an overpriced deal. If the car is below what we've seen for similar cars, we'll characterize it as a good or great deal. The important implication for dealers, and often when we first came out with this in the United States there was some confusion, concern might be another word, is that we talk to our dealer partners and say, 'Listen, if you have the white Ford Focus in your area that has the white color, tan interior, alloy wheels, chances are you're the only one that has that car that consumer is looking for in your 40-kilometer radius. So in our opinion, you don't need to get into a price war with someone down the road because chances are they don't even have that car.' The only advice we give dealers is: you don't need to get into a price war, but what you do need to do is look carefully at your marketplace and make sure that you're not the highest cost provider and the lowest cost provider. What we do tell people is you should price your cars so that they're fair. When we do our math models, we model the cars on a normally distributed curve. The two places where we tell dealers you don't want to be are the low end of the market or the high end of the market. You want to be right in the middle. So again, the only advice I give dealers is: don't try to be the lowest cost provider, don't be the highest cost provider. What you ought to be doing is using tools that are out in the marketplace. In the United States, there are all sorts of vehicle pricing tools that will help you calculate how to make sure your car is a fairly priced deal. It's really about pricing your cars fair, not necessarily low or high.

Yeah, as I mentioned earlier, when we do the price analysis, we look at the make, for Toyota, the model, the trim, the options on the car, how many miles are on the car, what geographic market it's in because we do local market analysis, we don't just do a market for the whole UK. And then we also do a vehicle condition lookup to see whether the car has been in an accident or not.

As I mentioned earlier, I don't think it's a race to the bottom. Well, I know it's not a race to the bottom because we've lived this for the last nine years in the United States. The only thing I think it's a race to is a race to the mean, if you use a mathematical analogy. If we price cars on a normally distributed curve, the two outliers are low price and high price. So what it is is actually a race to making sure the cars are all priced fairly, not necessarily to make sure that the cars are the lowest cost providers in the space.

My advice to dealers, especially in the United States that we've dealt with so far, and I think it would be the same advice here in the UK, is I would honestly urge people not to price their cars as great deals. You don't even have to price your cars so that they're great. Typically we characterize cars as either priced great, good, fair, high, or overpriced. My only advice to people is you don't need to be the great deal, you probably don't even need to be the good deal. What I would say is you probably ought to think about making sure your car is a fairly priced car, and then let the chips fall where they may. Because as I said earlier, chances are you have the low mileage Ford Fiesta or Ford Focus they're looking for that has a white exterior, tan interior, alloy wheels. You probably have the unique piece of inventory someone's looking for, so you don't need to compete with anyone in your local market. What I do think you need to do is make sure your cars aren't overpriced. It's not a race to the bottom, it's not about being the lowest cost provider, and I would strongly urge people not to price their cars as great or good deals, but probably to use pricing tools like we provide and other companies do to figure out how to price your car fairly in the marketplace.

In our ranking algorithm, how we rank the cars, it's probably worth talking about for two seconds. On our platform versus some of our competitors whose name I won't mention, we rank the cars based on a number of factors. The two most important factors being how is the car priced in the marketplace? Is it overpriced, is it a great deal, is it a fair deal? Obviously the cars that are priced the lowest to the market are priced the highest. However, there's an enormous ranking factor for the dealer's reputation. When we first came out with this platform in the United States, we noticed there were some dealers who were trying to play games. They would price the car really low, it would be a great deal, but when the consumer got there, they would say, 'Oh, it includes $5,000 down' or 'That's only available for military veterans.' So we had to figure out how to police this, because that's not fair to all the other dealers who aren't playing those games. What we did, I took a little page out of my life at TripAdvisor, we said, 'Why don't we have all the consumers rank the dealers after they've interacted with the dealer to tell us, on a scale of 1 to 5, did you have a good experience or not?' In the United States, we now get about 800 dealer reviews a day, and we bake that into our ranking algorithm. So when we rank cars on our platform, it's not just a factor of whether the car is a great deal, but that's multiplied by a factor whether the dealer has a good reputation. It's the quotient of those two numbers. So it's not just sufficient to have a great deal. If you have really bad reviews, your inventory is going to end up on page five of our search results and you'll get no exposure. If you have fairly priced cars but you have a really good reputation, consumers really enjoy working with you, you have a great service department, you treat your consumers with respect, your inventory will end up at the top. So it's really not just about price, it's about price and service, those two factors married together.

What we do is we ask the consumers to rate the dealers two weeks after they've interacted with the dealer. So it's not immediately, we wait a little while until after they've interacted with the dealer. In some cases, they may not buy the car, they may have just gone to look at the car. But we do think it's really important to bake into that concept of the dealer's reputation. As I mentioned earlier, we do it in the United States and I believe we're about to roll it out in the UK as well, as we do vehicle condition lookup. It's really about the car's condition, because there are some dealers out there who really do a good job of reconditioning their cars and they take pride in that. We think that's important to build into the mathematical algorithms, so we do take into consideration the car's condition as well.

There are databases we can look up to see whether the car has been in an accident, whether it's had flood damage, or been part of a rental fleet. Those are factors we do take into consideration.

You're talking about reconditioning. That's certainly a factor. In the United States, and I believe we're looking at it here in the UK as well, there's such a thing as certified pre-owned cars. If they've gone through a certain rigorous manufacturer authorized reconditioning process, we will give the cars a boost for that.

Much like my last life at TripAdvisor, we have an editorial staff and they read every single review that goes on the site. We do reject a whole bunch of them because there are ways that we can detect fraud. If we can tell that a dealer down the road is trying to submit negative reviews about their competitor, we'll reject that. The dealers do all have the ability to contest a review. They can contact us and say, 'Listen, I think that's a bogus review,' and we'll look into it. If we find out that it was false, we'll remove it.

Twofold, and most of this is from our experience in the United States but I think it's unfolding here in the UK as well. There are two things that dealers care about. One is scale: can you deliver a significant amount of lead volume to my store or not? Because if you're number five in their lead flow, if you're not a big player, they're not going to care, they don't have time to deal with that. So you need to provide scale. The other one is close rate, defined as what percentage of the leads that we send to a dealer turned into sold cars. The early returns in the UK seem to be mimicking our experience in the US, which is that we drive an enormous amount of volume to dealers. In the United States, we're typically their number one lead provider in terms of the volume of leads we can send to the dealer, and our close rate is typically around 10%. So 10% of the leads we send to our dealer partners turn into sold cars. We've been told that's typically the highest close rate they've seen of any partner. There's a reason for this, and to some degree I think we're at least today unique both in the US and in the UK. When a consumer comes off our platform looking at a car and they go to a dealership, there's something really unique going on. Number one, the consumer is going to that dealership with third-party validation by us that this is a fair, good, or great deal. They already feel comfortable about the fact that this car has been validated by a third party, CarGurus, as being a fair, good, or great deal. So they're more apt to buy that car versus if they come off a competitor site where there's absolutely no concept of price validation or the dealer's reputation, they're kind of going in frankly blind. Our experience in the US and the early returns in the UK have been similar: we drive an enormous amount of lead flow to the dealers, and the leads do close at a much higher rate than our competitors because they've got that third-party validation built in.

There are two ways that a dealer can be involved on our platform. We have a free service where they can upload their inventory to our site and they can get free leads. The only limitation to that free program is that the leads you get are going to be anonymized. They'll come across as saying 'Elstein369 at CarGurus is interested in your car,' so you won't get the consumer's email address, you won't get their phone number. You only get those if you're a paying dealer. But to be clear, this is free, you're not paying any money for it, and you are getting an anonymized way to converse with the consumer via our servers. You do have to kind of converse through our service, and I think it's limited to 30 days, you have five interactions, so you can communicate with the consumer. If you're on our paid platform, you get the consumer's email address, you get their phone number, you get your URL and your dealership address, and I think there's a map and directions application as well. So you do get more means of communication. But on our free platform, you can still communicate with the consumer, albeit through this anonymized email.

We don't. The way we deal with that is we ask our consumers to rate the dealers. If our consumers go to a dealership and they find that there are these hidden fees, they're going to rate that in the user reviews. That's the best way for us to police that because there's no way that we can go around and validate everyone's pricing. We have to take it at face value what they upload into our feeds.

No, not yet. We haven't seen anyone come up with the same platform we have. The only thing that I think it's changed in the United States is that I think dealers are probably more cognizant about how they price their cars. This goes back to the issue you raised earlier of 'this is a race to the bottom.' When we first came out on the marketplace in the United States, there was a huge outcry of 'this is awful, this is a race to the bottom, this is about price compression, this is going to ruin my business.' That never came to be. The only thing that I think did come to light is there are a number of tools in the United States, one called vAuto, one called First Look, we have a tool that we give our dealers as well, that allows dealers to upload any piece of inventory and then look at how you can price your car and get a sense of how it lays up against the competition. I think there's been a lot more awareness in the US, and I think probably there will be more awareness in the UK about how to price your cars. Again, not about pricing it to be the lowest priced car in the marketplace or the highest, but to price your car so that it's priced fairly to the marketplace. So I think there will be more dealer awareness around pricing your cars fairly.

That's the goal. To anyone who has any skepticism, I would probably give two pieces of data. Number one, when I first started this little company called TripAdvisor, it was myself, Steve Kaufer, and about four developers. It was like six of us above a sub shop. Now it's a $12 billion publicly traded company, I think it's the biggest travel site in the world. So we seem to have done it then. In the US, when I started CarGurus, similarly it was myself and like four or five developers, and no one had ever heard of us. Now, as I said, we're the number one player in the United States. So there are two data points that say we've done it before. I'm fairly confident we'll hopefully do well here in the UK as well.

The early returns, as I mentioned, have been really encouraging. The most important thing we look at is user traffic. The more users we're getting to our platform, the more benefit we can bring to dealers because we can turn those consumers over to dealers to buy cars. We've only been in the marketplace since December, and our traffic has essentially gone from zero to a million unique visitors. That's pretty good in the space of six or seven months. So the early returns are quite encouraging.

Some of the dealers we're already working with, as I mentioned we do have paying dealers already, they're telling us that the lead flow we're sending to them is getting upwards to the same level as Auto Trader already. So we seem to be sending a lot of customers to our dealer partners.

Very positive, yeah.

In the UK, I'm not all that aware of other companies doing a whole lot of what we do. There are a couple of different business models in the US that we've kept an eye on that I don't think are necessarily competitive to us. I think they're actually probably more competitive to dealers. There's one called Vroom, one called Carvana, where this concept of you can essentially buy your car online and it'll get delivered to your house. They're trying to essentially eliminate the whole concept of a dealer, which I personally think is a huge mistake.

I personally think that's a huge mistake. We talked about it earlier before we hopped on here. Two things. Number one, when a car breaks down, you've got to take it somewhere, and you need a dealer to service a car. So I think it's a little shortsighted to think that you can just eliminate dealers because then who's going to service the car? The other thing is that a car is an emotional purchase. I think people, it's not unlike a house, you need to touch it, feel it, test drive it before you're going to make what is often a really big purchase.

New cars, you can just turn up and buy it. You don't have to order it. You can order if you want to get some special features on the car, but for the most part, for new and used, you just show up and buy it. But to get back to this issue of some of these new business models, I think cars, much like real estate, are an emotional purchase. It's a big ticket item. I don't think people, for the most part, are happy to just spend a large sum of money without having touched the car. The analogy I would give is, would you go buy a house without ever having seen it? I'm not sure I would do that, especially if it's used. I would probably want to get the car inspected and make sure it's in decent condition, make sure the dealer has a good reputation, are they going to stand behind this car? There are many aspects to a purchase. It's not just about price, it's about service, it's about reputation. There are many elements to a car purchase, and I think dealers have an important role to play in that for the foreseeable future.

I think the reviews kind of take that into account. The dealer reviews are a huge ranking factor in how we sort the inventory on our platform.

They'll sell a lot of cars with us. That's really what it's all about. In the end, we either help dealers sell cars or we don't. In the United States, as I said, we have 16,000 paying dealers who are really, really happy with our platform and our ability to help them sell cars. The early returns in the UK have been similar in terms of our ability to send a lot of customers to the dealership store, and a large percentage of those customers turning into car purchasers. In the end, we either help them sell cars or we don't.

When we're bigger than Auto Trader.

I'm just kidding. I think a couple of years. We're in it for the long haul. The company as a whole has been profitable for eight straight years, so we're very profitable. There's no pressure for us to turn profitable in the UK anytime soon. We're more than happy to continue investing in this marketplace. We don't have the pressures of having to be profitable in the UK because we've got our US operation to fund the expansion here in the UK. So we'll get there.

No, no. We'll most likely be a public company sometime in the near future. No plans to sell the company anytime soon.

Fingers crossed. Good.