EverQuote, Inc. (EVER) Earnings Call Transcript & Summary

Good afternoon, everyone. John Wagner, CFO, as mentioned. I'll go through some slides, then we'll have time for Q&A. Slides will just give you an overview of EverQuote if you don't -- if you're not familiar with our story. Normal disclosure. So EverQuote is an insurance marketplace. We help consumers who are in-market shopping for insurance, connect with providers in order to get multiple quotes on insurance. Today, we service -- autos is our largest vertical and where we had our start. We also have home and life as well as health and commercial that we launched just this past year. Our mission ultimately is to make insurance simpler, more affordable and more personalized for the consumer. And we do this really running a marketplace that's based on data and technology in order to accomplish this. So a little bit about the market that we serve. Suffice to say, it's a large and expanding market. There's a reason why you know the taglines, and the mascots, and the jingles, of a lot of the providers, the insurance carriers, and that's because there's a massive amount of spend within the insurance market. $9.3 billion, as shown here, is the advertising dollars spent by the insurers in the markets that we cover within insurance. The vast majority of that is still spent offline, which can be somewhat surprising when you consider that most consumers now are coming online to shop for insurance. 70% of consumers report going online to shop for insurance. But yet, 70% of the actual ad dollars are still spent offline. Only about 30% of that $9.3 billion is spent online. But those dollars are moving toward the online category, and they're moving at about a 16% annual clip. So online advertising, specifically within auto, as that's our largest vertical and where I have the data, is growing about 16% per year. And so for us, that provides a very nice tailwind for our TAM and expanding market for us. If you, on top of that, layer in really what is the larger market, which is including all of the distribution and marketing spend within insurance and the verticals that we serve, you're over $120 billion. Much of that is in commission and distribution, but much of that is also accessible to us through agent commissions that are then recycled back into marketing dollars and also a larger trend, which is some of those dollars, again, moving online to online categories overall. So a very large market that is moving online, creates basically a tailwind for us. And then we've also seen kind of the change within the products and the workflow to support the insurance moving online. So products becoming more friendly to digital models as well as the workflow, being able to get consumers to quotes and to products more quickly and more efficiently on an Internet model. So our business model has really benefit, like any 2-sided marketplace, for both the providers as well as the consumers. For the consumers. The value proposition comes down to shopping for insurance, all right? We provide a single starting point for shopping for insurance. Consumers come to us to get quotes, they enter their information 1 time. In most cases, we can leverage that information to get them to a set of quotes, often only 1 quick click away or in many cases, with their information being used by the provider to get them to a quote more efficiently than doing it on their own. In addition, something that consumers don't always realize is, there is actually a set of carriers that are the right fit for them. Not all carriers service all consumers in the same way. Carriers make a living out of specializing and serving consumers in certain groups more efficiently than others. And they express that often in their pricing. So we will help align those consumers not only to multiple quotes, but also to the right set of quotes that's most likely to save them money. In terms of saving them money. Our studies, our survey data shows that we save, on average, $610 for a consumer who purchased insurance through EverQuote. So a strong value proposition for the consumer. For the carriers and the agents, the providers that we work with on the platform, the value for them is really about a consistent stream of high-intent consumers. It is an efficient way to acquire consumers. And specifically for the carriers, it is really that benefit of targeting those consumers versus their own dollars that they spend online, and of course, all of these carriers that we work with, have the capacity and indeed, do spend money online acquiring consumers. We give those carriers the ability to segment, filter and target the consumers that they truly want in their book of business and then be able to establish the pricing on those consumers that meets their ROI targets. We always monetize in this model on the referral. So we get paid when we refer the consumer to the quote, when the consumer generally gets the quote. But in all cases, the carriers backing down to a cost per sale in terms of how they decide what they're going to bid for a referral. So for carriers, it is a consistent and efficient acquisition source for new consumers. So a little bit of a drill down on the consumer journey in the marketplace, right? We are primarily paid on the initial arrival. Most of our consumers find us online. We advertise across pretty much all segments online. Most of that is online. We dipped our toe recently into some offline traffic sources. But for the most part, consumers find us online with the value proposition of shopping for insurance. They then complete -- generally complete a form online to become a quote request. So we see a conversion from the number of arrivals down to an actual form, which generally is a quote request. Quote request is the metric that we disclose publicly in terms of measuring our traffic to the marketplace. And quote request is -- effectively, because it's a consumer who has asked for quotes, it's a consistent -- it's consistent in terms of the level of intent of that consumer. They've asked for quotes, they're looking to shop insurance. From there, we take that quote request and that has the data that is necessary for us to be able to match that consumer with the right set of providers. From there, the consumer gets quotes from insurance providers. Again, some as close as 1 click away. In some cases, they're seeing their information go over to the provider to get -- make it so that, that provider can quickly provide the consumer with a quote on insurance. So a little more on the path, right? So consumers, again, come to us almost substantially all of our arrivals are online arrivals. We have over 300-plus traffic sources right across all the major categories. And in 2019, we've seen this kind of traffic mix go from some of our -- to some of our more performant traffic sources. So we've seen an improvement in conversion in 2019. Some of which is based on the mix. And this next bullet really points out that some of that conversion increase is also based on advances that we've made in the marketplace itself. So consumers come to us as arrivals, we see varying levels of intent on an arrival. Ultimately, consumers become a quote request at a rate that can differ between as low as 1% on low-intent consumer traffic. As high as 86% on higher partnership-type traffic or something like search, which is higher intent. At that point, a quote request is a consumer that we have the ability to monetize, right? At that point, we can offer them quotes through a number of providers. In general, we'll set them up with a number of choices. On average, we monetize about twice. So we see 2 referrals -- a little less than 2 referrals per quote request on average. In 2019 through Q3 of this past year, we saw our quote request grow 49% overall. So a little bit about how our carriers work with us. Our carriers work with us on a platform. They can -- within that platform, they can segment and target based on their ROI and based on the type of consumer that they would like to acquire. So this means that they can alter their bid based on -- not only on geography, but on profile. They can bid more for a consumer that has multi vehicles, more for a consumer that has higher credit score, and they can effectively bid that consumer to not only to the type of consumer they want to receive, but also to their ROI requirements. In some cases, we'll always coach the carrier in terms of how we best -- how they best use the platform in order to acquire consumers. In some cases, some of our carriers actually work with our smart campaigns product, where effectively, they give us their cost per acquisition target, their cost per sale target, and we actually handle the bidding within the marketplace for them in order to maximize consumers that meet their profile and at their cost per acquisition. So for consumer -- for carriers, once again, it's really about targeting the type of consumers that they truly want their book of business. All carriers are going to do broad-based marketing. Broad-based marketing is going to have some level of inefficiency because they're always going to bring some consumers in that they would rather not quote insurance to because that particular consumer profile might be outside of their sweet spot, and they express that sweet spot, often through pricing. We allow them to target just on the consumer that they are seeking to add and at the economics that they demand. So all of the marketplace is powered by data and tech. So we have what we believe is the largest multi-carrier data set of consumer quote and buying rates. And we use that data set and the technology throughout the whole process. So on the back end, we use that data set, and some of that data is based on how consumers actually perform with providers. So we can actually use that data to best align the consumers with that set of quotes that's the right set for them. So we know if you, as a consumer, have a certain profile, we know historically, where profiles of your sort have tended to bind insurance and therefore, where they're -- where they'll most likely get the right pricing. And so we'll make sure that you, as a consumer, get a right set of quotes. On the front end, we can use that same information to maximize and optimize on how we acquire traffic, which means we will -- although we always get paid on the referral, we'll look to optimize the marketplace downstream for carrier quote and bind rate, effectively saying that although we get paid on the referral, we know the carrier is going to back into a cost per sale for the basis of their referral pricing and bidding, and we actually optimize on that close rate on the front end. So effectively, we can take 2 websites that look very similar. And based on the downstream bind rate, we can change our bids to make sure that we're paying more for a consumer that has a high intent and likelihood of buying insurance, and less for somebody that is less intentful even if they travel through our own marketplace at the same rate. We're using that disposition data from the carrier in order to optimize on how we acquire traffic, maximize on the ability to acquire traffic that is intentful and in-market searching for insurance. So we think this creates a pretty powerful competitive moat. We have over 1 billion actual data fields that we've collected over the years. We've spent about $0.5 billion in advertising in order to collect this volume of data. We advertise across over 300 sources. We have the historic knowledge of how those sources convert going through our marketplace. And we've worked with about 50 million consumers over that period. And once again, we know the profile, where consumers are most likely to find the right pricing for them. And we're able to use that, again, throughout the whole marketplace. So a little bit about how we think about growth. We think about growth upon -- around several different levers. The first is doing more of what we're doing today, which is growing and attracting more consumers, adding more volume to the marketplace, expanding our relationship with carriers and agents, either adding new agents, adding new carriers or just as importantly, expanding their spend with us on the marketplace. And then expanding consumer and carrier engagement. One of the ways we do that is through integrations. When we hand that consumer off to get their quotes, making sure that we optimize on how much of the data that we've collected, the carrier is actually using in quoting that consumer. So it increased the likelihood that, that consumer can come through our marketplace and click off to see a quote and land directly on their quote, if nothing else, should at least see their data come and pre-fill the workflow from that carrier to reduce the friction in terms of getting that consumer to a quote and making shopping for insurance easier. And then ultimately, launching new verticals. This past year, we launched commercial and health. Previously, last -- about 2 years ago, we launched home and life as well. So expanding from being a single vertical provider to being truly a multi-vertical marketplace. So talking about growth. We have a strong track record of growth, roughly 31% CAGR over the last 5 years or so. This year, midpoint of our guidance that we expressed last quarter when we released Q3, was $243 million. So growing at a very good clip. And even stronger when you consider the shift from indirect revenue to direct revenue. Today, 95% of our revenue comes from our direct carriers and agents that operate on our platform. That was not always the case. As we started the business, we initially sold many referrals through aggregators and wholesale channels. As we've gotten bigger, added scale, we were able to build the relationships with the carriers. Over time, we've been able to expand that to the point where now today, the vast majority of our revenue comes from those carriers working directly with us. The importance of that is -- really is in this next bullet, that now, 98% of our revenue in Q3 actually came from carriers and agents that had been on our platform for at least a year. So in any given period now, we know where the vast majority of our revenue is coming from. It's coming from the carriers and agents that we're already working with, giving us better visibility into our revenue and actually giving our revenue a reoccurring-like quality. Variable marketing margin. This is really our north star of the business. This is how the business is managed. We manage the business for variable marketing dollars. Variable marketing is defined as revenue less advertising costs. Historically, we've seen that grow. Even as we've run the business for variable marketing dollars, which means we will spend on new customer acquisition, new consumer acquisition, right up to the point of marginal efficiency, where at least we generate more revenue from the next consumer arrival than we spend in advertising. So we manage for the dollars, but we've seen historically that this is also equated into an increasing margin. Today, we operate around a 30% variable marketing margin. And ultimately, we think this will continue to expand. Long term, we think this is a margin that can be 40% for us and part of how we expand our adjusted EBITDA profile into the mid-20s in the long-term is the expansion of VMM as we scale as well as leverage that we'll see elsewhere in the operating categories of the business. So it's something we've been very excited about is the launch of new verticals. Originally, home and life, which we were excited about because it really proved that we could transition from a single vertical company to a multi-vertical company, and do so more efficiently than we did becoming -- going into autos, by leveraging the existing technology to approach many of the same teams. This past year, we launched the commercial and health groups. And so ultimately, you can see that part of the importance of the new verticals is, it's really where the increased growth in the model is coming from. These represent more greenfield opportunities for us. So we should see the new verticals continue to scale faster than autos. And in addition, it provides an important source of diversity within kind of our revenue base, right? It is the opportunity when to lean in on certain verticals and diversify revenue so that we can maintain growth rates at or above our long-term growth rate. So some highlights from this past quarter. Very strong revenue growth, strong -- that much of that being driven by traffic gains. Quote request increased 81%. We had positive adjusted EBITDA. And for the first time, we actually had positive GAAP profitability within the quarter. We did launch in -- actually, in Q3, we actually launched commercial, the most recent vertical, and we've added significantly to some of our leadership, especially around the product and traffic areas. So a very strong third quarter that we reported. So key investment highlights. Much of this we've gone over, but we are an insurance marketplace. We service a very large market, and those have attributes -- the market has attributes that are actually -- create tailwinds, favorable for our marketplace solution. We think we're very well positioned competitively, and we'll continue to leverage the model to increase, both VMM in the long-term as well as add to our adjusted EBITDA profile as we grow the business. So with that, happy to answer questions specific on the business.

Can you talk about what happened in 2017? I noticed that revenue didn't grow as much as other years and just, in general, it looks like you grew quite lumpy as opposed to [indiscernible] ...

Yes. So 2017 was a very moderate growth year for us. Much of that was the result of really what happened in 2016 to the auto insurance industry. 2016 for auto insurance was the worst claims year, I think, for about 15 years prior. So it was really a perfect storm for claims losses for auto insurance: distracted driving, a strong economy led to more miles being driven, the severity cost to repair. All of these factors contributed to the auto insurance really taking it on the chin in terms of claims losses. What then happens is the insurance industry, the carriers themselves will then look and reprice and recalibrate where their rates are. And it takes time basically to file and see those rates flow through to their portfolio. In that time, they're not anxious to onboard a lot of consumers that are mispriced. So what we believe we saw was in 2017, a pullback from the auto carriers, specific to the fact that they were mispriced. And we saw that start to reverse in late 2017 as that kind of -- that new pricing flowed through and they returned to growth mode. During that time, some of the carriers also moved dollars from auto to other verticals. But at that time, we were in our infancy in our 2 new verticals, home and life. And so we really didn't capture those other dollars. So we saw a more modest growth in 2017, really focused on the cause being 2016 claims losses for the auto carriers. Even during 2017, if you unpack our growth, which was modest, you'd actually see a stronger growth within our direct category. So one of the things that we just -- I just pointed out was we've been growing direct revenue stronger. Direct-to-revenue continued to grow at an increased pace in 2017. It was really some of our indirect wholesale revenue that actually declined. Part of the theory there would be, working with us directly, we allow you to really target on the type of consumers that if you're mispriced in general, there may be still consumers that you want to onboard that are properly priced. And you can target that through the platform.

So the first question is sort of high level. So I believe that the Google, the Facebook and the modern day players are tightening up their, for the lack of a better word, the reins with regards to their search development and stuff like that. So although let's say, if it wasn't some third party who would [ swim ] here, I guess, what impact does this have on the kind of the [indiscernible] in driving the user?

All right. So I guess, a number of different factors there. First, we have a fair amount of traffic diversity. So Google is, obviously, our largest traffic party, our largest advertiser. But even at the time of the IPO, we disclosed that Google, combined in both search as well as in display, it was about 28% of our traffic. So Google is still not our largest source of -- well, it's our largest source of traffic, but not the majority of our traffic. Also important to understand that for us, Google on the search side is primarily paid, right? This model is primarily paid on that initial acquisition. So we're not subject to the same kind of algorithm changes that you sometimes hear about in Google. With regard to specifically kind of the announcement with Google that they're phasing out the use of cookies. We will be affected. We do some retargeting, but that is for us, still a relative small -- should have a relatively small impact. And as they phase that out over 2 years, I think there's time for a lot of providers to figure out how to react to that. As well, it's just not what we've built the business on, specifically on retargeting. And as well, there's really a level playing field. So it will affect all the players, both the carriers as well as the other marketplaces within insurance. So it's something that we'll be aware of, but we don't expect any over-indexed effect of the phaseout of cookies.

For the last 3 quarters, you have seen a very strong growth on your quote request, the demand side is working really well. Could you shed some light on the supply side and what the future agents and carriers are doing? Are they -- do they [indiscernible] one more? What about the agents? And what's on that supply side [indiscernible]?

Sure. So the growth through Q3 2019 has been -- we think of it as largely being driven by some of the progress we've made around traffic, both our execution, but also by the backdrop of the strong market conditions within auto. Looking at kind of the supply side, as you described it, the carriers and agents that we are working with, I think generally, were quite pleased that we've been able to scale that side of the marketplace to keep up with the volume that we're seeing increasing in traffic. So 81% quote request growth in Q3. There is still a -- there is -- we are basically still accelerating traffic at a rate slightly higher than we are actually scaling our distribution side. And you see that in our revenue per quote request ticking down slightly about 10% in the last couple of quarters, Q2, Q3. And that really is the effect of bringing so much volume to a 2-sided marketplace to an auction, where you're actually seeing us compress slightly the number of referrals that we're able to monetize per consumer. But generally, you see that distribution side keeping up, and we've seen tremendous growth on the distribution, both from expanding new partners, new providers, but really from expanding the budgets with the existing providers, especially within auto.

Can you contrast yourself with other players in the space? I'm guessing auto is an area [indiscernible] ...

Sure. I'd say the biggest difference between us and any other player is really our DNA around data and traffic. It is using data and technology to acquire traffic that is in-market and shopping for insurance, and do that more efficiently than others. That is, if you look at -- back historically at our founding story, we were founded by a couple of MIT guys that recognized the amount of volume of consumers online looking at insurance, the amount of spend in that category and really focused on how do we acquire these consumers more efficiently. We've built a whole marketplace that uses data from the bid, all the way to the disposition data with the...

Who are they?

What? I'm sorry.

Who are the competitors?

Who are they? So certainly, the one that's spoken about the most is LendingTree, with their acquisition in 2018 of QuoteWizard. Somebody that both in LendingTree and QuoteWizard, folks that we respected. QuoteWizard was also a customer of ours on the indirect side, that wholesale side. The difference between them and us, again, I think they've done a really great job of leveraging LendingTree's brand, their recent acquisition of ValuePenguin in order to kind of supercharge Quotewizard. I think the DNA is still very different. We are much more data-centric, much more technology-focused, much more acquisition-focused on traffic. We do that, we believe, more efficiently than anyone using machine learning and using our technology.

John, can you just speak to the short reports that came out recently, that's accusing you, that's knocked the stock down a bit. Have you seen traffic go up to your site? What percentage of your quotes that you had in the last quarter are coming from this verified partner? Will you just speak to that a little bit?

So I'll try not to comment specifically on the short report, but I'll certainly talk about kind of, if I will, I'll go back even to this slide and talk a little bit about the difference that you can see -- probably best on this one -- between a consumer who comes to us as an arrival, and ultimately, the metrics that we publish, which is a quote request, right? A quote request is a consumer who's actually taken actions to ask for a set of quotes. And so in that way, there is a consistent intent on that consumer. An arrival to the marketplace can differ greatly on where that consumer comes from based on -- at low end, we could be seeing a consumer come through an article page on a display ad where we're going to try to cultivate their interest in buying insurance. And that will tend to have a much lower conversion to a quote request than say if we onboard someone through either a partnership relationship or through search. So you can see a big difference in conversion rates based on traffic acquisition. During the year, we also increased the conversion overall within the marketplace. So we have seen both conversion gains in general and performance as well as the change in mix toward more performant traffic sources. Overall, I can say that through Q3 of 2019, we certainly grew arrivals. We do -- we did so at a more modest rate than quote request. But again, much of that is based on performance gains that we had in conversion as well as traffic mix moving toward more performant sources of traffic. Neither good nor bad, right, because we will bid all value -- we will bid all traffic sources to value. But ultimately, using a third-party tool to estimate an arrival, which is not a metric that we publish on a regular basis. So there's a reason why we publish quote request, because we think that is kind of homogeneous intent. That's a consumer that's actually asked for a quote, not simply a consumer who's landed on the marketplace through a variety of advertising means.

Going back to the competition. Can you just give us what the market share is? And what you think market share is? And any another [indiscernible]?

So I'm not sure specifically. So with regard to market share, with regard to the other marketplace models, so I'm not sure specifically how we break down in market share. I'd say we're both -- we think of ourselves right now as not competitively constrained. We think that the market and the kind of the tailwinds that we're seeing with dollars moving from offline to online basically means it's much more about our execution and how we grow the business than how we compete against others. Indeed, the largest kind of competition for traffic doesn't come from folks like QuoteWizard or LendingTree, it really comes from the other -- from the carriers that we serve. So I'm not sure of the exact breakdown of share. Certainly, you see folks like QuoteWizard and us has come up as, around the marketplace models, as 2 of the first names that the carriers mention in terms of folks that they work with. Certainly, it's still early days on the online spend increasing at a rate of at least 16% per year based on the studies we've seen. So it's early days.

We think -- I think 30% at $9.2 billion underlying?

Yes. Growing 16% per year.

Yes. And now that's a little over $3 billion. That's in new revenues of $150 million to $200 million quote. So where...

So if you look at QuoteWizard, they're similar sized. If you look at the other players in the marketplace space, they might, together, add up to something approximating $1 billion in revenue. So it's still a relatively small portion of their online spend.

Can you talk about the type of consumers that actually use EverQuote? Because I was always under the impression that users looking for better pricing on automotive insurance would just retire [ accident ] drivers into that space. So it's not an attractive market for carriers, right? That creates an indication of how much suppliers get and [indiscernible].

You can see differences in consumer kind of the risk profile of consumers based on traffic sources. That's 1 of the things you can see, for instance, search consumers tend to be very intentful. So if you bring a consumer arrival to your marketplace through search, it has a tendency to skew toward nonstandard. But that's really not -- that's an example of how traffic source can skew, but that's generally not all consumers that are shopping for insurance. I think the difference between now and 20 years ago is, consumers get a constant drumbeat that they should be shopping for insurance. And there's basically 2 types of consumers: it's those that shop for insurance and those that know that there's probably savings in it for them if they do. And either they haven't gotten around to it or they're loyal to their agent or some other reason. But I think really now, there's a general understanding that insurance is a place that if you shop for quotes, that you will save money and get the right carrier. That's kind of the drumbeat that we benefit from as carriers are even doing their own offline marketing. So all of that messaging tends to encourage shopping behavior and that telling the consumer it's a place to shop money. So I don't think it skews necessarily toward what would be considered nonstandard consumers. We certainly see all types within the marketplace.

Is the growth in nonstandard growing faster though, even though they'd be a smaller thing versus standard?

I wouldn't have any commentary around nonstandard other than to say, in terms of the growth rates of nonstandard, I would say that clearly, premium and preferred consumers tend to be more highly valued. We see that in terms of the bids that we see for those type of consumers and how we're able to monitor -- monetize them. But one of the benefits of a marketplace within insurance is, ultimately, we can ingest all types of consumers. And as much as premium or preferred consumers may actually monetize at higher rates, standard, nonstandard consumers, we still have carriers that are interested in quoting those folks and still can service that model efficiently. So really, allowing us to have kind of an advantage over any specific carrier. If you were a carrier that only focuses on premium or preferred, the nonstandards -- the standards that show up on your doorstep are inefficient marketing spend. We can segment those consumers and basically get them to the right set of carriers.

Just a question on the carrier's ROI on the ad spend. I guess through you guys, what's the difference between topline sources of, I suppose, your own?

So it would -- versus their own marketing spend? So working through us versus their own marketing spend, it's going to differ a lot by the carriers. Sometimes the carriers' other advertising spend is influenced by things like brand. We tend to sit more in the programmatic bucket. Certainly, all the carriers we work with are going to bid and work with us to their cost per sale targets and we're going to let them do that. In some cases, we'll actually do that for them, in the case of that smart campaigns tool. But otherwise, they're going to bid to their specific values that they see in terms of boiling lifetime value down to cost per sale, and we'll let them target on that. That's going to differ wildly though even again going back to the nonstandard comment based on what type of a carrier even they are.

We've just been through the QuoteWizard. So even though LendingTree spans several products not related to insurance, although theoretically they could, in order to create an increase in weight, they could reduce optimum pricing related to the Wizard. So maybe they're seeing each other as bad with unusual pricing? Is that sort of what happened?

No. We were both competing with QuoteWizard as well as partnered with QuoteWizard before the acquisition. We continue to be both competing and partnered with QuoteWizard after the acquisition. So we haven't seen any significant change. I'd say the biggest thing we've seen is a little bit of the validation that having LendingTree in the space brings to the space. I think you start to see the carriers think about marketplace a little more of it as its own segment of online spend. So certainly, there was a validation aspect of having Tree come in there and basically signal that insurance, in terms of its movement online, is similar to where LendingTree was a number of -- excuse me, where lending was a number of years ago. So I'd say that was the biggest impact from the Tree acquisition for us. For the ...

Now your competitor is more of a [indiscernible]?

No. Certainly, a competitor. They're the other very large marketplace out there. We don't think of ourselves really as competitively constrained at this time. So we don't think of it as, that the governor to our growth is how well Tree is doing and expanding. We think of it more as the idea that there are tailwinds in the industry that will allow multiple providers to do well in this space. We certainly want to be the biggest, but we don't think we're seeing an impact specifically from them. If we are, it's hard to parse out. Again, the biggest competition for consumers is really against our customers. Tree, through QuoteWizard, QuoteWizard was a customer of ours on the kind of wholesale side of the business. They continue to be a customer of ours. So we have a good relationship with them in that regard.

Sorry, what was that search referral that actually resulted in a sale?

So based on survey data that we have with our consumers, about 20% of consumers that shop with EverQuote report purchasing insurance through EverQuote. So the vast majority -- there's still -- this is still a shopping experience where there is an incumbent. There's an existing policy and so consumers are shopping out. Also leaves the ability for consumers to come back in the future renewal dates in order to shop insurance again. So we see a number of consumers that primarily paid on the initial arrival. We see a number of consumers come back for reshopping.

Can you tell me about updated health and the dynamic in the market? And the difference in [indiscernible]?

On health? So our launch of health is really geared towards some short-term policies. But obviously, we're very excited about the Medicare gap supplement and advantage part of the business. That has had a lot of growth. We just launched it. We launched it purposely early in 2019 so that we could go through the open enrollment period with the advantage and supplement plans in order to kind of have the learnings from that open enrollment period. We messaged that we didn't think that health would be a big revenue contributor this year. It was really mostly about us getting to know traffic patterns, setting up the marketplace, and we look forward to that growing in future years.

How many health care carriers are you working with?

I'm not sure that we've given that out. It's still a relatively small number.

Is your margin target long term, like Tree's? Are you -- where do you like your [indiscernible]?

Yes. So I'm not sure exactly what Tree's are, long term. But we think about our margins, again, on variable marketing margin. We think of that operating to around 30%, growing over time to about 40%. Then in terms of adjusted EBITDA, we think of that -- that is something that could be in the mid-20s today operating at low single digits. So we think we'll see expansion in adjusted EBITDA 1 to 2 points per year over the long period, sometimes more when we're able to do.

Until the mid-20s?

Until about the mid-20s. Yes. That's what we put out as our target margins. Okay. I think that's it.