MediaAlpha, Inc. (MAX) Earnings Call Transcript & Summary

And thank you for joining the MediaAlpha fireside chat here at the Citi Technology Conference. My name is Daniel Grosslight, and I'm the healthcare technology analyst here at Citi. I'm very pleased to have with us the MediaAlpha team. Now as we'll get into a bit later, MediaAlpha is sitting at the intersection of some pretty strong secular themes within the insurance and adtech and healthtech spaces, namely the shift towards performance marketing and direct-to-consumer distribution and most insurance products through auto, life and health some other products. So I'm very excited that the team can join us today to provide their perspectives and what they've built here at MediaAlpha. The goal of this chat is to be more conversational, and I'd love to get some audience participation. [Operator Instructions] So Steve, before we dig into the Q&A here, you're relatively new to the public markets, IPO-ed in October, and you've had a pretty volatile ride. I think you guys tend to fly a little under the radar because you don't fit neatly into one bucket, right? You could be seen as adtech or insurtech or even healthtech, which is what I cover. So I think it would be interesting and helpful for those listening in who are less familiar with the story to do a little kind of table setting here, brief intro to the MediaAlpha story and kind of how you've built this company.

Yes, sure. Well, first of all, thank you for hosting us and having us here. It's fun to be able to tell this story again. So going back, we started the company about 10 years ago. This kind of an outcome, where we're here now, is something that we really didn't imagine at that time. We've been -- we're self-funded and, actually, our IPO was our first primary capital raising event. We started really simply, there was 3 of us, 3 co-founders. We picked auto insurance for a simple reason, because we were good at advertising platforms, online customer acquisition, lead generation. We've done it in other industries, including for travel. And when the 3 of us got together, thinking about what business we wanted to start next, it was really about, "Hey, what can we do to apply our expertise and our skill sets and what industry should we apply this to?" And without quite honestly, a lot of thinking about it, I mean, we researched it, we picked insurance, and more specifically, auto insurance as a starting point, even though we're now in health insurance as well because we knew it was a big market. That's a mandatory product. We knew it wasn't a subject to some of the volatility that we've seen in some of the other verticals that we've been in, like travel, like mortgage, personal finance categories, et cetera. We knew that there was a lot spent on insurance distribution even from an outsider who didn't know a ton about insurance, you knew that. And we also saw that it was not a natively online industry. And so we made that simple deduction, I guess, that there must be a lot of inefficiencies when it came to online distribution and customer acquisition and that if that was -- if there was ever an industry that was probably, I don't want to say behind the curve because I still think there are some reasons why there's been some late adoption, and I think of online distribution channels. But if there was an industry where we thought we could make an impact and actually help bring a lot of efficiencies to the space from what we've done in other industries, insurance was it. And that was really the simple thinking behind launching within auto insurance. And we started again really simply as just the lead generation site. We still have this owned and operated property, but we've obviously more to be so far a lot more beyond that. But as an owned and operated lead generation site, we were able to be profitable since month 3. And again, we've grown profitably ever since and so we were fully self-funded. It was clear after about the first year or so that there was an opportunity to bring a lot of badly needed transparency and what I would call like platform-based programmatic efficiency to the way we define the space at the time, which was insurance lead generation. And if you know lead generation at all, it's an industry that really has not had a ton of transparency and a ton of [ exficiency ] or programmatic efficiency brought to it. And it's still the case in a lot of the other verticals that we're not in, like in the, first of all, finance category, et cetera. And so after our initial success as an owned and operated lead generation site, we've built the programmatic technology platform that led insurance carriers bid a on self-service basis on all the leads coming through our site and bid differently based on whether that consumer was a multi-car homeowner versus a single car renter, whether that person was a male or female, did differently based on time of day, et cetera, i.e., the programmatic granularity and control and the full transparency. And carriers loved it. They just hadn't had that type of a programmatic experience before as it applied to this type of advertising inventory. And somewhat ironically, there was a carrier who then approached us, that was Esurance, and actually asked us, "Hey, we love this platform that we can use to bid on all the consumers coming through your lead gen site. We actually offer this comparisons and experience that's showing ads from other insurance carriers, when we make a determination that a consumer on Esurances likely do not buy a policy from us. Can we leverage this technology to create a marketplace to have carriers be able to bid granularly in a real-time basis on all of the consumers coming through Esurance like they can on your site?" And to wrap things up, that's how we pivoted from being just an owned and operated site to being a marketplace platform where we now work with hundreds of supply partners like Esurance. And because we started with insurance carriers as the first set of supply partners, that's really an important area of distribution for us that we get, which is we work with 35 auto insurance carriers who are both buyers in our ecosystem and, importantly, sellers in our ecosystem who then offer their consumers on their website the option to go with another insurance carrier or get rates from another insurance carrier. If the data science tells them that consumer is highly unlikely to buy a policy from that, and that's a really important part of our business. But we've obviously expanded and we have other supply partners like auto insurance price comparison sites, health insurance price comparison sites, other lead generation sites. And I think, increasingly importantly, working with apps that have a personal finance-oriented comparison and monitoring experience and helping to bring auto insurance comparison to that category supply partnerships as well. So our mission, overall, is really to be the connective tissue between where consumers are shopping for insurance, be it auto insurance, health insurance or life insurance, on a carrier site or a on comparison site or on a lead generation site, and then be able to connect them with as many demand partners as possible. And these demand partners are insurance carriers, insurance agents and insurance brokers and distributors. And it requires a lot of integrations, a lot of custom integrations, knowledge of the space. But insurance is such a big area where so much is spent on distribution. It's more we look and more we uncover, the more opportunities we see. So...

There's a ton to unpack there. Maybe if we could just start at kind of a high level. You mentioned a ton of acquisition spend from the insurance carriers have at around $150 billion spend each year on just acquiring customers. And as you mentioned, historically, it hasn't been a digitally native field, and I certainly come across this all the time in health care technology, is you kind of bang your head against the wall and say, why aren't these companies spending more on the digital channel relative to other industries out there? So I'd love to get your view on, historically, why haven't we seen that digital spend for customer acquisition outside of a few, I'd say, tech-forward companies like Progressive and GEICO? And what do you think will change, going forward, that will kind of spur this investment into the digital channel?

Yes, absolutely. I'll skip over the easy answer and say, it's not that -- it's not really that insurance companies are traditional and conservative, right, by nature and slow to adopt change. I think that, that is largely true for the industry, and I think that's a factor behind it. And as I tell people, you don't necessarily want your insurance company to be on the cutting edge of things, right? You want them to be good stewards of policy revenue or your policy premiums. And so it's something to be expected from the industry as a whole. But I think the bigger reason behind that is that people do overlook the fact that insurance is a relatively complicated purchase process. And so when you look at auto insurance, really, the first insurance policy where you could have a full online buying experience, we forget that there's still some carriers out there, household names out there like an American Family and Farmers and others and State Farm who don't yet support a full online purchase experience. And so it's a complicated experience, I think where the conventional wisdom really was that you needed human intervention to have the policies or have the people be fitted to exactly the right policy, right? And we're talking about auto insurance. When you're talking about life insurance and health insurance, Jeff will talk a little bit more about this. But within health insurance, you're only now seeing really the first online-only, end-to-end policy purchase experiences now being supported and some of that being spurred on by a lot of the insurtech companies that are entering into the space. And so I would say the complicated experience being in a traditional industry, again, not natively digital, notwithstanding the recent influx of insurtech companies in the space. And I also think that particularly within the auto insurance space, some of the early adopters to online customer acquisition, right, there was some adverse selection there. And so I do think that there were some companies 10 years ago, 15 years ago, who first experimented with online customer acquisition. We like to start it to see that, hey, the customers that we're acquiring at this point really aren't good, and they're not sticking with us and they have higher loss ratios and the exact same type of customers that we're acquiring offline. And I think that was just the nature of consumer adoption and the types of consumers that were open to those type of experiences. And ultimately, I think that's what's changing. It's that consumer behavior is finally changing, people are expecting to be able to compare policies online, purchase policies online. And I think that you've seen that even grow during COVID with the growth of a lot of the price comparison sites like Insurance Zebra and Insurify, this is auto insurance. And so I think, ultimately, it's that consumers are becoming more comfortable doing this online. And I think that's really spurring a lot of the change. Because what that has meant is that the direct-to-consumer models like Progressive and GEICO and auto insurance are starting to win out over the traditional agent-based models. And I think that those carriers who have relied traditionally on indirect distribution through agents, be it captive or independent agents, really can't not react to this at this point. And I think the best, most notable well publicized example is the Allstate transformative growth plan. And if you look at that, that's Allstate's way of saying about 1.5 years ago that, "Hey, we really need to go after the direct channel and rely on something other than just our captive Allstate agent channel to be able to sell policies." And so I think that the change in consumer behavior, right, the online-only experiences and then the winning of the direct-to-consumer model is really what's spurring the change within auto insurance. And again, I think Jeff will talk more about health and life insurance and what's changing in those categories. But I think it's really the insurtech companies and similar trends that you're seeing there. And also, I think the lessons that they're seeing from what's happening in auto insurance, which generally tends to be ahead of those other 2 categories in terms of technology adoption.

Yes, makes sense. And Jeff, maybe that's a good segue to you on what you're seeing in health and life. Curious kind of, obviously, we've seen a tremendous increase in kind of the under 65-year-old marketplaces this year. Medicare Advantage continues to grow like wildfire. So just what you're seeing out there in health and life, specifically, and where we are kind of in the small game, what inning are we in, in those 2 segments specifically?

We said specifically for U65 and Medicare, correct?

Well, in general, but yes, I think this year, we saw Under 65 growing rapidly, and then Medicare as well is -- continues to be strong. So I would just like to get your high-level perspective on those markets.

Sure. I think Medicare is actually a little bit further ahead, just in general, because direct carriers are starting to focus a little bit more on direct digital distribution, which is positive for us and also starting to come out with more online flows and being able to bind consumers online. But we're still kind of at the early stages there. And then for U65, I think it's been a little challenging with kind of the different administrations coming in. But just in general there, I think we're starting to see a little bit more of innovation and some of these new companies, like a Bright Health, and some of these other carriers coming to the market. And kind of furthering down, I guess, the digital distribution funnel and focusing a little bit more on that.

Yes. Makes sense. So we have all of this demand, right, for referrals and buying those referrals in the digital channel. But we also have a lot of places where carriers, where folks looking for referrals can buy these referrals. We have the walled gardens like Google, Facebook and Amazon. We have the marketplaces like EverQuote, LendingTree, QuinStreet. Just curious how you kind of differentiate yourself from the pack here. It doesn't seem like there's anything that really would make a supply partner or demand partner captive on your platform, meaning there's no contractual obligation. So what differentiates you here? And maybe within that, if you can talk about the integrations you have with carriers and how that leads to stickiness of supply and demand.

Yes. So let me start with a -- with the bigger question first, which is how we're different in the advertising ecosystem -- well, how is our advertising ecosystem that we support, right, different than what Google does and Facebook does. And there, I think I'll describe it from a user experience, right? Because our ecosystem really starts when a consumer is doing a deep search for insurance, i.e., getting an insurance quote. And so they do this on insurance carrier sites, on price comparison sites, lead generation sites and again, increasingly, through these financial services apps, like Credit Karma is a big partner of ours. And so to explain it from end-user perspective, think about it this way, someone goes to Google, starts the insurance shopping journey by, let's say, Googling or searching for low auto insurance rates. Then they click on an ad for, let's say, the general or travelers, one of our carrier supply partners, and they go through the process of actually getting a quote from these carriers. And so we have the day site's capabilities that enable these carriers to then determine that, hey, Daniel is not going to buy a policy from me. The chances that he will are 0.02%, and so I want to generate referral revenue by helping match him with another insurance carrier. And then Progressive and GEICO and other insurance companies are able to bid for a position in this comparison ad table that's displayed on an insurance carrier site. And so you see that there's a couple of differences here. One is you started off the experience doing a keyword search. But then you do a much deeper search where you're providing a lot more information about yourself, how many cars you have, whether you're a homeowner or not. So you're doing this deep structured search where there's a lot more data available for targeting, right? And so that targeting is different, and the data integration that we then need to do with our supply partners to be able to take that targeting data and make it available to the advertiser, these are all typically custom integrations that we need to do with every supply partner. And so it's a heavy-touch, data-rich integration that we need to do, which is really like not much of type of integration that a Google or Facebook would do with their publisher partners. Then the second thing is that all of that data then is met and then sent to the buying -- to the advertiser, to the carrier who then wins the click, let's say. And so there's a requirement to be -- for the data passing integrations to be done with the buyers as well. And again, that's another set of integrations that need to be done that companies like Google and Facebook really don't do. And so because of the sort of the depth in the conversion funnel really where we play and where our advertising ecosystem plays are just different in kind from the keyword searches that are being done, the Facebook clicks that are typically a step or 2 beyond that, right, where there's higher intent. But then, more importantly, a lot of data for targeting purposes and data testing purposes to help with the back-end conversion rate, because when they use the clicks on an ad, then all of the data that they've already provided when they're doing the insurance quote search then gets passed on so that the form can be prefilled or you can even be taken directly to the initial rate page of the of the buying carrier. And so all of those aspects of this advertising ecosystem just makes it different from the types of systems that Google and Facebook try to support, right? Even if they were really to make an effort to extend beyond their wall garden. Then when you're talking about others who didn't play within this ecosystem, the EverQuotes, the LendingTrees and the QuinStreets, the easy one is really with EverQuote -- EverQuote and LendingTree through their QuoteWizard business, which is that they're lead generators. So imagine if we just did our O&O and that was it, and that grew to become a couple of hundred million dollar business. I mean, that's really, in a nutshell, what you have with an EverQuote and a LendingTree. So they typically do not work with third-party supply partners as we do. And so with QuinStreet, they're increasingly becoming that. So they used to have third-party supply partners. They were the type of black box, nontransparent advertising network that when we first came on that we wanted to displace the transparent programmatic platform-based model. And I hate to say it but like we've largely succeeded in that regard. So our market share for working with third-party supply partners, like insurance carriers, price comparison sites, other lead generation sites is about 80% to 85%. And so that means that we have a ton of scale vis-a-vis these other players, and we talked about that in our shareholder letter about how we're now 2 to 5x their size in terms of carrier spend with us versus in their ecosystems. And I think that goes to a lot of the competitive advantage that is starting to now inert to us because of our scale. I think we first developed those with our transparent approach, low take rate approach, right, and operationally efficient approach because we have about 7x the operating leverage of our competitors. And so that additional transaction volume in our exchange now, I think, leads to more data and the ability to apply predictive analytics tend to bring in just the next level of efficiencies to both the buying and selling of these customer referrals, which then leads to the efficiency that growth -- I'm sorry, that lends off the more scale and then more data. And so as much as I hate to sum it up as a flywheel effect, we are starting to see that at play now. And so there are data integrations that we have that no one else has. We get outsized allocation of budgets as companies like Allstate and others start to come online and make determinations of where they want to go heavy in on their investments. And you're increasingly seeing us as the preferred platform of choice by insurance carriers who really are seeking to consolidate where they're spending and how they're spending their money in our ecosystem. And so all of those things are just beginning additional advantages that, knock on wood, will continue.

Yes. And you mentioned EverQuote and LendingTree and even QuinStreet now are relying more on kind of their own properties to generate that supply, whereas you, I think you have a few owned properties, I think [indiscernible] Health and other smaller segments. But in the core segments like auto and home and health -- let's take auto, you're really relying on some of your demand partners to also monetize some of the referrals and supply partners. So can you talk about that dynamic a little bit? I think that's not that intuitive for folks who don't know much about you. You're a carrier and you're selling some referrals to a competitor. So why would a carrier do that? How many carriers are actually demanding supply partners now? And just take us through the journey of a demand partner, a carrier, actually becoming a supply partner on your platform as well.

Yes. Great question. And so within property and casualty, we have about 35 insurance carriers who act as both buyers and sellers. And then price comparison sites are working with like the Zebra and Insurify and Gabi, they're also buyers and sellers. So let's talk about just the carriers, right? You're right, it's not intuitive that an insurance carrier would choose to actually serve ads from other insurance companies or their competitors. At the heart of this is that insurance consumers are very valuable. There's a high expected lifetime value. So it's really expensive to acquire these customers, okay? And these customers, when they get to your quote page, when you actually show them the rate, you've oftentimes paid $500, $600 to actually get these consumers to the site and to the quote page. But there's typically a high single-digit to low double-digit conversion rate. So only, let's say, 10% of the consumers who are showing a rate -- who you show rates to are ultimately going to buy a policy from you. And so you have this other 90% of consumers who just aren't going to buy anything from you, even though you spent a lot of money, both in branding and performance advertising to get the users there. And so what carriers started to understand -- and this happened before we got there, by the way, because insurance is already doing this before we started to power this. What they started to understand was that there's so much data that the consumers are giving you, right, after they get to your site as they get a quote. You can use that data to then make predictions about who's going to purchase a policy or not. And you can make them -- not accurately, but you can make them in a very predictive way. And so you can identify, right, the 80% of consumers who are going to buy a policy from you and identify that, hey, those consumers are really coming from the top 20% or 30% of consumers are getting scored by us. And so for everyone else, we're far better off if we actually, in addition to our own rate, show ads from competitors who have -- and who are willing to pay $30, $40, $50, $60 for that referral, which then helps the consumer because it aids in the comparison experience. Because the alternative is that they see a rate, and it's not a great price for them, they go back to Google, and they click on the Progressive ad. And so that media revenue that then went to Google now can be captured by the insurance carrier who aided in that referral with one less step for the consumer than they otherwise would have had, right, that they can then reinvest in the customer acquisition. And typically, when we work with carriers in this manner, we're able to enable them to recoup about 30% of their original customer acquisition cost, which in a very competitive customer acquisition market as we're seeing, particularly within property and casualty, that 30% makes all the difference in the world that has the geometric effect in terms of just the customers that you can acquire to get to the site at the first place. And so it can have the step level effects in making your marketing that much more efficient. And so ultimately, that's what's going to drive this, right, which is the acknowledgment of carriers that, hey, the vast majority of consumers that we spend so much time getting to our site, getting so much information from, there's a lot of value there. Most of them aren't going to buy a policy from us, so what are we going to do about that? Are we going to stick to kind of a belief that we shouldn't be promoting our competitors? Or are we going to promote a comparison experience, pivot to more of a media comparison mindset and extract the value from that consumer that otherwise would go to Google or Facebook, right, and then be able to reinvest that into increasing the marketing budgets on our end. And more and more carriers are starting to opt for that. And the other dynamic is that a lot of the carriers coming online now, which are the traditional agent-based carriers who are starting to adopt the direct-to-consumer model, I think need that level of subsidy and this additional level of efficiency to compete with the more long-standing competitors who've been doing direct-to-consumer for a long time and have gotten quite good at it. And so you're seeing that being reflected as well and the increasing openness of these types of discussions with a lot of the traditional agent-based carriers who recognize that they need a little leg up to be able to compete with the long-standing direct carriers.

Yes. Makes sense. Okay. I think that's a good kind of foundation for folks who don't know you too well. I guess, now I'd like to pivot into some of the trends and drivers that you mentioned on your last earnings call that have changed relatively recently. And specifically in P&C on the auto side of things, you mentioned that you're seeing some price hardening from certain carriers that led to a bit of a cut back in customer acquisition spend. I think that is really the kind of the lower-cost DTC carriers that have been impacted the most by that and have since cut back on their digital acquisition spend. But can you kind of draw out the trends there? What's causing this hardening and over the next year or so, how should we look at the insurance cycle and its impact on digital acquisition spend and spend on your platform?

Yes, absolutely. And I'll make the caveat now that I'm not an insurance person, right? I mean I know enough that about insurance over the years. But -- and there's a lot of different opinions about exactly what's going on right now. So -- but at a high level, what's happening is that as the economy starts to open back up, as people start to communicate and resume more normal driving patterns, the frequency is up. Frequency is the number of accidents that people get into per miles driven. And severity has been higher than normal, in part because of inflation. And so, essentially, what's happening is that a lot of carriers or some carriers have misgauged just the losses that they were to incur for dollar premium that they've been writing, particularly during the COVID period. And so they're having to readjust rates as they have lost rate pressures, right? Again, more accidents and the accidents resulting in higher cost to actually fix the cars. And so I think you're right, generally speaking, Daniel, that a lot of the rate increases that we're seeing, the more aggressive ones, have been with some of the lower-cost, direct-to-consumer writers. I mean, Progressive has been one who's talked about this openly and their need to increase rates openly and that they're pulling back a little bit on customer acquisition spend during this time. What we're seeing is that they are pulling back, but I think our channel is one of the last places where they are pulling back. And so the pullbacks have been relatively contained. What we haven't seen are the other carriers, the more traditional carriers, who focus on standard and preferred like the Allstates and others, right, even though they're adjusting rates also. We haven't seen them have that pull back -- I'm sorry, those rate increase then translate into pullback in their marketing spend. And so that's the current status we see it now. Now I think for the going forward here, I mean I really think that the real answer is it's too hard to tell, right? I mean I think I just saw a report come through just in my e-mail today about how the Delta variant and how -- what that's going to do to actually reduce driving and maybe take some of the time pressure off of some carriers to actually raise their rates because their profitability is going to go back up as people drive less because of the Delta variant and some of the delays of return to office that you've seen as a result of that. Certainly, we've delayed the opening of our offices because of that. And so it's a dynamic situation. And I think people forget that no one has actually ever lived through a reopening of the economy from a pandemic like this. And so there are going to be rate adjustments that are going to be necessary. And you're absolutely right. There are different levelings of rate -- levels of rate adjustments that are being made by different carriers. Our thesis, though, is that the Allstates, the Liberty Mutuals, the State Farms and others, right, as they shift budget from offline acquisition to online and that secular shift, that's been accelerated by COVID, that that's going to offset any pullback in marketing that they would have if they did need to increase their rates by more than what we're seeing happen today. Right, so...

Got it. Okay. Okay. And I guess, sticking with those agency-based carriers, you have been buying referrals off. You directly, but also there's an opportunity that you've mentioned to actually sell referrals to agents. So you're gearing up the sales force to really attack kind of the agency market. Can you talk about the strategy there of kind of a new product set geared more towards agents and the agency-based P&C carriers?

Yes, absolutely. The -- so agents -- so we've never worked directly with agents before, and we're starting to now. And I think you'll start to see that business contribute meaningfully in 2022. It's a market that's predominantly been served by companies like EverQuote and LendingTree, as well as some private companies in DMS as well. Our view, at a high level, is that, that market is broken. The quality of leads being sold to agents is exceptionally poor, in most cases, and that there's about a 1% to 2% conversion rate for leads sold to that channel, and that's just an abysmal conversion rate. And one of the problems in this industry is that there isn't a feedback loop of which lead's working, which leads or not. And then there are some agents who just aren't equipped to buy some of the leads that are coming because they're not making contact and they don't really know what to do with these leads. And so for us, we've approached that, hey, this is broken. If we're going to get into it, we need to find a way to fix it. And so we're testing different products. We spent this year primarily focused on product development to be able to connect insurance agents more efficiently to consumers who are shopping for insurance on one of our supply partners' websites, focusing on calls as a way to connect these consumers with agents because selling a lead to an agent and expecting them to call and make contact and be competitive with the call center that insurance carrier has is a little unrealistic. But we've seen that business ramping up slowly as we work on product development. And I think in Q4 this year, you're going to see us start to ramp up on the hiring side to be able to make a big push into this marketplace next year with a better product, right? And with the different types of sales team that's focused on going, and we're putting new agents on the platform and working with them directly as opposed to the sales team that we have now, which is focused on working with these enterprise-level insurance carriers and the media buying teams of these insurance carriers.

How does the contribution margin on that product compare to your core products?

The margin would generally be higher on that product. It's also operationally a little bit more intensive because you would need an inside sales force. I think, net-net, from an EBITDA perspective, it's -- it would be accretive.

Okay. And we'll start to see the impact of that in 4Q of this year?

In 2022.

Of '22. '22, okay. Got you.

And we talked so much about the secular shift, right so to direct-to-consumer, online. I mean the reality is that still over half of auto insurance policies are still sold through agents. And so we anticipate that it's still going to be a strong market or a strong channel. And I think you're going to see more independent agents, going forward, both online and offline. And it's a marketplace. If we want to serve and be the pipes to insurance distribution connecting, right, everywhere where consumers are shopping for insurance, right, with anyone who can actually sell them insurance and provide that best match to the biggest ecosystem that's powered by data science, we have to bring these agents to the table and enable them to find a way to compete for these types of consumers. And so in a simple way, that's why we're going after this marketplace, because it will be an important channel going forward and we want to provide all of those connections and be that connective tissue between the suppliers within the consumers as well as everyone who can actually sell them a policy.

Yes. Makes sense. Jeff, maybe I can get you back into the discussion here, turning to the Health segment. A lot of the e-brokers have had a pretty tough COVID recently, talking about eHealth and SelectQuote and GoHealth. I know a couple of those are partners of yours, maybe SelectQuote, more broadly. Just curious to see -- to hear what you're seeing within those -- that specific segment of the kind of comparison tools within health. They've really had trouble with churn, with high labor costs. How do you think that's going to impact their spend on your platform this year, particularly as they ramp up for this year's annual enrollment period for Medicare.

Sure. So -- we don't believe it will have much of an impact on spend through our platform. These e-brokers still have to grow their customer base, and we provided a very measurable and scalable way for them to do that. And so what we're hearing specifically from our e-broker partners is that they continue to ramp and spend with us as it makes sense. So heading into Q4 for Medicare, we feel pretty good with the e-broker partners.

Okay. That's good to hear. Now turning back to travel. Steve, you mentioned the Delta impact and the unknown on auto insurance, but I think that even more unknown is how it's going to impact that travel segment, which had a nice rebound last quarter, same with education. How is the Delta variant impacting your outlook for travel and education for the rest of the year and heading into next year?

I'll be honest. We don't -- for education, we're uncertain of what that impact is going to be. And education is really primarily one major partnership that we have through our private marketplace model, right, with a major player in that space. And so with travel, a market that we're more actively involved in, right, through our open marketplace as well as seller private marketplace partnerships, there, I mean I think you summed it up. I think as we start to see the market return -- and the market return, by the way, was still lagged in comparison to just how much domestic air travel returned. Because the dynamics of the travel space is that a lot of the profits come from business travel as well as international travel for the domestic, and even for most domestic carriers. And so without that, I think that led the carriers to be more conservative when it came to marketing and customer acquisition. But we were starting to see that come back. We have seen that go down as people are traveling less because of the Delta variant. And so I think it's just, at lending, a little bit more of uncertainty to that space. But what we like is our position there, where we have a very small team ready with the partnerships that we have to really recapture the growth of that marketplace whenever it does return. And these are the other competitors in the space serving this market. It's a fairly contained part of our business. It has the most operating leverage there, which is obviously a big theme for us. And so I think we'll be in a market leadership position when that market returns. And we have the ability to actually have that bet there, right, with a low option cost, almost indefinitely just because of our inherent operating leverage heavy model that we have.

Yes. Got it. Okay. I think another big trend that's been apparent this year is just the increase in pricing that everyone has seen on -- if I look at this on a transaction value per referral, it's had a pretty good uptick. As it becomes more expensive to acquire referrals on the demand side. Obviously, you're a marketplace and you get your big -- regardless of pricing. But do you think you'll see any type of reduction in your take rate because it has become increasingly expensive to acquire referrals in the digital channel?

Dan, I don't think so. With respect to the pricing, the key aspect of our platform that's important to keep in mind is that, that pricing set by the demand partners on a very granular fully sourced transparent basis, right, ultimately measured to ROI. And from a supply side perspective, that's a revenue share-based relationship. And I don't see the pricing there from a margin and take rate perspective getting commoditized, right? And one of the things that we focus on, as we've discussed, right, is building deeper integrations and continuing to go to that next level and next generation of integration, be it data passing, conversion tracking. And all of that is with an eye on helping everyone become more efficient, right, and the average is find those pockets of our lot that work for them and enabling that for our demand partners.

Got it. Okay. Okay. Good. Last question, I think we're just running up on time here. But I want to focus on that margin question a little more granularly. I think one of the more impressive aspects of your model, and you mentioned this previously, is just the capital efficiency. Most of that contribution margin drops down to EBITDA around 60% in my model. I'm curious, as you look to the future, you're obviously building out the agency product, which is going to have a little more OpEx, but as you mentioned, should be EBITDA-neutral to your core product. But certainly, it's become more expensive to hire data engineers, and the sales force is going to be expensive. So how do you think the cost pressures are going to develop over the next year or so, specifically with the talk of all these labor shortages and the difficulties hiring these more specialized employees?

Yes. Dan, well, it's a competitive labor market. But we really haven't seen that cost pressure impact and play a big role in our business today, right? And that's really reflective of our focus on automating wherever possible, right, taking a technology-driven approach to solving problems. So that's just core to our culture. And that's what has helped us, over the years, kind of drive this industry-leading cost structure and advantage, right? That's what enables us to be a low take rate provider and share greater economics with our partners and really kind of consolidate transaction value, which is important. It feeds that flywheel of getting more data, right, clearing more transactions. And so it's core to what we do.

Yes. I mean it's -- yes, I mean, in terms of the size of the marketplace, as we mentioned, we're 2 to 5x bigger than the other marketplaces. But then our headcount is 1/4 to 1/3 of that of some of our competitors. And so I think that is a good combination for us. We do focus on that. And some of that will have to change as we go after different opportunities. But we've also been relatively distributed even pre-COVID. And I think that mindset that we're just going to go after people where they are, we're used to collaborating virtually, which I think a lot of companies are getting good at, is helping to offset some of the tightness that we see in the marketplace in which we're seeing, as everyone else is for sure.

Yes. Yes. Makes sense. Well, I think we are just about out of time. I want to thank you guys, Steve, Tigran, Jeff, for joining us today and sharing your perspective. Very interesting. Thank you all who have joined. I really appreciate your time. We'll talk to you later.

Thanks, Daniel.

Bye.