Good morning. I'm Justin Patterson. I lead the internet research team at KeyBanc. I'm excited to have David Schwarzbach, the CFO of Yelp, with us here this morning. Before we kick into questions, David is gonna kick it off with a safe harbor.
Thanks, Justin, for having us at the conference. We'll be making some forward-looking statements during the conversation today that are subject to risks and uncertainties. Please refer to our SEC filings for more information on the risk factors that may affect our results.
Great. Thank you for that, David. To kick things off, you know, it was roughly a year ago where we were last talking up here, and Yelp feels like a very different company over the past year. You've actually been one of the top performing advertising companies really over the past year now in there. We'd love to hear about just some of the, the product initiatives, some of the trends you've been seeing that's driven some consistent, profitable growth for you.
We have... we're obviously pleased with the Q2 results, which, in a way, reflect what's been a 5-year journey, and I think it's obvious you don't, you don't deliver 13% growth in the 2nd quarter and 25% adjusted EBITDA within the 90 days of a quarter. That's a lot of. That has to happen leading up to that. I would say one of the things that we did about 2 years ago, and we've talked about, is we've really shifted to this product-led strategy, and we did add to that team both in 2021 and 2022. It takes time for people to land and ramp up, but I think you're starting to see the results of that investment.
The consistency of performance comes out of, I believe, of course, the four areas of strategic focus, which we'll talk more about, but each of them demands product. I believe that that team is really operating very effectively. Obviously, we have our new Chief Product Officer, Craig Saldanha, who investors have now had a chance to hear from a little bit. We are really pleased with our ability to deliver value to consumers and to advertisers and make that work across what is Yelp, which is really a broad-based consumer app.
Got it. One of the things that I think also fits with that, too, is just consistency. Each year, you start out framing your strategic pillars for the year and then hold yourself accountable to those throughout. If I go back to this year, we can talk through each strategic pillar a little bit, but let's just start out with one of the very core ones, growing qualified referrals for a lot of your advertisers.
Yeah. Several years ago, we really zeroed in on value. For growing advertisers, you obviously need to be able to deliver value to them. One of the things, just on the theme of why we've been able to consistently deliver over, particularly the past, say, eight quarters, is Yelp is down funnel, it's measurable ROI, it's high-intent audience. Equally important is ensuring that they're seeing that ROI, and we have invested heavily in that ad tech stack in order to consistently improve how we are matching folks, which also adds some consumer side to it 'cause you want more information, and, and we'll get to that. It has a little bit of: How do you go to market? 'Cause how do you pitch that?
At, at its heart, if you're not delivering value to an advertiser that's measurable, especially in an era where people are more broadly pulling back on ad budget, then obviously you can't consistently deliver. I think that's o ur place in the market, particularly around local, has proven much more resilient than I think people expected. Outside of Yelp, we were confident that we could be consistent, and obviously, the results speak for themselves.
Right, they certainly do. Home service is a very good example of that, consistent 20%+ grower for several quarters now. Talk about just some of the changes that have occurred on the product side that have really just unlocked that business.
As part of this strategy, we wanted to monetize service leads, and to do that, Yelp of five, eight, nine years ago was very much around restaurants, and it was a single, unified experience. What we have done over the past several years is really differentiate the experience because what you need in order to figure out what restaurant you wanna go out to eat on a Friday night is very different than finding a plumber or a locksmith. Differentiating the experience, also product-driven, has been very important. Equally important along that is gathering more information from the consumer along the way to understand what it is they are exactly looking for. The more information that you have around their intent, their need, then obviously, the better you are at matching.
You build up the infrastructure that can absorb that information and then obviously deliver that right advertiser to the consumer at the moment when they're looking for something. That goes particularly to things like Request a Quote, which enables us to gather more information, our Message Center, where we're connecting service pros with consumers, and overall, just not assuming that, that a, a single, unified experience will work for Yelp.
Right. Right, teasing that out a little bit more, one of the interesting products you talked about this last quarter was just the ability to, of service providers to actually call the consumer, which seems like something that could have, you know, very meaningful benefits to both sides, have some monetization implications.
Yeah.
We'd love to hear a little bit more about how you're thinking of scaling that product, whether you start small and move gradually, or if this is something we should expect to be a, kind of, bigger contributor earlier than later.
I would put all of this under the theme of improving the Message Center.
Message Center, at the end of the day, is just connection between, obviously, consumer and advertiser. Again, on this theme of Yelp being a broad-based consumer app, privacy is very important. We're a very trusted brand, we believe. The reason consumers come to Yelp is that they believe that they can rely on these reviews to make a decision. We want to be respectful of privacy. At the same time, what does a service pro want? They wanna get on the phone with the consumer, but the consumer doesn't necessarily wanna give their phone number, and they don't wanna be called over and over. We have moved forward with building out this ability to mask the phone number, whether it's gonna be a phone call or a text message, and we're in the process of scaling that.
We think this is very, very helpful on both sides because it enables the conversation, but it gives control to the consumer, around their phone number. Then there are lots of permutations that we can play with there, whether it's, enabling them to, set an amount of time that their phone number's available. For the service pro, maybe there's some differentiation of how soon one service pro gets the lead versus another. In terms of, the technology itself, it's now well proven. If you get in an Uber and you are trying to reach the driver, or the driver's trying to reach you, it's the same kind of intermediation. We're really using tried-and-true rails in order to build out the product from a connection perspective, but we're making it very much Yelp about the way it's being used by people.
Got it. Perhaps to step back just a little bit on the product side and to revisit a theme from a conversation you, Craig Saldanha, and I had about a month or two ago. Would love to hear more about just how you're thinking about that pace of product experimentation on Yelp. Since you do have millions of people visiting at any point in time, this great canvas to really experiment with ad load, new product types, what are kind of the factors you look at when really deciding a new product and what to bring to market?
As you all know, Yelp has tens of millions of visitors, and that enables us to do experimentation at scale. Experimentation at scale means that we can run lots of experiments and get results quickly. We are very much experimentation-driven, agile development. You start with a hypothesis. You wanna find a way to test that hypothesis. You A/B test it. If it's working, you invest more. If it's not working, then you either refine your hypothesis, or you decide not to pursue it. That's very important to product velocity because you don't wanna put resources against things that aren't working, and you wanna double down on the things that are working. Craig Saldanha really brings that discipline from Amazon around product development, and one of the things he mentioned was one-way and two-way doors is very important from his perspective.
A one-way door is: You make a change to the site, you're not gonna reverse it. Something like Yelp Guaranteed, which we can talk more about, where we have a money-back guarantee on the services side, that's a one-way door. Once you start offering that, it's very hard to pull that back. There's lots of things that are two-way doors from a product perspective that you can test very quickly and then pull back. I'd say with his leadership, plus the foundation of experimentation at Yelp, the overall product velocity is increasing. I would say the yield is also improving because we're getting smarter and smarter about the bets we're making by using A/B testing at high velocity against these tens of millions of visitors.
Got it. That's a great segue for AI, but I'm gonna put that on hold for a second and stick with that one-way door of Yelp Guaranteed. You've been very thoughtful about which markets you're rolling that out to right now. Talk about just kind of what your initial learnings are and what you're really thinking about to really make this a broader product.
Yeah. It did roll out nationally in a, a certain set of categories.
We are now at scale nationally, and one of the things that's very important, of course, is around trust. If you're coming to the site, and, you know, you're picking, let's just say, on this theme of differentiating the product between both sides, why Yelp Guaranteed on one side, you don't need Yelp Guaranteed for the restaurant. Like, there's no money back guarantee when you go to Italian.
Yep.
Like, it should be good. If it's not good, you go to a different Italian restaurant. On the other hand, if you're doing a project at home, like re-roofing, doing a new roof, it's gonna be a little hard. We wanna be able to communicate that you can trust Yelp when we're providing those recommendations for the job that's gonna be done. Now, this is particularly helpful if there are no reviews, and we do have a lot of service pros who don't have reviews on Yelp. You add that Yelp Guaranteed, and so you're enabling a consumer to have confidence that they can use them. Also, if a service pro is not getting leads, they're not gonna stay with us from a retention perspective.
As we did do A/B testing on various parts of Yelp Guaranteed before rolling it out nationally, we saw a positive signal across these different dimensions. Those signals will also help inform how we evolve the product. Overall, still a bit early. It's clear good signals that support the rollout, made us confident on this one-way door, and so far feel very good, and in fact, confident enough to roll it out nationally.
Got it. Got it. I'd love a guaranteed chicken parm-
Exactly.
I'll settle for the home service project.
Exactly.
Jumping over to AI, obviously a very hot topic lately, to say the least, how do you just view that being incorporated in your business right now? Is this a potential monetization, engagement vector? And I guess related to that, is this an existential threat on the side?
Yeah. We overall definitely see it as a positive for us, and I think it's a positive across different avenues. We have been using AI for some time. It's not new. Now, it's sometimes it's machine learning, sometimes it's neural networks, sometimes it's large language models. There's a lot of different types of methods that can be applied that fall under this rubric of AI, but it's not new to us. We use it for, obviously, search. We use it for deciding which reviews to surface. We're not new to it, and we have the infrastructure, importantly, to apply these new methods, and then we have the ability to deliver it to the consumer. We have that under-the-hood matching between advertisers and consumers. Very versed in this technology. Overall, a positive for us in a couple different ways.
Large language models are expensive to develop, but we get the benefit of using them after they've been developed, so the cost to us is modest. It, it enables things that really haven't been possible before, which is, say, summarizing large amounts of text. One of the ways we're doing it is really picking out parts of reviews that really are representative of what people have written about. Then, of course, they can still go and read the review. We think it enhances the experience on that side. It is being applied on the matching technology. We think in Service Center, it's gonna enable us to make it more conversational. There's also just the opportunity to increase developer productivity. We're running tests right now and seeing how developers can use this in the right way.
Overall, we personally, I think, and we broadly see it as much more of an opportunity than an existential threat for Yelp.
Got it. Got it. We've talked a lot about just monetizing service providers, monetizing the leads, the ad tech stack, the product. We haven't talked a lot about channels so far.
You made a lot of progress there this past quarter. I believe 50% is now self-serve or multi-loc.
Yeah.
Congratulations-
Thank you.
... on that accomplishment. You know, just talk about what you've really done there to scale those two channels and how we should think about the growth there going forward.
For us, one of the things that we recognized a number of years ago, and maybe it's to state the obvious, which is, if we can acquire an advertiser and they can manage their own spend, it's much less expensive than having a salesperson involved. We really invested in self-serve, which is a product effort, and I'd just underscore, it's not just about acquisition, it's enabling the person to manage their advertising themselves after they've chosen Yelp. If you've chosen Yelp, you've sold yourself on Yelp in the first place, so we do see higher retention. That's gone well and has continued to grow nicely, mid, about 25% growth in the second quarter.
On the enterprise side, a few things had to happen. Again, to state the obvious, very large TAM, very, very, a customer that knows what they want and is very discerning. You have to build out the attribution. You have to be able to give them all the products they want, not just at the bottom of the funnel, but up and down the funnel, on and off Yelp. We can talk about Yelp Audiences. That's very product-driven. There's also a human component, which is clearly sales is very involved in that. Now, they're more productive because there's larger revenue opportunity. Those enterprise sales... That enterprise sales team took a number of years to build out. We now think it's high performance, well-functioning, and obviously, the performance at 15% overall for services has done well for us.
In general, we've seen really good growth on our multi-loc business. When you put those two together, they're growing faster. We're at 51% now coming from multi-location and self-serve, and we think that was a, you know, a great threshold to, to pass.
Got it. Are there any incremental margin dynamics we should think of as you move towards more of these low-touch locations?
We do definitely see a margin pick-up because they are more efficient channels for us, and obviously, 25% adjusted EBITDA margin in the second quarter, we were very pleased to see. I just maybe as a theme on that, because it's a product-led strategy, we also said that we'd hold head count about flat this year.
Mm-hmm.
We'll get there at the end of the year. We can talk about that as well, but we are seeing the results of what we thought could be the case, which is, as we exceed on revenue, we can flow that through to adjusted EBITDA.
Yep, yep. One more high-level revenue one before we flip over to that margin conversation. Yelp Audiences, which you just brought up, has had a lot of progress over the past year. What type of advertisers are you bringing in versus who you've seen historically?
One of the things that we really like about Yelp Audiences, this is where we are providing an audience to an advertiser off of Yelp. That opens up the opportunity for us among categories of advertisers. You don't have to have an obvious reason to advertise on Yelp to use Yelp Audiences. We learn a lot about consumer intent from their search queries. Say you're looking for a car. Well, we, on Yelp, it's about the dealership, but for an OEM, it's about buying their particular car. Obviously, OEMs care a lot about ensuring that consumers who are considering switching brands actually have the opportunity to go down that journey. We consult to the auto OEMs. You've got the financial services companies. You've got a lot of folks who don't have an obvious way to advertise on Yelp.
That being said, the brands that are advertising on Yelp now have a reason to spend more with Yelp. They see that return, and so we're able to get more share of wallet.
Great. Two more from me, and I'll then kick it open to the audience. Just sticking with Yelp Audiences for a second, you know, we do have the deprecation of the third-party cookie coming up next year.
Yep
... or so Google says.
Yeah.
We'll see the exact timeline if it sticks.
Yeah.
As you see just more, more of these measurements end up getting deprecated, how does that change the advertiser demand for Yelp Audiences?
We overall believe that there will be an industry solution. Obviously, this is a massive industry. Google's delayed it a number of times, working to find something that's gonna work for the entire ecosystem. One of the themes that's definitely emerging is CAPI, and so we're also looking at that. We believe that we will be able to continue to participate in this attribution that is foundational for Yelp Audience, as well as just generally, large advertisers wanna be able to see that they're earning a return. One of the things that's very important about Yelp is first-party data, so we can generate a lot of that data ourselves, and then also we have something called Yelp Store Visits. When you're using Yelp, it's local, so you're willing to share with us location. That ends up being very important for attribution.
We have our own ways to also get the attribution for these large national advertisers, and we see ourselves being able to participate in whatever emerges from an industry perspective.
Great to hear. Margins, we've danced around it a couple of times now. We've shown a lot of progress this year, like you said, mid-20s margin, headcounts being kept flattish, and you're also starting to see the benefits of the action to reduce SBC.
Yes.
I think your target is sub 8% by the end of 2025. As we kinda put all this together, just how should we envision margin progression at Yelp?
Yeah, w e have been very focused on driving margin expansion. We believe that the product-led strategy will deliver margin expansion over time. Obviously, 2021 and 2022 were about investing. As I mentioned, we believe that we're starting to see the benefits of having made those investments. In terms of stock-based comp, we're committed to lowering stock-based comp as a percentage of revenue to 8% by the end of 2025. The way that we're doing it, and it's something that's already been underway for a number of years, is hiring, particularly in product and engineering outside the United States, where engineers and product managers do have a preference for cash compared to equity.
That being said, we're in the midst of planning for 2024, and we will continue to shift compensation at Yelp away from equity towards cash in order to achieve the goal, which we remain committed to. The implication of that, of course, is that we believe that we will be able to deliver higher quality adjusted EBITDA. For investors who focus on GAAP EBITDA, then obviously, you're we, we believe we can deliver improvement over time as well. Just to say, the other thing is that if we're issuing less stock from a share repurchase perspective on a net basis, we'll see, I think, larger reductions in overall share count.
Net, lots of ways to see GAAP earnings growth over the next few years.
Exactly.
Great. We've got about five minutes left. I can keep on going, but I'll pause here to see if there's any questions from the audience. Yes, sir?
[audio distortion]
It's a great question. It's an important question. 19 years of experience doing this, and I would just say broadly, I think one of the things that's underappreciated about Yelp is our ability to curate content, of which fake reviews is one element. Of course, large language models introduce a new vector for people to create these. We think that we're good at detection. We use, also use AI, you know, to do that detection. We're very well known, for instance, for mystery shopping and, you know, trying to ensure that we're discouraging people from, or businesses from using fake reviews. Out of all of our reviews, we don't even show 25% of them because we're not confident that they're genuine.
That is expensive to do, but I think it's resulted in this high level of trust that people have in Yelp's reviews, and we get that in the brand. One of the things that an FTC economist did two years ago was compare the distribution of reviews on Yelp to other platforms, and it's much more even. That's hard to achieve, 'cause people generally, they're either super happy and they wanna leave a glowing review, or they're really annoyed and they leave a terrible review. Really getting that differentiation is a part of the way that we prompt people to leave those reviews. So in a way, seeing that evenness of distribution or more even distribution is also a measure, because if you're writing a fake review, you're not leaving a 4-star, a 3-star, a 2-star. You're leaving 5 stars.
That's also a measure of it. Overall, we think that we have a lot of experience doing it. It's an arms race, so you have to keep... You know, you don't let up on it. You stay active around it. Overall, we feel confident that we'll continue to be able to detect and deprecate the reviews that we don't think provide real information to consumers to make a decision.
Great. I guess, you know, sticking with just kind of the overall consumer side of the, the business for a second, you're starting to incorporate richer media into the business.
Yeah.
Video's becoming a little more common on there. How do we think about just, you know, what that can do as both an engagement and monetization vector for you over time?
Absolutely. Obviously, broadly, consumers have moved towards more visual experience, whether it's a post and a photo or video. We are incorporating much more video. As part of review, of a review, you can now upload video. We have not gone down the path of video reviews alone because we think that the, the star rating, but much more importantly, the text is still important. We see a lot of uptake around people wanting to leave video around those reviews. Ad formats enabling video are very important. Obviously, advertisers are creating a lot of assets using video, so we wanna enable them to use their best-performing assets on the video side. On Yelp, that's another thing. Our feed is becoming much more visual. We're actually applying AI there to really pick out the right things to show. It enables ad formats in that flow.
Actually, very under-monetized on Yelp is just our incredible library of photos. We're also looking at ways to intersperse ads within photos and videos. There's still a long way to go as we adopt more video. Just broadly on the theme of, you know, people are leaving video reviews on other platforms, we just think completeness is very important. It's fine to leave, you know, a video review on TikTok, but that's gonna be one restaurant out of possibly hundreds or even thousands in a large city of restaurants. On Yelp, you've got the completeness, you've got the reviews, you've got the ratings. We haven't seen that being something that erodes people's desire to use Yelp to make those decisions.
Got it. We've got roughly a minute left. Perhaps we end with a big picture one. As you mentioned, you're nearing that, your end of the year budget planning process. You've made a lot of progress with the strategic pillars. As you think about just kind of the next iterations of product that you could do into 2024 and beyond, what gets you, Craig Saldanha, and Jeremy Stoppelman excited?
There, there are definitely a few things. First of all, we are continuing to see a lot of opportunity around the four strategic areas. We're not letting up on that. I think we've been very disciplined in applying the resources against them, and they've clearly been working for us. It's within that context that I'd answer. Clearly, large language models are exciting. I think that we can have a much more conversational experience on Yelp. I'm personally excited about that. Message Center, I think, is really important. Broadly, we talked about being able to mask phone numbers and connecting folks in an, in a, in a, in a way that's privacy, respects privacy. I'm very excited about that. You heard from Craig Saldanha.
I think what Craig Saldanha is really bringing is that opportunity to do two things at once. It's, hey, we had a capability. We haven't talked about we're looking at, using SEM now as a way to acquire customers. Well, we're building a lot of componentry around that that's gonna enable us to do other things, like making it easy to create an account becomes very important. Passwordless login reduces friction. Then you can use these building blocks to do other things to create consumer experiences. So we can, I believe, continue to execute on our strategy. We have a, a, a good roadmap, a large roadmap, a long roadmap around that. Then you build components that enable you to continue to evolve the experience and meet consumer expectations along the way.
When you put all that together, then you come up with hypotheses, you test it, and then you can do more with Yelp. I think we're all very excited about where we are and the investments we've made and the opportunity ahead of us.
Great. David, thank you-