Yelp Inc. (YELP) Earnings Call Transcript & Summary

Welcome, everyone. My name is Matt Cost, the Morgan Stanley U.S. Internet team. I'm thrilled to be joined by David Schwarzbach, the CFO of Yelp. I know you have a statement to read, I'll let you begin.

Yes. Well, thanks, Matt, 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. And then I'll do mine for important disclosures, please see the Morgan Stanley research disclosure website at morganstanley.com/researchdisclosures. If you have any questions, please reach out to your MS sales rep.

All right. The [ legalese ] out of the way. Thank you again so much for being here. So for those in the audience, maybe just coming now to the Yelp story, maybe just start with a quick overview of what you're doing at Yelp -- where you fit into the ad ecosystem and maybe how the business has changed over the past couple of years?

Yes. So Yelp has changed a lot over the past several years. I'll come back to that, but maybe -- just to start, we're obviously a broad-based consumer Internet brand and site and destination. We serve many, many categories in local. So we're focused on local businesses. And I would just say broadly, when you think about ad platforms in digital, you have the top of the marketing funnel, which is much more brand oriented than you have at that point of conversion. We play very much at the point of conversion. And then there's engagement. And we're query based. So when you think about Yelp, you're really thinking about someone comes, they have a very specific search in mind. They're telling us they need a locksmith and we're able to match them with locksmith. That's the focus. Now who's coming to Yelp, I think, is equally important. More than 50% of our visitors come from households with more than $100,000 in income, very educated more than -- or just about 80% of folks have some college -- a college degree or an advanced degree. The majority are homeowners, I believe. So when you put that all together, you have folks who are highly educated, affluent, they like to do research, and they are ready to purchase. That's an audience that advertisers want to reach. And what we found over the past several years as we transformed is we grew up on restaurants. We're known for restaurants. But now more than 60% of our business comes from services. And that has been a very strong driver for us. About 85% of our traffic still comes from those restaurant retail and other categories, 15% from services. But if you're generating more than 60% of your revenue from 15% of your traffic, what does that imply? Those [indiscernible] are extremely valuable, and the CPCs are obviously much higher. So the final point I'll make is that Yelp really grew by driving growth through adding local sales headcount. And that's a very difficult way to grow. So over the past several years, and I think we went through a transformation. We're now very focused on a product-led growth strategy, and we've been able to drive a lot of leverage with that. So that's a very quick overview of a few aspects of Yelp.

That's great. I mean maybe sticking with kind of the theme of what's changed over the past couple of years. I think you're 4 years into the role roughly at this point. So maybe talk to us a little bit about when you came in the areas that you feel you've made the most progress on since then -- and then what your biggest priorities are still going forward from here?

So I joined on Valentine's Day 2020. And a month later, we are all working remote. And then I think it was more than 3 years before I saw anybody in person in the management team. So there's a few things going on in that era. But pretty immediately after I arrived, we moved to reduce cost by reducing the headcount and particularly local sales headcount. And the company had already set this path of moving towards self-service and away from this sales headcount driven growth model, but this turned out to be an opportunity to really accelerate that process. So if you ask me what has been the predominant theme over the past several years, it was accelerating this transformation of the business, moving away from that local sales headcount driven model. And then the second part was really increasing product velocity and innovation at Yelp. And last year, we released nearly 60 features to the site, we're going to talk about LLMs and AI and what we've been doing a lot there for some time, but that has also been an area of focus. And I would just say the last piece was just getting much more efficient in making good trade-off decisions in terms of resource allocation. So what are the priorities going forward. We are very, very focused on services and being the, I would say, premier destination when you're making a decision about hiring a local pro. And we're doing lots of things to improve that. When you come to Yelp to look for a Service Pro, you typically will go through -- well, you start with a query and you can choose to go through our Request-A-Quote process. We've put a lot of effort into gathering the right information from consumers to match them with service pros. And that has worked very, very well for us. Home Services has grown approximately 20% per year for the last several years. It grew at that rate in 2023. So we're going to continue to double down on the services front. And I think there's a lot of opportunity for us to intermediate the interactions between service pros and consumers.

Great. Maybe 1 more high level and then I do want to zoom in on a couple of things you mentioned up until this point. But let's talk about the ad markets and more specifically, what you're seeing from your advertiser customers? Are there pockets of strength or weakness that you're seeing as you move into 2024? And then what are you hearing from your advertisers year-to-date in terms of what they care about? Because it seems like it is a mixed backdrop out there.

It does seem like it's a mixed backdrop. So one of the things that we said on our Q4 call was we saw in the second half of December and then through January, and my comments are restricted to January, we did see softness in consumer traffic across restaurant, retail and other. And I think this is -- other companies have reported this. Obviously, the retail sales numbers for January were soft. It seems like restaurant reservations had been soft, even sign-ups for gyms and fitness were very soft. So there seems to be something going on in this retail segment from a consumer traffic perspective. What we also know is that input costs, which have been going up over the past several years have gone up even more and putting pressure, particularly on restaurants. So minimum wages, I believe it's 22 states increased their minimum wage on January 1, California will go to a $20 minimum wage in Quickserve in April. And those restaurants are less able to pass through cost increases to consumers. And so margins have been squeezed there. And so that is definitely something that we've seen in January -- on the services side, it performed. So there's definitely different dynamics going on there.

Got it. So let's talk about generative AI. I mean, first, maybe we'll begin by talking about AI/ML investments you're making to make Yelp -- making orders -- excuse me, to make Yelp more relevant to users to become more personalized and complete experience? And then second, how do you think about your position in the ecosystem as a more personalized AI assistant that people can use to, to discover businesses to the platform?

So we are applying AI and particularly large language models across the site, but you'll hear neural nets as being a type of method that people are using to improve things. So we've applied neural nets very broadly, particularly in the search algorithm and in the matching algorithm. We rebuilt those with neural nets recently. We've revamped our home feed. Again, we use neural nets in order to choose which photos to show in which order for consumers for review writing on Yelp, we do a lot to ensure the quality of those reviews, we actually don't display about 1/4 of the reviews that we receive because we don't think they're trustworthy for one reason or another. Again, we're applying AI there. Large language models, in particular, we've applied to, for instance, it turns out in search that improving your set of synonyms is very important. And so we've used large language models to pick better synonyms for matching. It sounds trivial, but it's actually very powerful. We are now summarizing consumer sentiment from reviews as a sentence at the top of a listing so that people can have a quick sense for that business. We use large language models for that. We're certainly exploring ways to make the entire experience more conversational. I think that's where -- when everybody is certainly focused on, and so we continue to experiment around that. So we feel like we're very well positioned to take advantage of these technologies as opposed to seeing them as an immediate threat.

Great. So let's talk about some of the core drivers of the 2 segments of services, you mentioned growing 20% than RR&O. I mean I think service has been a big focus, but maybe unpack for us, which of those has the bigger runway? And what are the key drivers that we should be focused on the 2 sides of the business?

Yes. So services definitely has a tremendous runway in home services, in particular, is obviously a very, very large category and we feel like we've really been able to deliver value to advertisers and consumers alike. So I'd say, first and foremost, we're focused on home services, but services categories are much broader than that, and there's areas where we'll have the opportunity to focus in the future. But from the perspective of wanting to be very disciplined about resource allocation, we're going where we're seeing success and performance. And again, I mentioned that the value of a consumer that comes to service is much higher than restaurant, retail and others. So that's going to continue to be a focus and the roadmap is very, very robust there. That being said, we still care about engagement. We enjoy a tremendous amount of organic traffic. People come frequently to look at restaurants. And so we don't want to ignore that, but it's not going to be the principal focus for us going forward. Either way, we need to deliver value to all advertisers. And the way that we measure that is by reporting external metrics, growth in CPCs and growth in clicks. And what we want to see is clicks growing faster than CPCs, that gives us an indication that we're delivering value to advertisers regardless of the category. And the way the system works, it's agnostic on category.

And maybe talk a little bit about how the competitive landscape differs between services and RR&O?

So I'd say the services landscape certainly has fewer competitors. Obviously, Google is a competitor. ENGIE is a competitor, there's others out there. But we think that we've really been able to gain market share, growing at that 20% -- about 20% level in 2023 and so we're going to -- as I said over and over, we're going to continue to focus the effort there. The restaurant landscape, in particular, is much broader. You obviously have delivery players who have gone into advertising, you have more sites dedicated to food and to finding restaurants. And so that's a much broader landscape. But then I think really importantly is that trusted brand that Yelp enjoys. The reviews are perceived as, and I believe they are very indicative of what the experience is like to be for you. And there's so much detail in -- so it's not just the ratings, I should say, but the reviews themselves, like what dish should I order? When should I go? What was my experience like? Is it good for a date? Is it good for a family? All of that detail really differentiates our experience and why I think that we remain very relevant.

So maybe we can cut the business a little differently other than services and RR&O. Let's think about it in terms of multi-location self-serve and local. How does the strategy in focus differ across those 3 different categories? And then kind of similar to the question about services versus RR&O, what is the length of runway in those 3 areas and the level of focus on each of them?

So just for context, our services categories tend to have many more small businesses. And the restaurant retail and other tends to have many more national accounts. So in general services, not particularly concentrated in the United States, and there are obviously many, many large chains. So they are very different in that sales motion for SMB and really regardless of whether it's restaurant or services, but it's much more predominant on the services side, is that inside sales motion or the self-serve motion. And the inside sales motion is you have sales reps, [ cold calling ] folks. It's important to recognize that when you start a business in the United States, you sign up on Google, you sign up on Yelp. And we're obviously going to reach out to you. We have some contact information and we want to take that opportunity to introduce you to advertising. So that's very much services, SMB. National accounts very, very different, Very disciplined. They do their own attribution work. They're very sophisticated in attribution. They use different metrics, different methodologies. They are much more attuned to whatever the macro outlook is. They have obviously, CFOs who have a point of view and they're looking at margins over the longer-term. And so that is very much the enterprise sales motion. You have to have engagement, you have to build relationship, you have to be able to communicate value, and you have to be able to measure the performance of that advertising for a campaign over the course of the year. So extremely different sales motion between those, but at the end of the day, you're either very good at matching those businesses with consumers and that converts to valuable leads to them or it doesn't. That is the same for each of them. So one may be more sophisticated than the other. At the end of the day, if they're not seeing value, they're not going to spend with you.

Maybe let's flip it around and talk about consumer engagement for a second. So how are you thinking about what's changing in consumer engagement on the platform. I think the way that consumers behave for services and our RR&O can be pretty different. And what should we be monitoring to get a sense of what's changing and what the milestones are for you there going forward?

Yes. So obviously, you have to remain relevant and format matters a lot, and people have become much more used to interacting with content from either visually so it can be photos. It could obviously be video. So we've really revamped the home feed and the underlying technology and platform to be able to serve up a much more visual experience. We recently revamped our home feed in order to do that and the types of things that people typically engage with when you come to the home feed, it's going to be like photos of great food. That's going to be much more engaging than like seeing a very nicely done deck or I don't know, for plumbing, which you would even show. So the engagement matters a lot in the app for restaurant, retail and other and for organic frequency. So we tend to that, and it's important to us. And we want to make it easy for people to leave those reviews. We encourage them to leave more photos. We encourage them now to leave video. And we want to make sure that they are writing interesting reviews. Here, we're using LLM. So we have little buttons at the top as you're writing a review, prompting you to talk about, say, if it's a restaurant, service, quality or price and ambience. So we're doing things to make sure that the content is relevant. We're doing things to serve that content in a way that's engaging and that has been utterly revamped not only to remain competitive but enable us to continue to do more. On the services side, it's very, very different. It's Request-A-Quote. I come and actually figuring out the right service Pro for you is not always that straightforward. And so asking the right questions, getting you through the flow quickly and then serving up pros that you have confidence can do the job well is an entirely different experience from a consumer perspective because it doesn't stop there. You may have chosen to go to the Italian restaurant or not. But if you need a service pro to come to your home, you've got to communicate with them. You have to understand the price, you have to schedule with them. You have to be confident that they're going to do the job well, and you have to meet them in your home. So there's a lot more interaction with that service pro than there is necessarily going to be with -- I went to that restaurant, I had a nice evening.

So I guess, is there anything that we should be focused on that might be changing in the consumer experience on Yelp. I mean along those lines, I think you went through in great detail kind of the process of what it's like to interact with either the RR&O or the services side. But is there anything in the pipeline consumer facing that you're excited about?

Yes. So one of the things that really improving our message center, but importantly there, it's masking phone numbers. We think that it's much more consumer-friendly and privacy focused for consumers on the services side to be able to say, "I do want to talk with or text with the Service Pro, but I don't want to give my phone number. So we've added a lot of capabilities around that, but in particular, phone masking is an important capability. Now there's a certain point where you're going to exchange your phone number, if you're going to work with that Pro because you want to be able to do all the other things in order to get the job done. But we think that's very important. Another element, again, as I said, is we want to -- we're looking at ways to make that all more conversational, and we're also looking at ways to prompt the pros to be better at engaging with that consumer from the get-go because a lot of pros are just like call me. And that's not the best experience for consumers. So we have a lot of work to do on that -- on the Pro side. And then I'd just say the final thing is we've rolled out something that we call Yelp Guaranteed so that a consumer has confidence that when they hire that pro if the job doesn't go as expected, they can apply for and receive money back if it meets the criteria that we set. We do limit that to $2,500 lifetime, but that has turned out to be very important to enticing consumers to consider, say, a service a pro that has no reviews or 1 review because we're really lending the trust halo that we have with our brand to those new advertisers. That's been something that's been very important.

Maybe switching to the advertiser side. I mean, what are you changing there from a product and tool and an ad format perspective that you're excited and think you'll see incremental adoption and ad spend as they come to market.

Yes. I should have mentioned this when you asked about national accounts and what are they looking for. We have many more ad formats for the national advertisers than we do for the SMBs. And let me just cover a couple of those because they end up being very important in that they enable these national advertisers to reach consumers up and down the funnel. So it's the awareness, the consideration the conversion and the engagement pieces matter. And I'd say, in general, at least my experience with national accounts is that once they take a platform to spend on, they do want to funnel as much add dollars through that funnel because there's a lot of set-up costs, you determine how to return for your business. So we want to be able to offer that to them. So we've built out the product set for them up and down the funnel. And then equally important, not just on Yelp but Off-Yelps. So we have a product called Yelp audience, which enables us to build an audience for a national advertiser to reach people who have visited Yelp and told us something like I'm interested in cars to reach that audience off Yelp. And just as a theme, as we look at our opportunity, we believe that there's lots of ways for us to engage with consumers off of Yelp, not just that organic traffic, and that's led us to invest heavily in building out our capabilities to actually buy leads through SEM, something that we have not done in the past. So traffic acquisition cost TAC for Yelp has basically been 0 since its founding. Clearly not spending anything on traffic acquisition is an optimal. So we've been spending a lot of time around that. So as a consumer settlement, you see Yelp in search results for Google and on not just SEO, but SEM piece, and then if I can, one last thing that's been very important to us is this idea of ubiquity. So we do license our data. And while we enjoy that revenue, when we show up on Alexa or in Apple Maps, or in the Discover tab on Uber or in the infotainment system for the OEM system -- OEM infotainment system in a new car, that is putting our brand in front of people. So we're finding more and more ways to make Yelp front and center to folks and not just, "Hey, I heard about this thing, Yelp. I think they do restaurants and maybe I should consider them for services." Like we're trying to be much broader than that.

Yes. I guess when you think about the business in terms of brand versus performance and DR objectives, how has that split changed over time? Where do you see it going? And where in the marketing funnel ideally does Yelp drive the most value for advertisers?

Yes. So the Yelp audience product is really our brand product for those national advertisers. It's a lot harder to do brand on Yelp. First of all, for most advertisers, it's the -- it may not make sense in the Yelp environment for -- to pit someone a credit card. That's not why people come to Yelp. They're not expecting that, it's out of context. So they're not typically looking for brand spend. That being said, one of the products that we added among these various products for national advertisers was the way to co-brand. So national advertisers could sponsor ads for local businesses and show that they are committed to a community or to supporting veteran owned businesses or black-owned businesses. So there are -- that ends up being brand spend for them, but it's in the context of what is really bottom of the funnel engagement for that local business itself. So in general, I would say that we are very much on the performance marketing side of the ad spectrum with some brand products.

Maybe we'll talk about the life cycle of your advertisers, how you acquire them, how the relationship tends to scale over time and really just overall your go-to-market, which probably is changing now as you're focusing more on the growth in the services side.

So certainly, we want to show value, and we need to show value early when you first sign up and then we need to show value over time when you be able to communicate that to you and I would just point out that there's really 2 groups of advertisers on Yelp. There's new businesses. I signed up or I registered on Yelp. And we pitch them on ads, and I'm going to try Yelp. And it's very important for us to give them some performance out of the gate. Now that's a little bit difficult. That's why Yelp Guaranteed is important because they have reviews and people come for reviews and ratings. But that's definitely a significant proportion of our advertising. The majority of our advertising though comes from businesses that are 4- and 5-star rated. So there are already good businesses that are trying to grow even more. And what we find is that people really accelerate their spend on Yelp as they see it working, in particular, in services. Because you really -- it's obvious, I got this lead from Yelp. I won the business and that was worth x thousands or -- hundreds or thousands of dollars to me. So when it goes well, that ad budget can grow very, very rapidly. The other thing is that for a lot of these services businesses that have more than one truck say, they do use local advertising agencies. So we care a lot about engaging with those local ad agencies who do have sophistication and do run metrics. And to the extent that we can show them value, then they want to choose to spend on Yelp because their advertisers see value out of that. So we're really very wide in our approach of making sure that we're talking to local service pros and communicating value to them in a way that is relevant to them, having enough products for national advertisers with the required attribution and in the mid-market working with sophisticated multi-location businesses or working with their ad agencies. So you got to be able to do all of those things. And I'm repeating myself, but at the end of the day, in advertising, if you're not showing demonstrable value to an advertiser, they will leave you. And if you're showing demonstrable value, they will grow.

Right. Talk about cookies. Obviously, a big, big talk in the advertising space this year. So just remind us of Yelp's positioning with regard to use of 1P versus 3P data, how cookie fit into that? And then are you adopting any techniques? Do you need to adopt any techniques to work through the loss of cookies potentially this year?

Yes. So because we're a first party and because you come to Yelp and most interactions are on Yelp and we generate the vast majority of our revenue on Yelp, this is less of an issue, obviously. It's also very helpful that it's query-based. We know exactly what you're looking for, whether you're logged in or not I need that locksmith, we now to show you locksmiths. So that insulates us. Now our Yelp audience product definitely depends on that, and national advertisers do want attribution, and it appears that there's 5 emerging solutions. And so we are focused on being able to support all of them, though I would say in general, national advertisers are just now getting serious about shifting to new attribution methodologies. So the 5 -- you have clean room, very expensive. Really it's only going to work for your largest national advertisers. That's something that we're not going to probably support anytime soon. The 4 others are you have Google's Privacy Sandbox. You've got Trade Desk and others trying to come up with a universal idea -- ID. You've got media mix modeling and you've got one more that is conversion API. And so it's not obvious of those 4, which is going to be preferred. So we are investing to be prepared to go in whichever direction the industry goes in. But overall, we're very confident that there is going to be a solution because obviously, ad businesses depend on it. But at the end of the day, advertisers do want to reach those audiences and they know -- they want to know that they're spending well. I would say the one tricky thing with media mix modeling is that there is a lot of variance in how different analytics companies and ad agencies and analytics teams and national accounts approach that. So that's, I would say, probably the least defined of those 5 solutions.

And just to stick with that for 1 second. Is there a reason leaving clean rooms aside, which I think you've explained are a little more involved in expensive. Is there a reason that you can't try to use multiple of those solutions in conjunction with each other. Is that -- are there challenges there?

It would -- it's certainly possible, but there's also only so many minutes in the day. And so I think that most folks are probably going to settle on an approach. For your largest national advertisers, because they run so many campaigns, they may be multidivisional, they may be working with multiple agencies. -- any given team may be using one method and another team may be using another method. And certainly, most sophisticated players will want to triangulate across these methodologies. But I think in general, when you're selling as you're sort of like trying to get to a point where you've agreed on a pretty clear set of metrics and more importantly, a very clear methodology for measuring and it's just too many -- if you have too many things going on, it's just hard. So I think people are going to push for a, an approach for simplicity at the end of the day. But it's still super early. So we'll see how it plays out.

So I think shifting gears to kind of to Yelp's operations. When you think about your most important investment priorities from here, how are you thinking them through? How are you defining them? And what are the biggest ones that you would call out for investors in 2024?

Yes. Services, services, services, value, value, value, trust, trust, trust. Like at the end of the day, we have to serve and deliver value to consumers in a way that they want to engage. They have to have confidence in the content that we generate and they have to feel like when they made a decision using our information that it was a good decision. That remains central to everything we do. But at the end of the day, with this product-led strategy, we have to go where we see the most ability to generate incremental revenue and drive margin. And we think we did that in '23 with 12% growth and 25% adjusted EBITDA. We had flat headcount last year. We had flat headcount in '22, and we guided flat headcount here in '24. So being efficient with our resources remains paramount.

So maybe just sticking with investment for 1 minute. You did mention SEM a few minutes and I want to circle back to that. I think the amount that you earmarked on the earnings call to put into SEM is pretty small right now. So it feels like it's kind of testing the waters. I guess how are you -- or do you agree with that? And then how are you defining success? What are you measuring.

Yes. Yes. So we have to build out a lot of capabilities in '23 in order to do this. And then I'd say we spent to run water through the pipes, just does it work? Can you bid on keywords in categories, in geographies, can you land the person to the right experience? Can you get them through your question , "Oh, first, you got to get them to sign up for an account, "Oh, you need to make it easy for them to log back in with simplifying and going password list. And then you need to be able to engage them in the message center. That's a lot of different things that have to happen because if you don't do all those things, you don't get a lead. If you don't have a lead, it wasn't worth the cost. And clearly, we saw enough performance there for us to share that we would increase our budget here in Q1 by $5 million from what we were spending for what we call experimentation. Now this is optimizing the experience on the way to suitable economics. And again, limiting my comments, we've obviously been encouraged enough to increase that investment. Now how much can we spend depends on whether or not we can make this work economically what we are not doing and something that comes with buying that lead as we do get new consumers. And so those consumers are coming in the most valuable category to us. And ideally, they're going to do more services searches, but also those stay for restaurants. And so we aren't trying to ascribe value to that today. We're saying it has to work as a lead in the first place. So what are the metrics that are going to matter most as we scale spend or decide to scale spend this year. The first thing is, can we get incremental leads to Yelp. This is the #1 thing at a price that we think we can make work economically. That is the first thing. So we're discussing ways to communicate that to investors and to you as we go through the year. Beyond that then, what we want to start to see, of course, are measures that show that it's working economically. And the way that we work, we are CPC model. We get ad budget, we run an auction. We find the market clearing price for that visitor in that category, for that geography at that moment in time. So it's not like we're just going to monetize by buying a lead and sticking for advertisers against it and make that breakeven. That's not the way we work. So there's going to be some lag in revenue as we perform on bringing in these incremental leads. So another measure of that can be what's going on with CPCs. If we're able to bring more demand into the system, how will that show up in CPCs is something that we've been thinking a lot about. But we'll have more to share as we progress.

Great. Maybe just circling back on the AI point, any progress that you would point out on the AI-driven tools and then adoption and feedback from advertisers as you roll those out into the offering?

I would just point to, we really see value by more clicks, clicks growing faster than CPC. That was upside down in '22, very upside down. And we were able to reverse that. And a big component of that was the investment that we made in AI in the matching technology. No doubt. So that is the output metric of some of that investment, and we'll continue to invest there. To the extent that we make the experience more conversational, I think you're going to see that as you use the app. Obviously, we need to see that perform. There's also, obviously, on the cost side, can we make developers more productive, like us quite a range of imputed productivity improvements from using get Hub copilot or ChatGPT or other services. We do think there is going to be leverage on that. So how would that show up? We'd be able to point to increase developer productivity. I think it's still very early on LLMs for post-sales. There's obviously -- I think Klarna had announcement last week about how well that's working for them. For us, I'd say it's still much earlier. And then I think there's a lot of folks who are looking at ways to use AI and large language models to enhance the selling productivity of reps, I think that's actually still quite early. So I think all of us are very optimistic about the leverage this technology can bring not just from creating the experience for the consumer, but in operating the business more efficiently. But I think it's going to unfold over several years.

Got it. Taking at the point of leverage, I guess, on the financial leverage side, talk to us about the margin trajectory for the business from here? And what sources of leverage you would call out, frankly, not just in '24, but beyond?

Yes. So we do believe that this product-led strategy provides leverage over time. And fundamentally, we believe that because as you improve those improvements stack up. And if you're making the right improvements, then you are able to generate revenue and you're able to do that more and more efficiently. So I don't believe we're the first business to do that, and you obviously see in other product-led companies that they've been able to achieve very strong margin profile. So we continue to have confidence in that we've grown margin over the past 5 years from about this adjusted EBITDA margin from about 19% to the 25% that we saw last year. So we don't think we're done. Now I do want to caveat that from an adjusted EBITDA perspective, we've been very focused on stock-based comp. That's feedback that we've received from both the research side but as well as investors, we've committed to getting stock-based comp down below 8% by the end of next year. We expect to grant 65% fewer shares in 2024 than we did in 2023 to employees. So we are taking very, very meaningful steps against that. But in the short run, that does mean that we're shifting some of that compensation from equity, which is adding back in adjusted EBITDA into cash expense. So we're going to go through a little bit of a transition here on the adjusted EBITDA. But underlying it, I point out, like we've guided again to flat headcount here in 2024 at about 4,700. I think that imply -- and we're growing revenue. So that implies under the hood that we're continuing to become more and more efficient.

Got it. Maybe we'll close on capital allocation. I guess how are you thinking about capital allocation broadly when you think about the impact of balancing dilution, which you're just talking about driving higher operating income and then other opportunities like M&A in '24 and '25?

So our approach is to return capital in excess of a target cash balance, and we've been doing that. We do hold capital on the balance sheet for operations, obviously, a buffer above that and then for acquisitions. And so we'll continue to hold some portion of capital. But I think at the end of the day, fundamentally, we're committed to returning capital in excess of that target cash balance. And we did -- the board did just authorize incremental share repurchase of $500 million, which brings the total authorization since we started repurchasing shares to over $1.9 million.

All right. We will close there. Thank you so much for being here.

Thanks.