Okay, good morning, everyone. I'm Mark Murphy, a software analyst with J.P. Morgan. It's a great pleasure to be here this morning with, I'll start on the, on the far end with Brian Mulroy, who's the CFO of Semrush. And seated right next to me, we have Eugene Levin, who is the president of the company. So first of all, gentlemen, I just want to thank you so much for taking the time to be here with us.
Thanks for having us.
Always a pleasure.
Maybe we could begin with you just giving, for the benefit of anyone out there in the audience who doesn't have familiarity yet, just a kind of a quick one-minute, you know, sort of overview of yourselves and Semrush.
Sure, yeah, so I can start. I'm Brian Mulroy, CFO at Semrush. I've been in tech and finance my entire career. Spent a number of years at Nuance Communications, which is a voice and speech technology company, then at Microsoft for a few years, and now CFO at Semrush. And Eugene?
Yeah, I've been with Semrush for eight years. Actually a little bit more. Before that, was in venture capital. Did everything from, you know, seed stage, late stage, pre-IPO. Invested in many companies that you probably know, but I'm not going to advertise them today. It's all about Semrush. All right.
This is an advertising-free session here?
Well, I'd like to sell a couple subscriptions, of course, if someone is interested. Actually, we have a lot of investors who buy our software, so if you wanna learn more, I'm happy to, you know, answer your questions later.
Okay, okay, great to hear. I'm sorry, do you have something else to say?
Yeah, I was just gonna say, in Semrush, just to give everyone a quick overview, we're a SaaS recurring revenue business, highly profitable. Our focus areas are on marketing and in particular online visibility technology. Our channels or our major focus is on any technology or data that helps companies to enhance their online visibility. Key technologies around search engine optimization, search engine marketing, social media, competitive intelligence, digital PR, content marketing, and in local advertising. I'll give you two examples. If you're not familiar with the term SEO, everyone knows, if I took a survey of this room, I can guarantee you none of us click on the second page when we're searching for results.
It's sort of death for companies and businesses, so we help them to create the awareness, to create content, to figure out what keywords and information is gonna help them rank on that first page for relevant searches. And then for local marketing, I'm sure we've all searched for, you know, the best pizza shop or the best painter near me. My daughter was showing me a TikTok the other day and said there's actually a pizza shop called Pizza Shop Near Me now, apparently. There's a local or geographical component to searches. We help companies to improve that as well. A couple of things from a financial perspective. We just delivered our first quarter results for 2024.
Q1 2024, we had $85.8 million of revenue, up 21% year-over-year. Our gross margin was 83%, also up year-over-year, and then non-GAAP operating margins at 11.3% and up 2,000 basis points year-over-year.
So, I'm thinking of changing my name to research analyst near me.
There you go.
What do you think?
So let's... You know, you've executed well since the time of the IPO, but I'm wondering what you think might be the most important, you know, to understand about this longer-term evolution of Semrush. Because in the beginning, Eugene, it really was an SEO tool. It really wasn't anything else. And now it's this comprehensive, you know, SaaS solution for online visibility. And when we look at what is in this platform, the breadth of it is pretty unbelievable. So it's content marketing, it's advertising, it's social media management, it's competitive research, and the list kind of goes on and on. How wide do you want to open up the scope here on the product side?
I, I think in terms of, the breadth of offering, we really ask ourself, only one question: If I'm an average business, like, you know, normal business, what am I interested in? And today, the answer is, I'm interested to be everywhere where my customers are. And as internet evolves, and there are more and more places for people to discover new content, new products, and new services, businesses have to optimize for all of those places. You know, when, internet was young, you could have a website, and suddenly people just come to your website. And today, you create new website, nobody shows up, only bots. So to attract traffic, to attract attention and eyeballs, you need to be visible in all those places where people spend time as-
Mm
... and there are more and more of those places, so it's much harder for marketers to navigate this landscape. And as long as complexity grows and as long as there are new and new places where people discover content and learn about products and services, we wanna help marketers and business owners to optimize their presence in those places. So expansion will keep going, until there are new ways for people to interact with content and, you know, consume information. And as of today, key areas for us are, like we already discussed, you know, search, where people find a lot of things and spend a lot of time, you know, social media, YouTube, reading digital media, and, interacting with maps and other kinda data services and so on. We help businesses to be visible in all of them.
So we've always been impressed by the sheer volume and variety of data that you've been able to ingest into the platform. It's quite the engineering effort. And those numbers go into the billions. I think sometimes they go into the trillions. We look at the number of backlinks that you're indexing. We look at the number of URLs that you're crawling, and it seems to make a difference with the product. Could you help us understand the nature of the data sources, and then how does having the better accuracy and better volume create value for customers?
Yeah, absolutely. So I'll start with just an explanation of what we do and then explain why it's a good thing. So, what we do really is we crawl the internet the same way search engines do that. So we go in every single hidden corner of the internet, and we do that to understand how different websites are connected to each other, and which website drives traffic to another website, and so on. Who shows advertising on different websites? You know, who advertises on Google? How different businesses are presented on Facebook. And then we connect the dots to build what Brian, for example, likes to call kinda map of the internet. And if you have this map of the internet, then Semrush becomes kinda like a GPS of the internet.
So if you need to go from point A to point B, which in marketing would be, you know, getting more traffic or getting more visitors, then we can build a course and tell you what to do. And if GPS tells you, "Turn left, turn right," then we can tell you, "Oh, advertise here, or optimize for this, or improve your website, or create new content." And we can give you this advice only because we have this underlying map of the internet that we collected through our, kinda crawling and other data collection methodologies. And if you think about this analogy, GPS is really only as good as a map. You know, when Apple Maps first launched, they had, you know, bad maps, and a lot of people couldn't get from point A to point B.
The same way, if your map of the internet is not very accurate representation of reality, then you cannot really give people good advice, or you give them advice and they end up not there where, not where they wanted to be. So, if for maps, high accuracy means having every street, then for what we do, high accuracy means having large volumes of data and crawling internet as frequently as possible and as deep as possible to make sure that what we know about internet is the most accurate representation.
Why isn't that a proposition that is too expensive? Why wading that deep into all the data, crawling it, searching it, right? We're in an inflationary environment, third-party data, clickstream data, that you have the infrastructure to store it and process all that data. What does the cost profile look like?
So I think, and Brian can talk a little bit more about margins, but I'll make it a little bit of a shortcut. We have phenomenal margins, but the reason we have it is because since the very beginning, we actually didn't have that much money, and we were kind of bootstrap company, so we had to figure out how to get more with less, and we optimized our infrastructure and servers, specifically to work in this kinda high-load environment. And another thing that is unique about data businesses is that you have very high cost of getting into the business, just building the asset in the first place, but once you have it, you don't necessarily need more money to serve new customer. Again, you can think about Google Maps.
It was a big heavy lift to give, to get initial map of the city or Street View.
Mm-hmm.
It's a big investment upfront. But once you have it, it doesn't cost you more money to serve another 1,000, or 100,000, or 1,000,000 customers. And, Brian might want to talk about, the margin profile.
Yeah, I mean, we, this is a structural advantage for us, so we have an incredibly efficient platform that we've built up over the last 12 years. The other key, which Eugene didn't get into, is we are able to get data from our own customers as well to supplement it. So we have 112,000 paying customers today, about 1 million free users. They're all using this extensive platform in a very efficient way, and we're able to capture their data to improve it and create a flywheel effect, where we're getting even more customers and creating even more efficiency in the platform. That advantage has allows, allowed us to create 83% gross margin in the business. What's in that gross margin number?
We have credit card processing fees, our technical support, and our CSMs, and of course, our cloud infrastructure, but it's very efficient and something we believe we can continue to scale over time.
Okay, so economies of scale off the initial investment, and then it's kind of inherently a very efficient infrastructure.
Mm-hmm.
Okay, so let's talk about, you know, what's happening in the, in the SMB market. So Semrush recently reported a healthy quarter. You, directionally, you took the guidance up, you know, just a touch. There wasn't much of that happening-
Yep
... in the software landscape. And you did this while SMB software providers really have been suffering, right?
Yep.
I mean, most of them have been calling out softness and demand. So what's the difference, or what's driving, you know, relative health for Semrush these days?
I think there's three fundamental things that we're seeing that gives us optimism about our potential for the year, and of course, there's a reason for our success and ability to grow over the last few quarters. So key thing is, there's a secular market shift that's happening. A lot of companies are realizing that they just can't have a website, and expect to generate traffic and convert that traffic to paying customers. They have to engage and invest in areas that allow them to get exposure to where their consumers are, and that's, their consumers are searching, they're reading blogs, they're looking at videos, they're going on social media, and we have data and technology that allows them to understand that and to optimize their website for performance within those channels.
So key thing is, while companies are, of course, being efficient and watching the bottom line, they still need to invest for growth, and they're shifting dollars from areas that are inefficient to areas that are more modern, and efficient, and that's where we play, so we're benefiting from that. The second key thing is, while there are macro pressures, and of course, we're feeling it, we have an extensive portfolio. So we already have 112,000 paying customers, but we have an extensive portfolio that allows us to continue to drive cross-sell and upsell throughout that extensive base. And we've been able to grow our average ARR per paying customer over the last year. And then third key area is we're moving up market into the enterprise.
So we've had about a little bit more than 5,000 of our 112,000 paying customers are enterprises that have more than 500 employees. They've been asking us for more enterprise rich feature set. We've delivered that to them, and we believe we can inflect ARPU upward in the next year to actually continue to drive growth for the company.
So you mentioned that you are feeling the macro, Brian, and I'm wondering if we can touch on that for just a moment. Do you sense any different type of business health, maybe developing into this year with your customer base? Part of what I'm wondering is, might they actually lean in on digital marketing or online visibility this year, if only because they were kind of kicking the can down the road? They were deferring a lot of those projects, right, in maybe in the last year or two. So in other words, like, normally, the marketing spend is going to go along with the economic cycle. Do you think it's going to do that?
Yeah, I mean, we're just like every other company, we're not immune from macro. We're weathering through it very well. We just grew 21% in the first quarter-
Right
... and have been above 20% for a number of quarters in a row. And, you know, key for us is we have an extensive platform that customers, companies need, and they'll continue to invest in it. I don't think they're necessarily pausing. They may be prioritizing projects and subscribing to certain portions of our platform, and then, moving on to something else, for periods of the year. So, you know, key for us is we see that there's an extensive need. Customers continue to look for us and subscribe to our platform. We just added another 4,000 customers in the first quarter, so we're weathering through it well. But we believe if macro turns into a tailwind, of course, we would benefit from that and potentially see even better growth than what we're experiencing.
Can I ask you, when you look at the behavior across your customers, is there any chance that what they're doing is, you know, the budget for this technology is remaining, you know, is not changing a lot, just given the macro, but the legacy tech, you know, the legacy marketing spend is down-
Yep
... the modern marketing spend is actually increasing?
I think that's exactly what it is. I think, you know, I was asking some colleagues the other day when the last time they saw one of those car dealerships have a commercial on TV, you know, where they go from, like, national advertising down to local. I haven't seen it. Mostly, you see national advertising. So most companies are keeping their marketing spend flat or slightly increasing it, but not at a significant rate. But there's a significant shift happening, where they're finding efficiencies and figuring out what's going to work for them and, of course, shifting their dollars to more optimized channels, which is where Semrush plays, and it's why we're benefiting from it.
Okay, so relating to that, maybe we can touch on your dollar-based net retention for a moment. I think when we look back a year ago, it was 116%, this year it's 107%. But you mentioned on the earnings call that this could be... we could be getting around a trough-
Yep
... at 107%. I think you even alluded to a possibility it could gradually increase.
Yes.
Maybe we don't know when, but is it more a function—so again, coming back to just the macro, is it a different customer buying behavior? Or is it you're mixing upmarket, maybe that has something to do with it? And then you had said something on the earnings call about changing sales incentives-
Yep
... that could help this too. Could we maybe double-click on that?
Yeah. Our net retention rate is a very important measure for us. So it measures, of course, our gross retention, plus our ability to expand, and drive cross-sell and upsell, in the business. It's been very strong and above 100%, since we started as a company. We just achieved 107% in the first quarter. We do believe with all of the investments we're making in enterprise and continue to extend our portfolio, we did a lot of AI investments, which Eugene will talk about, where we're monetizing AI in the platform. So we believe that's going to be an inflection upward for our net retention rate. The other key thing is net retention is a function of mix for us.
So our retention rate and expansion rate for the enterprise is of course better. They have a higher propensity to adopt more of our platform and to renew their subscriptions every year. We also do play pretty heavily in the SMB market, which is affected by macro, where the retention and the ability to expand is a little bit lower. So the more success we get in the SMB, the more impact that potentially has on the retention metrics. So we're always looking internally on: What's the retention? What's the expansion rate for each of the segments of our business, which are very strong and healthy? And then what's the mix of each of those businesses and the effect that it actually has on it? So have to look at both.
Okay. Now, you also effected a, you know, relatively substantial price increase, I believe, at some point last year. Can you walk us through... We've been in an inflationary environment, and that's been a common thing, right, for software companies to be doing. How do customers respond, and how do you think about maybe potential for pulling that lever again? Because, again, the breadth of the platform is expanding so rapidly.
Yeah. Pricing's a very big lever for us, so we did just do a price increase for our core portfolio in the third quarter last year. It was an 8% increase on our new customers, and then we rolled it out to a small cohort of existing to test to see what the impact was. A couple of quarters later, we're very pleased with the impact. We've actually seen an increase in gross adoption. So, we're still seeing strong demand for the platform and not seeing any impact from the price increase. And then a retention on the cohort that we rolled out the pricing to, also just as strong as as the rest. So we're really pleased with the elasticity of the platform and exploring potential for future price increases.
Pricing for us, though, is a function of a number of different things, though. It's not just raising the price for a single product. We play in a lot of different channels and have a lot of different products. So we're looking at pricing, we're looking at changing the way the tier structure works, so we have different upsell paths. And then at certain times, we'll monetize features separately so that there's a separately monetized SKU that we can sell, and drive growth in our average AR per paying customer. So we're always looking at all three of those as options to just generally increase the value of our platform out there.
Okay. In, as one example, and kinda continuing on that thread, you recently released an enterprise product.
Yeah.
I think you've been working on that for quite some time, right, Eugene?
Yeah.
That that saw general availability recently, and on the recent earnings call, you said it had a good customer response, which is nice to hear. Can you tell us a bit about that product? You know, it was not a focus of the business, even as I recall, you know, two or three years ago. It is now. What is different about this feature set, and what's the kind of customer that it's gonna appeal to?
Yeah, I think everything you said, absolutely right. Haven't been a focus for a while, but because we had such a good data, we always had customers who are large corporations because they just couldn't find anything that would be better for any kind of money. So they would settle with our SMB product, which is, you know, very affordable, and they would pay a little bit, and then build additional infrastructure around our data to satisfy their needs. And, you know, those custom installations, they would become quite expensive to maintain for them.
For example, typical customer would have, you know, Semrush, and then on top, maybe couple, couple workflows that are automated through Python and spreadsheets, and maybe they would use data warehouse to move data from Semrush to somewhere else, and then they would use Tableau or Looker or Microsoft BI for data visualization. And they would have a couple analysts and, you know, sometimes developers, data scientists to work with all this, and it becomes very pricey to maintain. And often they would write us, "Guys, why don't you just build all this stuff for us, and we would just pay you one subscription, and we don't have all this additional infrastructure to maintain?" And since they were relatively small percentage of our customers, we usually wouldn't do much for them.
Only recently, when we did the math, we started seeing how big of an opportunity that would be, and we started talking to people, "How much would you pay for this?" The numbers they, they were giving us were very generous. Right now, we estimate our typical large company could pay us 10x-15x average check. We started working on it roughly two years ago, and we focused on three primary things. Number one, customization. Those big companies, they don't want one-size-fits-all product. They want something that is adjusted to their specific needs, and they can also, unlike a lot of smaller businesses, they can explain what exactly they need. The second big one is automated workflows.
There are a lot of things that you can do with a small website, relatively easy, maybe with little or no automation, but it's very hard or almost impossible to do them with a big website. Example that I usually like is internal linking, so figuring out how different pages of the website should be connected for, you know, both user experience optimization and you know, to optimize Google crawl budget, so Google indexes your website better. For a small website, if there are 20 pages, you can put all those pages on whiteboard and connect them using Sharpie, and takes maybe half an hour if you know how to do that, so very simple. Now, let's say you have a big website with millions of pages, so there is no whiteboard that can fit all this information.
Actually, even if you try to put it in Excel spreadsheet, it crashes, especially on Mac, and you end up, you know, building some kind of custom solution, paying engineers, and so on. Or you can just buy this automated workflow from Semrush Enterprise product.
Hmm.
And the third part that we focused on was services. A lot of big companies, they need to combine software and services. They want someone to onboard them, to hold their hand, to, you know, give them quarterly review, to explain them what they need to do, how to understand data. And Semrush is, of course, a software company, and we are very passionate about our margin profile, so we don't wanna do services. So the compromise that we discovered is we've built communication workflows inside our enterprise product that helps our brand customers to connect with freelancers and experts and our agency customers to facilitate those services. And as a result, it's much easier for them to get the job done and partner with the right service provider, and we take a small commission for that, and it's all margin for us.
So it's a kind of win-win-win. Customers get better service, our service providers and agencies get more customers, and Semrush gets a reward for this, you know, in the form of...
So, it's data management, it's analytics, it's reporting, it's the internal kind of connection points of the website, and then it's working with these third-party service providers, it's that kind of aggregation?
That's correct.
Okay. You know, so it is intriguing when you throw out numbers like 10x-15x , you know, obviously. What is the size of company that you are targeting? I mean, are we talking about more than 1,000 employees? Are we talking about we're thinking about the Global 2000? Should we forget the number of employees, and just think about the number of marketers that they have?
We like number of employees because it's probably the most objective metric you can measure. Of course, you're right, measuring number of marketers would be better, or maybe marketing budget would be better.
Mm-hmm.
At the same time, I think if you use metric that is not very clear for the outside world, I think it just becomes a challenge to manage. So we do focus on businesses that are 500+ employees. Of course, some of them would have big digital footprint, and they would need more, and some of them might have small digital footprint. Maybe they're a manufacturing company or a logistics company, so their website is not that huge, even though they have a lot of employees. But in a nutshell, I think it's a very good proxy to kind of understand who's a good buyer persona for a particular product. And then, of course, internally, we supplement it with tons of scoring metrics.
You know, in general, I think saying that we focus on 500+ employees is a really good way to frame the narrative.
Okay, 500+. Do we know how you're pricing the product yet? Is it... Should we think of it, you know, is there a price per individual subscriber? Is there a price per... Is it just you have, you know, 1,400 employees, and here's gonna be your, like, we're kind of pricing it per customer? How are you thinking about it?
Right, right now, we just have several subscription tiers based on how much you use, what you use, and how many people you have on the team, and their roles. There are certain roles where it does not really impact price that much, but there are roles that impact price quite a lot. It's kind of like, am I a contributor or just a, you know, viewer? So, those are parameters that impact the price. But for, for a large company, almost every deal, you, you're going to end up with some kind of custom pricing, where it's adjusted specifically to their needs.
And who do you think you would see, you know, competitively as you move upmarket? Because you're wondering if we should think about visualization. Excuse me. I'm wondering if we should think about, you know, the cloud-based data warehousing providers. Should we be thinking about Similarweb? Should we be thinking about something entirely different?
Yeah. I think, you know, when we talked about typical installations, so right now, when we pitch accounts with a new solution, what we see is they have this kind of patchwork.
Yeah.
Like, a lot of things poorly connected, you know, those connections break down all the time, so they need to fix them. Very hard to maintain, you know, especially if they have some kind of time crunch. Maybe they have a board presentation; they cannot pull all the data together fast enough. So people just operating under a lot of stress and using a lot of, you know, solutions that are not bad software. You know, Excel is a very good software, but it's not a purpose-specific software. It's not designed to do SEO; it's designed to do any kind of knowledge work.
Mm-hmm.
It breaks down at certain tasks.
Mm-hmm.
If you have multiple components that are not purpose-specific components, then you have multiple points that will break down when you try to use them for a specific purpose. That's primarily what we are, you know, fighting with, this sort of, you know, entropy of solutions, and we're trying to unify it under one purpose-specific solution. Of course, if you think about patchwork that initially existed, there could be other SEO vendors. There might be a BrightEdge or something else.
Mm-hmm.
And often, as people transition from this patchwork to Semrush Enterprise, they don't need them anymore. But for me, that's more of kind of like collateral damage type of situation. I don't go and pitch against other people. I go and pitch better solution that is much easier to maintain, that provides higher ROI and enables them to do things they couldn't do before, and sometimes that will also would include replacing other people.
Okay. Very well said. I wanna do a quick call for questions in the audience. If you have one, raise your hand. We have one right here. Could you get this gentleman a microphone, please? And then we'll go on to our topic of AI.
So I guess as you're expanding more into the enterprise space, I'm sure larger organizations, right, you can probably charge a premium. But as you kind of balance that with customer acquisition costs, which I would imagine also are higher-
Mm-hmm.
Can you just kind of talk about the profile of margins kind of over the life cycle and how that all balances out?
Yeah. Great, good question. No question that costs are higher in the enterprise, and it's a very different sales cycle. We have played in that space for a while, so we already have more than 5,000 enterprise accounts where we've had to rely on a demand generation engine, a more sophisticated seller that knows how to establish a relationship and build rapport with accounts, versus just facilitating a transaction, a deal desk that has to navigate through legal and procurement, and ultimately, that post-sales customer support model, where there's a bit more of an intimate experience there. So the, you know, costs for sure are higher. Our average for the company right now, we see about six to eight months' payback for our customer acquisition costs.
Ultimately, from a lifetime value perspective, we get about 5x-6x what we're investing from a acquisition cost perspective. While the absolute dollar investment for enterprises would be higher, we do actually expect that we'll be able to maintain it and probably expand the multiple on lifetime value.
... Okay, let's spend. We're down to three minutes. Let's try to cover generative AI in three minutes, and it's gonna be impossible.
Good luck.
So, I think you were one of the first companies we were aware of that was leveraging the GPT engines. I think you might have been doing that before the world had really was even aware of the concept. So you've had these models, you've had ContentShake, you've had AI Writing Assistant, you've had an AI Social Content Generator. You were early with this. Can you give maybe a 30-second example of just what you know, try to make it tangible for the audience, what these products can do for a marketer?
Yeah, I think with the generative AI right now, the main question is not, like, should I use it, but how I use it. And especially considering that people can just go and use ChatGPT, you know, how much better you can do if you use the same API. So what we do is we actually not just use large language models, we also contribute our own data and pre-train those models, or we engineer prompts on behalf of our customers to get better outcomes. So, for example, let's say you wanna write content about some topic. A large language model can write you this content, but it needs to know what to write, how long should it be, what should be subtopics, what words they need to use, what should be readability, and so on.
And you end up with a lot of information that you need to provide to the system, and you don't really even know where to get this information. Well, Semrush already has it. If you wanna create something that performs well in search engines, we can tell you all the instructions, like everything you need to know about this new piece of content you need to develop. And then one way, of course, we can just give instructions, you know, to human writer, or we can engineer a prompt and give it to large language model, so human editor then can, you know, verify, double-check, maybe make couple edits, but will probably be the 10 times more efficient in content creation than before.
We still think that it's important to have human in the loop to verify and, you know, ultimately sign off on the quality of content. Large language models still often hallucinate, but efficiency improvements are phenomenal, and without Semrush data, this content would never perform well in search. And with Semrush data, you get best of both worlds. You get improvements in productivity, you can create way more content, cover more topics, help your customers, and you can create content that ranks well. And one more thing that I really like when engineers show me typical prompt that we feed into large language model. This prompt is actually bigger than the output of the model because you need to clarify so many important things, you know, in order for content to perform well.
That's amazing. Well, that's a great note to end on, and we are exactly on time. So Brian, Eugene, can't thank you enough for being with us. Really appreciate it. All right. Thanks, Mark.