Datadog, Inc. (DDOG) Earnings Call Transcript & Summary

Perfect. Thanks for joining us. We had one small change in the lineup. Yuka, thanks for still kind of doing this. David Obstler is sadly kind of ill at the moment. So we kind of had like a last minute change. Since Yuka and I have been going back for many years, I'm kind of very confident and looking actually forward to that conversation. Yes. Thanks for making the time. Thanks for being here, Yuka. The -- just to -- before we kind of -- I wanted to get the kind of the shorter-term hedge fund questions out of the way. And I know a lot of the ones you will not answer, so a short answer is fine.

Super.

But I wanted to make sure I ask them. So if you think -- and part of the situation is, obviously, you're kind of on cycle and had like all the off-cycle names. And so a lot of stuff happens. And so it's kind of important to kind of ground us all again. So if you think -- so a lot has happened since the Q3 results, like has there been any changes to what's going on? Obviously, the world has changed a lot. We had election and whatever. But any changes from your perspective that you want to call out?

Yes. So first of all, Raimo and team, thank you very much for having Datadog here at the conference. David does send his regards that he was feeling a little bit ill and wasn't able to make the trip out. This is also my great opportunity to say, as we usually do at about this time of the quarter, that we don't give any intra-quarter updates. And so anything that I say is going to be the commentary as of when we reported earnings in early November. And then as to elections and other things that happened, just remember that the revenue at Datadog really relates to the usage of our customers, which relates to their usage of the cloud and how they observe that with Datadog. And as a result, it really is the usage that their customers exhibit on their applications, which won't tend to change based on election cycles and things like that.

So then that kind of makes my next question also redundant but I'll ask it anyway because it comes up like Q4 budget [ flush ]. It doesn't sound then that you guys are really thinking about that.

It wouldn't show up in directly, again, cloud usage or use of Datadog. Maybe it would show up in kind of indirectly and later if that causes people to kick off more cloud migration projects or more applications that they want to monitor with us. So maybe we would see that in that sense at some point but I think it'd be pretty hard for us to see in our revenue numbers.

Yes. And then last thing on this subject. In the early years, I remember post-IPO, David was always really happy about like every quarter, as the CFO, he could sleep a little bit better because overages came in. The -- it does look like customers learn kind of how to kind of size their commitments, better, et cetera. Is that still something that comes up? Or is that kind of a thing of the past?

Yes. This is a great opportunity for me to keep kind of trying to remind everybody the way our business model works. So I think when I hear people say overages, I think investors are thinking about the opportunity to price at a different and higher level on usage above the commitment. I think -- what I want to make clear is that our customers do use above their committed levels by design, that is their choice, right? Our customers are thinking about how much they want to commit with us. They don't want to overcommit, right? So they generally plan to use above their commitments. But our larger customers then are going to commit to us and generally get kind of a rate card, right? A price per unit for the various products that they will use. And that will not change, even if they use above their committed levels, let's say, the commitment per month that -- the dollar commitment that they would give. Their dollar commitment with us can be used on any product at any time. And that's why we -- it's really a usage-based model. But if they use a lot in 1 month, that rate card does not change. And so it's not that there's an overage. It is that they will use above their commitment. And by the way, there are some smaller customers for which there might be that kind of a price changes relative to their commitment. But our larger customers, most of our revenue will not have that kind of aspect to it and it's a good opportunity for me to make that clear for investors. But what is true is, our revenue will relate to the usage that the customers use. And again, by kind of nature by their own choice, they're going to tend to use above the commitment over the period of time.

Okay. Perfect. And then the -- I mean, it almost like the answers -- you almost gave the answer already, like the next question I'm getting a lot from investors is around, if you think about potentially better times ahead, like no kind of 2026 guidance or anything now and the question here intended but how would that show up in your business?

Better times ahead would be more demand and more activity on the cloud, which would be more monitoring and securing by Datadog, would be more kicking off of cloud migration projects, more adoption of new technologies and the things -- and hopefully also more adoption of more products with Datadog. And so yes, part of our growth does relate to conditions in the marketplace, demand conditions. And so if demand for cloud improves, then we would see that as part of our business.

And you mentioned workload migrations because that's the one big thing we're all kind of looking out a little bit. And one of the things is like -- and I don't know how much you guys are using it is looking at the classic kind of hyperscaler performance and like what's Azure core doing or what's AWS doing, if they -- if you strip out AI, et cetera. Is that kind of -- is that something that you guys pay attention to, to think about workload migrations because that kind of new workloads coming to the cloud is going to be a good revenue driver for you. Is that something that you see as a thing that can be correlated? Or how do you think about that?

Yes. We know that investors look at these and it's publicly available data and we know that you guys all correlate it. We would just point out that it's something that -- we think it's -- what's difficult is to know what portion of that cloud -- those hyperscalers revenue relate to what Datadog would observe, right? Because those revenues are going to include all kinds of different things, some of which will not relate to us. So more recently, AI, if you're building a lot of infrastructure, renting a lot of infrastructure for training, that probably doesn't have that much to do with Datadog for observability. Similarly, within those hyperscalers, I don't know, network ingress, egress fees, IPV4 address pricing. There's any number of things and none of us know the mix of those things, right? And so we are believers in the medium- and long-term opportunity that workload migration, that cloud growth represents for us. What we don't know exactly is how that plays out every year and how that relates to the hyperscalers. So for sure, we think that it's a good thing for us if the hyperscalers are growing healthily, definitely. What we think would be difficult would be for people to try to precisely understand exactly how that would relate to our business in any given quarter. So that's where I think people need to just be careful about how they think about that data.

So it's more like one is the long-term trend, one is like a short term kind of...

We're very enthusiastic about the medium- to long-term trend, right? We do see it as a secular underlying driver of our business. But again, like how it plays out quarter-to-quarter is probably where it gets a little too difficult.

Yes, yes. And then you mentioned one thing. And I have to say we were all -- say, positively surprised but yes, surprised last quarter when you kind of talked a little bit about AI and the AI contribution. I mean, conceptually, from my perspective, I would have seen -- expected from you something actually later in the AI adoption cycle because -- and it's more like AI apps coming into production and then you want to monitor it. You gave already numbers, like can you maybe speak to those numbers and why you give them out?

Yes. So I agree with you, where we think we'll see the bigger opportunity in AI is when the broader group of our entire 29,200 customer base start deploying AI as part of the service to their customers. That will be more cloud usage. We'll see that and that's where we've been communicating to all of you that there's this opportunity that we think is one the come. However, it is also true that cloud-native customers tend to -- we think we have a really good opportunity to monitor and observe and secure workloads with cloud-native customers, including AI-native customers. So there's definitely a hunger and interest in data around AI, obviously. We do think it's still pretty early on in that first broader trend of everybody getting to production. We think it's pretty early days there. But we can also look at our customer base, try to identify customers who are AI native, who are -- their entire business is delivering models, tools, services that are next-gen AI. And what we've been telling you is the ARR that those customers represent. Now that will mean all of the ARR that -- all of the things that they're doing with Datadog, which is typically the observability and security that they're doing with us. Like many of our customers are going to take multiple products and understand how their systems work. So that's the nature of that. We've been giving that stat for 4 or 5 quarters now. We have noted that it's grown pretty healthily. So clearly, at least some of those customers have gotten into production, are using a lot of cloud and are monitoring that usage with Datadog.

So in a way like what we see at the moment as the revenue is more like an OpenAI or someone like, not to name customers but like the AI specialists basically kind of using you. But it's not necessarily Barclays doing some AI projects and kind of using Datadog for that one.

Probably not, right? Because again, like there's some people who have gotten there with in production. But we think a lot more people are still experimenting. So just to give you some data points around that, right, we told you on our last earnings call, hundreds of our customers are using LLM Observability. But actually, we also told you about 3,000 of our customers are sending us some data about AI via one or more of our AI integration. So again, a lot of people are looking at it but in terms of people really using a lot of volume of LLM, it's still something that feels like it's in early days.

And how do you think about that like conceptually more longer term, like the AI opportunity for you, is that just inference workloads that are in production and you need to do that? You had like LLM Observability, as kind of an area, like how do you think this is kind of playing out for you?

Yes. So in general, people use Datadog for their production application environment, for their mission-critical, customer-facing, high-volume workloads, right? So that would relate to that inference type of workload in the AI sort of environment. AI training would be more analogous to test and dev type of situations where you're not yet using Datadog but you're in kind of preproduction type of phases, so yes. So for us, the opportunity will be, again, when people get to production, they start including AI as part of the service delivery to their customers and that's when we would see a lot of that usage occur.

And then the -- are those -- a lot of them -- those AI workloads are -- from what you see, most of them will be cloud because, I guess, it has so much compute, which kind of should favor you over like a one of the classic or the old observability [indiscernible] that were more on-premise?

Well, we think these are going to be cloud applications. In some ways, although this one seems to be a bigger, more durable trend, this is just another new technology, right? And our customers come to Datadog because they want to deploy to new technologies with confidence. And so -- and that will be true with this AI part as well. And so in that sense, for us, we would hope that, that kind of customer would want to use Datadog and we would expect that they will want to build that on the cloud.

Yes. Okay. Perfect. And just moving a little bit away from the AI, like -- if you think the one big theme that played out for you over the last few years, like how much broader your product set did get, like how much more -- how many more sources of revenue kind of you're realizing, can you maybe speak to that a little bit? Like and -- you and I go back quite a few years when Datadog, like when we thought about Datadog as an infrastructure monitoring vendor kind of, the observability but then actually there was more observability need now or it's really observability. Can you speak to that evolution a little bit?

Yes. So for those of you who were around 5 years ago when we were IPO, like Raimo said, it was primarily infrastructure hardware monitoring. And I think the jury was still out on whether or not we could be successful as a multiproduct platform company. These days, I think about 1 year ago, we told you all that our infrastructure monitoring product had exceeded $1 billion in ARR but also our APM suite and log management had each exceeded $0.5 billion in ARR. And this last quarter, we told you that those 3 pieces together are now over $2.5 billion of ARR. So we have continued. And I -- and we also give you multiproduct adoption statistics every quarter. I think you guys can get the sense that I think we feel good that we have developed a platform that our customers are using fairly broadly and getting value out as a platform. So today, we have 23 products, many of which are in observability but we talked at our February Investor Day about cloud security, developer experience, software delivery, product analytics, cloud service management. So expanding beyond observability into these new areas. And so it's a very exciting thing for us to solve more problems for our customers in more different areas now outside of observability.

Yes. And then the -- if you think about it, like it did get broader, like last quarter, I remember on the call back, I was talking with your CEO about like security event management and things like that. It's like, if you sit together with the team and think about a product evolution like, is that kind of how does it feel in terms of like what's still out there that can still be covered in terms of how broad this can get?

I think that we think of the opportunity is quite broad for us in that there's a very large problem set that our customers face. And if you want to think about Datadog more broadly, you can think of us as a real-time data platform that is meant to be able to take in all sorts of data from wherever you want to send it to us from. We have over 800 integrations. We have an agent. We can accept OTel data. And then that we can take that data and correlate it and analyze it and visualize it and express that data in many different products, right? And so if we think about that as what Datadog does broadly, you can see how it might be a fairly natural idea for us to extend beyond observability to provide that same analytics in security, to provide the ability to take action in cloud service management, to extend the analytics to product and business analytics. And so those can be in that frame sort of fairly natural pathways for us to expand our -- what we do for our customers.

Yes. And if I think about your data footprint, I mean, there's still a lot of like vendors out there that started before you, where there's a lot of on-premise and like crazy amounts of data still sitting there. Like -- and it gets me to my next question, like in your industry, there was a fair amount of consolidation over the last few years and some very large vendors like Splunk, for example, kind of went away as public kind of competitors. How do you -- how did it play out so far? And how do you think about that conceptually in the long run?

Yes. So the opportunity to consolidate to Datadog has been going on for many years now and we hope it will keep going on for many years. Hence, to this point, in the past, we were consolidating in the observability space. And in particular, as we added more products in observability, we had this opportunity to add more value to our customers and consolidate against those products. And more recently, we've had opportunity to consolidate against new use cases as we've put out more products like Cloud SIEM, for instance. And so the consolidation, I'm sure there's maybe tactical opportunities but the overall dynamic of us having these consolidation opportunities is probably really unchanged. And by the way, there are still competitors, right? They've just changed ownership hands. And then the final thing to remember is our -- the customers will have multiyear deals often. So it's not like just because they change ownership hands all of a sudden, there's just like there's this burst of it, right? I think the opportunities come when the customers are ready to consider other partners. And hopefully, at that time, our product set has expanded and improved in a way that we're going to be part of that conversation and we have the opportunity.

Yes. I mean, like on that, you mentioned product set because like there's like new things like Oracle Database, Oracle OCI. Like I mean it's like where you think like, oh, yes, you should have had that but actually, there's only something recently like -- is that kind of -- like even if you look at the basic stuff, there's a lot more you can do with these.

There's always this long tail of technologies that our customers ask us to monitor on their behalf. So to Raimo's point, we recently announced for general availability, Datadog monitoring for OCI. So what that means is at least there's some number of customers who wanted to be able to do that and we want to be wherever our customers want to deploy. So we've done that. Our database monitoring product today monitors 5 different types of databases, including Oracle Database. But I guarantee you there's any number of other databases that our customers would like us to monitor and we'll -- I'm sure we'll keep extending that over time. And even our oldest products have this, right? And then new technologies pop up and you want to be able to monitor those.

Yes. And then talking about -- sorry and I kind of should have asked that one question before that. Talk about the -- if you think about the legacy guys and thinking about them as a migration opportunity, is there -- was there anything you guys did internally in terms of sales, et cetera, to try to kind of expose yourself better there or like kind of maybe work on these new logos?

Sorry, in legacy like on-premise or...

No, no, like in terms of the big accounts that are sitting out there that are sitting with legacy vendors. Is there anything you guys did like in terms of having new sales guys that are kind of hunting those logos or something?

Sure. So I would say -- I think I'll put this in the frame of our overall expansion of our enterprise sales capability. We have continued to expand the sort of sophistication and capabilities of those teams. And so several years ago, we stood up a major accounts team. So fewer accounts for reps to make sure that we covered those larger customers well and explored expansion opportunities where there might be new business units that have never used Datadog. There's more products that we could help them adopt and they're focused on that. More recently, this year, we stood up a key accounts team. So the idea there was for prospective customers for whom maybe there's a multiyear sales cycle. Maybe it is a more top-down sale. We tend to be more bottoms up, where we want to support our salespeople and also provide intermediate targets for them to go after that type of customer, which may be our previous enterprise sales teams were kind of structured and supported in a way that went after those the best way. So we continue to go after those types of opportunities as well. At the end of last year, 42% of Fortune 500 companies were our customers. However -- so that's good. We continue to make progress. I think the year prior was 37%. However, the median spend on Datadog of those customers was less than $0.5 million. So we think of those largest enterprises as still very large opportunities for us, both the ones we don't have and the ones that we do have that we could do more for.

That was my next question because like the one thing I do notice when I talk to guys in the industry is that you do show up more upmarket. And it does -- is that kind of something that, from your perspective internally, that you wanted to do, that we kind of -- yes, we did bottoms up and a lot of guys are using it but we could do a better job. So you talked about the sales changes as well but your product got more powerful as well. Do you see that in the numbers that you're kind of -- you're doing more stuff upmarket?

So the number of our $1 million customers continues to increase. It was -- we give it yearly -- at the end of last year, it's 396. Some of those are $10 million annualized spenders and we obviously hope to grow both of those numbers over time. But I think we're just always trying to meet the needs of customers and our customers have gotten larger and more sophisticated and those larger legacy customers have moved to the cloud, right? So we are -- have capability to be PCI compliant, or HIPAA compliant. We have set up a GovCloud for our public sector companies and have FedRAMP Moderate authorization. We have capabilities that really only the largest customers need, like role-based access control. Our -- one of our newest capabilities in log management, Flex Logs, is really for just the biggest customers who have a lot of logs, right? And so I think as we've heard that need from our customers, we're seeking to make sure that our platform is ready for -- ready for that, so that we can meet their needs and be ourselves more sophisticated against their advances.

And the other thing I wanted to do and there is like to show the success you have already. I threw out a number yesterday and you kind of answered it really well. It was like, while the Dynatrace ASP is $400,000. Yours is like $100,000. So there's still a lot of room but you gave me a number that actually kind of trumped that looking at the bigger accounts for you guys.

Yes. So we also give you guys every quarter the number of customers that spend $1,000 or more annually with us. That -- last quarter, was 3,490. They represent 88% of our ARR. So if you play around with the numbers, you're going to find that the average ASP of that group is about $700,000, I believe. And so we feel like we do quite well with enterprises but we want to do more as we've talked about. And also just remember that it has not been a winner-take-all market in observability, right? Enterprises notoriously will often have dozens of tools that they use in observability because, historically, it was a product market space, right? I think Datadog was on the vanguard of developing a platform that included many of these types of products. But even still today, it's clearly something that enterprises do. They will use not only Datadog somewhere in their environment, Dynatrace, Splunk, as well as open source, hyperscaler cloud tools and new things that come across, right?

Do you see that in your statistics, if you look at deal statistics, in terms of like we would think out there there's your top 20 deals and it's going to be like, oh, Splunk, Dynatrace and Datadog will show up all the time. Is that really what you see in reality?

There's definitely going to be some competitors that we see more at the enterprise level. And then when we look at our commercial or smaller customers, we'll see some different competitors certainly. But it will also be that we will see in that deal, let's say, in that expansion, we're trying to get into a new business unit and sure, we displace a previous competitor there. In the next renewal, there will be another business unit over that, that same competitor was still in, right? So it's very common, I think that there's multiple observability tools represented in a customer.

Yes. Okay. Last couple of minutes, I wanted to switch gear a little bit. So you've been able to grow at a healthy clip and it looks like that growth is now stabilizing in that kind of mid- to high 20s. But at the same time, you still show like good leverage. Can you think about like how you're threading the needle there in terms of like doing both, like kind of making sure you're kind of growing but -- and maybe get ready for better growth even and still kind of have that profitability, how do you plan that?

Yes. So I think looking back over time, Datadog at IPO had about a 0% operating margin back in 2019. Since then, we have balanced our intent and desire to invest aggressively against our opportunities as well as delivered financial performance in the form of margin improvement. And so today, our margin is somewhere in the mid-20s. We also, in February established a long-term operating margin -- non-GAAP operating margin target of [Technical Difficulty] long term, no timing on that. We're obviously kind of already in that neighborhood. And then by the way, we also expressed in the near term that in 2023, we were a little more cautious with our head count growth. And that in 2024, it was our intent to get back to investing to the medium- to long-term opportunities. And I think you can all see that in year-over-year OpEx growth accelerating. And so I think we have tried well to execute against both of the sides of the coin on growth and margins. I think that's our goal over time still as well but that we are -- what we are mindful of is looking to maximize our long-term opportunities. And that means that we are going to keep investing as well.

I mean how do you think about like you just kind of went through the budgeting cycle, and it's more of a David's question, so I apologize. But if you think about -- is profit and outcome, like in a way you could think about growth, there's your OpEx that you build to that. And then if you do better revenue, the OpEx is built and you get ahead of profitability, like what happened in 2023? Or like how do you think about that as you start the year?

Yes. So I think as a usage-based model, we don't know exactly how our revenue will turn out. So there's going to be different scenarios at play. As far as our guidance to the financial community, we're always -- as a result, we're conservative because we don't know exactly how things will happen. We do have to budget and plan for our expenses. So we're trying to grow appropriate to what we see. And again, trying to balance these dual goals of growth and profitability. And then we can always change our minds, right, as the year goes on and as we reflect what's happening. And so I think that the good thing for Datadog is we have grown at a fairly robust rate over the years, which has supported our ability to grow our investment. And so in general, that's been the case for us, right? So we will continue in general to pursue investment to support that growth so that we can keep growing our investment and so on.

And it's like how is the -- I mean, David feel -- to me feels like he's the old statesman in the organization that kind of have to say no a lot of the time in terms of the -- if you think about investment returns that -- I'm sure the R&D organization or sales wants to invest in that [indiscernible] How do you see him in that?

I think our whole leadership, including David, has done a great job on this. And I do want to kind of bring it, although I do think that David does a great job in kind of grounding everybody in the aggregate sort of needs of the business because I think you're right, any given department will definitely have ambitious goals themselves. But I would say that all of leadership is mindful about our investments and really trying to get a return on our investment. And this is kind of part of our origin story really, our founders who are our CEO and CTO today, they came out with this idea to develop this observability platform. Not every VC thought this was a great idea. It wasn't necessarily easy for them to fund raise and they didn't want to be dependent on the next funding round. So they were very mindful from the very beginning about spending wisely, about investing wisely. And I think from founding of the business to IPO, I believe they burned less than $30 million in cash. And so by the time we went IPO, we were about operating margin breakeven, I think about free cash flow breakeven and then we've improved on that over time, obviously. So I think David definitely does a great job certainly in helping our teams think about this all together but I think our overall leadership has always been mindful about not getting over our SKUs on investment and being conscious of the fact that we want those investments to yield.

And then last question for me and then I need to let you go. You just did a convertible. Like how does that fit in? Like what was the idea? And how does it fit into your overall capital structure?

Yes. So for folks who aren't aware, we issued a new convertible debt instrument on Monday, $870 million with a $130 million green shoe, 0% coupon, 35% conversion premium. So just as a reminder, we had about $750 million par convertible that is going to mature in June 2025. We retired a little bit of that in conjunction with this new issuance. The rest of it is outstanding and we will need to retire that. So you can primarily think of it as being related to that. Overall, though, the convertible debt allows us -- just gives us the financial flexibility for our leadership in case they want to invest.

Okay. Good. Perfect. Yuka, thank you. I really enjoyed our conversation. Thank you.

Thanks, Raimo.

Thank you.