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

All right. Great. Thanks, everyone, for joining us. My name is Sterling Auty. I'm a software analyst here at JPMorgan. Thank you for joining us for our 48th Annual -- I can't believe it's 48, and I think I've been involved in 18 of them already -- Technology, Media and Communications Conference. Kicking off this track, we have Olivier Pomel, who is CEO and Co-Founder of Datadog. Oli, thanks for joining us. We really appreciate it.

Thank you for having me.

All right. Listen, just to get started, maybe from the highest level, give us just a brief overview of Datadog, what you provide and where you compete for those that -- I'm sure there's still a few that maybe are not as familiar with your company.

Sure, yes. So what we do is provide observability and analytics. And by the way, if you can overhear the toddler in the next room, I apologize for that. Let's just go around it.

As a father of 4, we've all been there. And if you hear my golden retrievers barking in the background, same thing.

Yes. So as I was saying, we do observability and analytics for applications and infrastructures at running cloud environments. And the good news is that basically every single company, big and small, is ongoing in the cloud migration at this point. So we sell to engineers. We start serving enterprises when they get to initial footprint in the cloud, and then we grow with them to serve all of their observability needs across infrastructure monitoring, application performance monitoring, log management and now security as well, security management.

All right. Fantastic. And just as we get started here, I should have started with -- in my little intro. [Operator Instructions] So we had compute workloads moving to the cloud. Do you feel like we're still in the early days of those cloud workloads being implemented or instrumented, I should say, with even things like your infrastructure solution, which is kind of the core of where you guys started?

Yes. So just to backtrack a bit. And as I mentioned earlier, we start with customers as there is some initial scale in the cloud, and then we go with them as they move more and more workload. So to a large extent, we track our customers' cloud migration. I think it's still very early in this cloud migration. The most companies still have their -- the bulk of their systems in legacy IT and on-prem environments. And in parallel, this cloud migration, what's happening is the broader digital transformation. So what's interesting to understand there is that when you think of digital transformation and the footprint and the impact of these applications, we're going to a world where DC is a lot larger, I guess, a lot more footprint with more impact. So instead of being a transition and migration from 1:1, it's a migration from 1:10 or 1:100. So it's going to be very interesting to see as it develops. I would say also that more specifically about Datadog, we're still fairly early in our go-to-market. So with at scale in all segments and all geographies, so for example, we barely started in government at this point. So generally speaking about the market, very early and more specifically about that given earlier. So we're basically at a stage where we're investing as much as we can, and we see a lot of growth ahead.

How about when you think about your APM solutions? What about penetration there in terms of the work -- those workloads into the cloud and your penetration in terms of instrumentation of those solutions?

Yes. So it -- I would say it's even a bit earlier than infrastructure. I mean, for one thing, it's a product that's more recent for us. It's only a few years old at this point, even though it's already got scale and it's going very fast. I would say, if you were to compare the adoption trends of our various products, if you think of the 3 largest products we've released so far, I mean one is infrastructure monitoring, one is APM and one is log management. Infrastructure monitoring tends to track the cloud migration pretty closely. APM tends to start a bit more slowly and have a longer fuse because you actually need those applications to reach some scale and to be instrumented to appear at APM. And the log management one can be a bit lumpier because companies can switch everything to a new log provider from one day to the next, and that gives us a lot of growth very quickly with some of the customers. When you think of the target for APM in terms of the penetration, it's a lot higher in this new digital and cloud world than it is in the traditional Gartner [ court ] of 5% of workloads billings [ to manage ] with APM. And that's actually what we see today is already a little bit higher than that in terms of what we see from our customers, and it's growing very steadily. So overall, earlier in the infrastructure, a lot more growth in a destination where APM is going to be fairly central to businesses.

Do you think that there's a big difference in the size of the overall opportunity in terms of your revenue opportunity, infrastructure versus APM versus logging versus some of the other newer areas?

No, we think of those -- in general, we -- so we don't try to slice and dice too much within observability because we think quite a bit is going to be fungible in the end. So even though we all think in terms of the existing categories, for a customer, it doesn't really matter if they sold a problem using something that's part of the infrastructure products or part of their log management products or part of the APM product, and they become more and more interconnected. That being said, I mean, we -- the 3 categories, you can think of them as being roughly equivalent sizes. I mean infrastructure is probably a little bit bigger. But give or take, they're the same order of magnitude. When you think of the other categories we've entered, so we started developing products in security, security monitoring and protection. And that market, in and of itself, you can think of it as being the same order of magnitude that has the observability as a whole in the long term. So that's what we think about these various components.

One of the questions I get from investors is they're trying to figure out the competitive moats and the long-term competitive advantages. Which one's harder? So as you think back in the development, if you had to start again, which of the solutions is the most challenging and why?

Well, if I had to start again, and we'd do it the same way. I think -- I don't know if it was timing a lot, but I think we've got the mix right from the beginning. And the way we've done that is we've built a product that was built, one, to get as much data from as many different data sources as possible, which means it was very easy to open into more use cases as we grew. But also, we built a product that was really you can see from day 1 to be adopted by as many users as possible [ with the company ]. So we didn't build it for 5 specialists who need to be heavily trained. We built it so that all of the 5,000 engineers can spend their day basically in the product. And so when we combine those 2 things, we get that out from all those different places. And thanks to our infrastructure product, we can deploy it everywhere. And the fact that we get used by every single engineer every single day, that's what gives us the competitive moat. And that's what -- that gives us all that [ surface ] contact with our customers and lets us solve a bigger and bigger problem for them over time and grow. And that's what we've proven over the past few years since we've taken the company public, is that we can actually solve more problems, build more product. And with those products to be adopted without any friction by our customers.

So if you look at infrastructure, in particular, one of the worries is that, geez, why don't the cloud providers themselves, the AWS, the Azures, et cetera, have capabilities? And I'm finding that maybe smaller companies, when they first move to the cloud, maybe they'll use that. But why aren't they a bigger competitive threat as companies get larger?

Yes. So I mean, they've always had these consoles and these things that live inside of the cloud. And it's been the [ history ] for service providers ever since there have been service providers. I would say there's at least one structural reason why it makes sense to go to a company like Datadog to do this. And that's most companies will have several service providers. They will have hybrid deployments where they use a few clouds, but also maybe they have some on-premise deployments. And they actually don't want to be -- to run those things in different ways depending on where they run. So that gives them the independence and the, I would say, the cross-cutting integration that they need. And in general, even if cloud providers say they're going to give access in [ one or the other things ], they tend to be very myopic to the things they build themselves. And that goes not only for the services they operate. Say, for example, if you're AWS, you're going to be [indiscernible] versus Azure, but also to the software that runs on those services because not all the software -- actually a very small fraction of the software you run on a cloud provider has been developed by that cloud provider. So that's, I would say, one structural reason. I would say more practically in terms of how we run the company is -- when you think of the way the cloud providers are structured and the [ products they sold ], they end up being built to create all those APIs and let the customers act as the [ blue ] in the middle of it, and that's the way their teams are structured internally with the opposite. Our job is to actually take all of those different PCs and bring them together and [ paying ] something that is consistent and makes sense and make it as accessible as possible to as many users as possible. It's a fairly different job from what all of the cloud providers do. So they keep providing building blocks and we keep integrating those building blocks. But in terms of serving the end-to-end problem, we are fairly differentiated in what we do.

Is there -- when you look at -- we're seeing customers that, to your point, have pieces still in private cloud but have pieces in multiple cloud vendors. Is it important to give like a single pane of glass where you can almost red light, green light across all of your infrastructure in those environments? Is that something that is a priority for customers or not really? Is it still kind of managed a little bit more granularly?

Well, it's important because it reduces complexity. And at the end of the day, every single customer is on the race against complexity. So there -- as their businesses become more digital, they end up building more applications to have more scale in those applications, they have more engineers, they make more changes more quickly. And as a result, the complexity is shooting up. So our job is to reduce that complexity for them. And you can see a complexity when it comes to operating and building software, you can also see it when it comes to securing the applications. That's even harder and that's where the -- we said it's the next frontier in the race against complexity.

Got it. Let's switch over and talk a little bit about APM in terms of what's kind of the competitive dynamics or what's the competitive advantage? Because it feels like APM is one of those that takes a little bit longer to mature to get deeper knowledge and capabilities within each of the development languages. Will you tell me, is that the right way to look at it? Or is there something different that it really kind of drives?

I mean, look to -- for APM, it's typically easy to have a product that has a very narrow applicability. So basically, you deal with a specific version of a specific language for 7 [ cloud ] companies. You can do something that's very useful fairly quickly. It's very hard, and it takes a lot of time to actually make that work across all of the languages and all of the run times and all of the types of customers and classes of problems. And as it turns out, large enterprises end up having a collection of just about everything. So they need products that are broad, that work with all of their environments for that to make sense and for -- the promise of getting the complexity to be fulfilled for them. So I would say over the past many years now, we've actually been developing and investing in the APM product. And we've taken it over time from -- a product was fantastic but had narrow applicability to a product today that is still fantastic but not works for just about every single situation or [ customers ] we come across.

So how does -- that says a lot because I think it changes the competitive dynamics in terms of how customers will look at you relative to Dynatrace, correct?

So without speaking to specific competitors, I think the product -- we said in the -- actually in the last earnings call is that the product is very mature today. It's a product that is at a significant scale, that is growing very fast. We had disclosed in the last -- on the last earnings call that the -- if you combine our APM and log products, they added more AR in the last quarter than the whole company the year before. So this speaks to the scale and the success. And if you also look at the gross retention of those products, like the gross retention of the APM product individually, it's about the same as the gross retention of the company as a whole, which is in the mid to high 90s. So this tells you those products are mature and sticky and they actually bring a lot of value to our customers.

Security is one of the modules and newer areas. What is it that you would provide today in security? Where do you envision kind of directionally taking Datadog?

Yes, the security is -- right now, it's an area of investment, it's something we're building. So it's not a meaningful part of our revenue today, although it is growing very quickly, and we're very happy with the traction of the security product. But today, we're in a mode where we're trying to make those products really excellent with a fairly narrow applicability still and also similarly to what I described earlier for APM. And we're in a situation where when it comes to getting it commercially adopted by customers, we're going from the customer pool as opposed to pushing it ourselves. The products we have today handle a few different areas of security. I mean we think there's a few building blocks. You need to have a credible cloud security offering. The first one is the system of record and analysis, which is our security monitoring product that does threat detection. And is -- we're building it as a seam. Basically, it's not a full seam yet, but it's -- you could put it as a cloud [indiscernible] seam in the future. There's another building block, which is infrastructure security, and we'll have 2 different products there, one that secures the actual workloads, one that looks at the cloud compliance and posture -- the [ security ] posture. And then there's a third area, which is application security. And for that, we've actually announced an acquisition 2 quarters ago that we're in the process of integrating today, which is going to be called [ screen ], and that does security and application. And what's very interesting with it is that it's a security product that plugs into the application exactly the same way the APM does. So for us, it's a very, very straightforward extension of what we have today, and it's a way for us to bring security to our customers with very, very little friction.

We had a question from an investor here that just wanted to understand how those security products compare to CrowdStrike in terms of securing the workloads?

So this is a great question. So we are in a completely different space. So we're not dealing with end points. We're dealing with cloud applications -- sorry, production applications, security in the cloud. So we're not going after the corporate market. We think -- but the corporate market is mostly fully assigned today. So it is a -- anybody who got a seam for their -- what's happening on their corporate environments and their e-mails and things like that. And we're not trying to disrupt that. We're trying to go after the cloud environment, which are production applications, and this is fairly new. There's no winning product there. There's no -- it's basically wild west and it's completely open. And we think that market is going to reach some form of maturity in the next few years. And what we're doing today is we're building to have our products ready for when that's the case.

But I think -- so the heart of the -- this person's question is, CrowdStrike also talks about -- a big part of their growth is actually securing the cloud workloads themselves are not corporate end points. So if you have an application or a virtual machine running in the cloud that they would actually be inside that instance protecting it. Do you want to be the malware protection inside? Or -- because I think this is an area that investors, to your point -- it's so very new in terms of cloud security. They're still trying to figure out what's what versus on-premise, people understand, "Oh, you've got firewalls. You've got end point." The cloud part, I think, is still very under -- misunderstood.

Yes. So I would say the way we present what we do to give you a bit more idea of how we differentiate from the rest is that our goal is to bring together security operations and development and unify these 3 teams' approach to building and securing the cloud. We think the biggest problem in security today is that those teams don't speak. And we actually think that it's very similar to what was the case between dev and ops 10 years ago when we started Datadog and the problem is [ set up to solve ] initially. So that's where we start. What our strength is, is that developers and operations folks are on our product all day long, and all of them on our products all day long. So if we can weave the security use cases into what it is they're doing all day long in there, we can solve that problem. And that's fairly different to what you'd expect from a malware management platform. Like developers and operations folks are not on the malware management platform all day long. So the kind of problems you will solve from that perspective are going to be fairly different from the ones that we can solve at Datadog with our integrated platform.

There's an awful lot of marketing and press around CPSM, so the cloud posture, the configuration management. It feels like there's a ton of companies that are offering something. Do you think that's actually going to be a big area in terms of an important revenue generator? Because so many companies are doing it, it becomes a check box and all the value comes in that holistic approach that you just described?

Well, I think the comment you made applies to pretty much all the businesses we're in, right? So we -- when we started the company with infrastructure monitoring and when I was fundraising for Datadog in 2010, 2011, anytime I say infrastructure monitoring, the [ same times ] I would come back was in the private market because there were so many different companies. And when you look at the marketing websites of everyone, it looks like everybody's doing the same thing. The reality is there's a difference between having -- saying you have a few capabilities on the website and actually delivering on that value and having everybody use it and impact change. And I think our ability at Datadog is to do that. And that is built into the -- that comes from the 2 things I mentioned earlier, which are -- we build to be deployed everywhere, and that's what we achieved with our infrastructure monitoring product. And we build to be used by absolutely every engineer every single day. This is how we get to it.

One of the other questions that came in from an investor. [Operator Instructions] Just very generically, when you look across your solutions, especially infrastructure, what kind of moat does Datadog has?

Well, I think there's a couple of things. I mean, the -- again, looking back at what I just said, the fact that we get deployed everywhere and that we can use by everyone, nobody actually replicates that. It is a key differentiator, and it has to be -- it has to do it not only with the way our product is built as a broad data integration platform first, but also as the way -- it comes back also to the where the company is built itself. When you think of getting used by everyone or getting adopted by everyone very easily with our professional services, we actually built a company so that we sell have a very wide customer base. We have -- we serve everyone from individuals and students that compares anything on one side, all the way up to the largest enterprises that pay us more than $10 million a year and will have everything in between. And obviously, the bulk of our revenue, as we've disclosed before, comes from the larger customers, right? So the bottom half of our customers only account for a few percent of [indiscernible]. But the reason we serve them is that they pull us towards simplicity. They keep our product and our [ usable banner ] models, which means that those 10 new models in large enterprises can get to use our products on their own, they can adopt it. And we can, in the end, deliver the full value to give enterprises for that reason. So that's the, I will say, the core differentiator in the way we build our product and the way we've built the company. And in addition to that, because we're fully SaaS, we have incredible access to the data basically that our customers send us so we can use it to train algorithm and get smarter and smarter over time, which is one where we'll be innovating and building more and more for those customers.

Let's switch over and talk a little bit about some of the growth drivers. How should we be thinking about the growth from here split between new logos, existing customers using more as well as cross-sell, upsell into additional modules?

So the way -- every single quarter when we talk about our results, the picture is currently similar in that. About 1/3 of the growth comes from new logos and about 2/3 come from existing customers and their growth of the products that they had adopted in the past. But you have to keep in mind that every single product -- even when we add a new product for existing customers, that product is going to start small and it's going to go over time. So our model is heavily land and expand.

So heavily -- I'm sorry, say that last part again?

Heavily land and expand.

Absolutely. So how often -- I guess what's kind of the catalyst for that expansion? What do you see when -- how frequently does the customer come back? And what -- are they saying, "Oh, I need more because of this"?

So the main catalyst, like the broadest catalyst is digital transformation and cloud migration. And that's what is going to drive our customers moving more and more of their infrastructures and application and data from whatever they had in their legacy data centers to these cloud environments. And by the way, to a large extent, we can fully control that rate of migration like this is a broader industry trend, which last year led to a surprise in Q2, where the growth was slower than we thought as everybody was scrambling at the [indiscernible]. But as I said earlier, we're still very early in this transition. We think it's actually a transition that we haven't seen since the industrial revolution, and I think we [indiscernible] of it. So it's -- we see that continuing for many, many years. Beyond that, as I described earlier, we have this position where we have incredible [ surface ] of contact with our customers, where we deploy everywhere already, we get all the signals from them and were used by all their engineers. So it's only logical for us to solve the bigger and bigger problem over time. So as these customers realize, hey, we also have a problem understanding our networks. Or hey, we also need to understand the behaviors of our users or hey, we want to understand the security threats that relate to this part of our application. They naturally come to us to expand and add up more products from us.

When we talked about the beginning of the pandemic, I think we talked about it at that time as overages. But to your point, I think overages normalized. What have you seen -- what are overages and have they normalized? And can that be something that we see pop back up?

Yes. So we don't actually have a concept of overages. We have a lot of usage. But the usage is actually recurring in nature. And like when you think of overage, you think I went over my plan, I need to reduce. In our case, the usage all relates to things that are going to come back in the future. There are cloud instances that are up, their applications that have been deployed and things like that. So what we do see though is that the usage is growing in ways that is not necessarily planned. Nobody understands ahead of time at which peer, at which rate they're going to move into the planned environments. And pretty much systematically, all customers are going to under-commit when they start this migration. And then they're going to outgrow their commitments, and they're going to recommit over time. That's what we've seen throughout the history of the company. What did happen last year is that the growth itself slowed down as opposed to keeping to the rate we had historically and then reaccelerated after Q2.

Got it. A question for -- investor wants to know what trends have you seen in pricing? And maybe I'll add on to that. Is there any difference in pricing trends across the different areas, infrastructure, APM versus log then?

I mean, overall, there haven't been any changes in pricing. And we keep track of the -- which [ outflow ], there's no pressure with retail. The other data point I can give you there, I can point you back to the gross retention numbers and the room for visible disruption there. The thing I will say, though, is that when you think of the parts of the business that are priced per gigabytes or things like that, it's guaranteed that 10 years, 20 years from now, the volumes are going to be [ orders of managing ] more than that. Many others of managing more and when those are -- prices are going to be different. There's no way around it. So the way we see it is maybe the price per gigabyte are going to go down over time, but the wallet share of the customers are going to go up as the footprint and the impact of this application goes up. The way we think about our pricing in packaging is that it has to align with the value our customers get from it. And our job is to deliver more and more and more value. So we basically get a larger share of their spend.

So how has, if at all, changes in pricing from like Splunk and from New Relic to very different parts of the market in terms of what they're offering -- but how would those changes impact your business, if at all, even just the conversations with customers?

So there's no impact that we can tell today like there's no -- I would say no more impact that -- what you get every year when you have conversation with customers where you get competition and things like that, so -- but just more and the same. As far as our business is concerned, like we haven't changed anything. And the way we think about it, though, is not in terms of what [indiscernible] are doing today but in terms of where the market is going and where do we want to be 10 years from now and how do we get there? So the pricing is not the determining factor in those decisions.

Back in the fall, you had announced the partnership with Microsoft. How is that progressing? And what do you expect to culminate from that arrangement?

So it's progressing. I think we -- it is almost live. There's still a few things that need to be added. Technically, a lot of it is live, and we have great [ proof ] points from customers. It's exciting for us because historically, the agile environment was not the ones that we added the most value because if you look back at Azure in the early days, it was very Microsoft centric. And you probably didn't need Datadog to monitor your share point on Azure. But this is changing very quickly. The Azure platform is now turning into a really general purpose cloud platform that is adopted as an equivalent to the AWS and GCPs of the world. And so for us, it's an occasion to really get deeper into that ecosystem and be a first-class [ season ] on the Azure platform. We -- there's a lot that we can do for it. For us, it's all upside, and it's not something that's baked into any of our numbers. But we see a lot of potential upside, and we have great proof points from the early customers that have been to this integration.

Do you think that will allow you to get an outsized market share for the Azure platform because of that relationship?

Well, it is what we're looking for, right? So we always want outsized market share and we're -- I think at the end of the day, this has to come from having the easiest, the best experience for customers. We did see though -- I mean, we actually talked about a few of those on the last earnings call. We do have customers that even if they don't go through the specific Azure integration to get to Datadog -- are extremely happy to see us being a first class [indiscernible] Azure as it aligns with their strategic plans to have Azure as being their first cloud and the others as being their secondary cloud. So we've mentioned the supermarket chain, I think, on the last call. There's others in that situation where we're very happy with that.

You mentioned -- you announced last quarter that your head of sales had decided to take a long needed rest after, I think, being out on the road for 27 years. What changes do you see happening in kind of the go-to-market and how do you minimize any type of disruption as you go through management changes?

Yes. So we're not actually looking to make any drastic changes. From where I stand, we have a great team. We have a great process. We're growing very fast, and we're doing so with an efficiency that few companies have when it comes to going to market. Really, what we're shooting for is scaling the team. So we need to -- there's still many more areas, geographic segments where we need to get to initial scale. And the ones where we have initial scale, we [indiscernible] to scale them up. So there's still a lot of work there in terms of getting to fully getting to market, and that's all focused at this point.

And does that actually open up an opportunity in terms of the type of people that you might bring in to help scale that out that maybe have different experiences to maybe build that presence in some of those geographies? And along those lines, what would be the priorities in terms of those international geographies this year?

Yes. So there's plenty of opportunities, right? I mean, we've grown the team. We already have, I would say, a fairly differentiated approach to go to market. And various people inside the company are involved in that, including our new hire, new CEO, Adam Blitzer, who's helping with some of those functions. So in terms of the geographies and the segments we're going after, like government for us is a brand-new build. We have very little going on there. And we've been investing in building the -- and getting to FedRAMP and things like that. We're still waiting for FedRAMP moderate certification that should hopefully come soon. And there are just a few more that we want to do after that. So I would say that's one of the major ones that we're building. But then there's coverage we need everywhere -- at this point, every single team, every single team's main objective across the world is hiring.

All right. So we've got just over a minute or so left here. The one question I get from investors consistently, I wonder if you could answer for us, is I get this comment that you're a mid-market vendor. You're not an enterprise vendor. How would you answer that? Do you think you're an enterprise vendor or a mid-market vendor or both?

Well, so here's what I would say. If you look at our business, roughly speaking, like with a healthy margin of error, we're about 1/3, 1/3, 1/3 in SMB, mid-market and large enterprise. And the cutoffs for us are mid-market starts at 1,000 employees and large enterprise starts at 5,000 employees. So most of the -- most of the other companies might call our mid-market enterprise basically. So when you do the math, our 1,000-plus employee business is about the same size as any of the competition and is growing significantly faster. In fact, if you look at last quarter, we probably added more ARR in that dozen-plus category than anybody else in the field, in the whole business. So I'm confident that we're a business that, while we serve all kinds of customers, also serve the most efficiently the largest enterprises, and we're the ones that deliver the most value for them. And in the end, we want to be -- to have the outsized share of market for that part of the market as well.

Fantastic. All right. with that, you hit the mark right on the spot. Oli, thanks for joining us. We really appreciate it. Stay safe, stay healthy, and we'll talk to you soon. Thanks again.

Thank you.