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

It is -- how is everybody holding up, by the way? This is day 2, and we're not done yet. We've got 1.5 days more. Plenty of AI. So this is going to be an exciting, hopefully, session with Datadog. David, welcome to the conference. Third year in a row. It's great to have you back.

Thank you for having me. I appreciate it.

Absolutely.

I always enjoy coming to this conference.

Most happy to have you. So let's start off with this. How is the conference treating you so far?

It's been great. Lunch was pretty good, and I got a cup of coffee. So it's okay. I came in on the 7 a.m. flight from New York. So I was saying that I'm not often up at 5:00 a.m., but yes.

Awesome. Okay. Well, thank you for putting up with the long hours. So I asked you this question 2 years ago, last year. I'm going to ask you the same question. Maybe the answer will turn out to be slightly different. What do you want Datadog to look like in 5 years? And how do you define success for the company?

Yes. I think Datadog has always been a product-led growth company with a firm focus on DevOps. And I think our CEO, Oli, put it very well at -- I believe it was our Investor Day, that the vision is that Datadog is the platform that, that customer base turns on when it comes in, in the morning and spends their whole day in it and does their job in it, which is essentially the monitoring and remediation of mission-critical modern apps. And that's the way we think. We think about what the customer is doing in their day and have relentlessly tried to expand the platform so that they spend more time in the platform, and we make their lives easier to do their job. So that's something we always have in our mind.

And so if you can translate that into some sort of a tangible goal 5 years from now, how would you boil that down into? It doesn't have to be financial, but what are the kind of companies that you guys aspire to be like in terms of role models that...

Definitely. And I think that's a good way to put it because we've always thought about other platform companies, ServiceNow, Atlassian, Salesforce in the Sales Cloud, and thought about how we can create a solution that is easy to use, very flexible, ubiquitous, meaning everybody is in it and keeps innovating and in order to have product-led growth that's maintained and essentially does anything the client wants as applications evolve, which includes AI we can talk about. And so that is the North Star guiding the investment philosophy. Datadog spends more in R&D than all the other companies in observability combined.

That's a good...

And that's the epicenter of the company. Yes.

Wow. Okay. And to achieve these goals, what are the things that you need to do operationally, decision-wise, hiring, scaling? What's the plan to get there?

Yes. And it is very much related to hiring, educating, grooming people. So I think what we need to do is if you look at -- it was in our Investor Day presentation. We put out a vision of service management, which is really a holistic way of looking at DevOps and their uses. And then it looked a little bit on different cases around it, including shift left into getting involved in the life cycle of developers earlier, security, business analytics. So I think what we have to do is we have to continue to innovate in the platform and add more functionality like I'm speaking about. There are some really good examples of that, which we can get into, from our most recent DASH where we added -- where we talked about a service management function for on-call management of the platform, LLM monitoring, et cetera. And so that's the product side. And then we have to continue to grow and evolve our go-to-market in a number of ways, geographically, channels, Fed, governments and centralized selling, enterprise upselling, in order to be able to deliver this product to our increasingly sophisticated customer base.

Got it. Yes. You recently talked about how enterprise was strengthening at the margin in SMBs and not seeing strengthening but showing stability. Can you just give a little bit more in-depth read into what's going on for business trends?

Yes. I think what we saw a year ago or a little more than a year ago was, call it, a rationalization and optimization. And we saw that around cloud natives, who had expanded very rapidly, and that went across our customer base. That was in the cloud natives and SMB and mid-market as well as the cloud native size of enterprises. And then we saw more of a pro rata recovery in the previous quarters. And in the last quarter or 2, we saw the larger enterprises, particularly the more traditional industries, get back to what they had done before investing in digital projects and experience more rapid growth than the SMB side of it. We see stability in SMB. We see growth in SMB. But we saw a little more of the investment impetus go be correlated not with size and with enterprises.

What -- I'm just curious, what could be holding SMBs back? Is there any execution thing or a macro thing? What...

Yes. It's probably a couple of different things. They've probably over-expanded more in the bubble. The sort of business model changed from invest at all costs to growth with profitability. And the venture capital investment cycle got suppressed. So I would say the environment has been steady, stable but maybe not as robust. Now I would clarify that our SMB is probably more of what you think about M because if you're going to have a digital application and you're going to have a DevOps group, you're very unlikely to be a 50-person mom-and-pop. So that may be one of the reasons why I know there's been a lot of discussion about SMB, why our SMB has held up more in this whole cycle. Also because it is related to delivery of the mission-critical apps, if you're alive, you're not going to turn it off. So that creates a lot of stability.

Got it. Got it. We also talked about in recent quarters how customers are signing longer-term contracts. What is driving longer-term contracts?

Yes, definitely. So a couple of different things. Customers have gotten more confident, one, in that they can get a good understanding of their business and their capacity needs. So they're able to plan out longer. They -- I think Datadog has been winning market share in observability. So more customers are deciding that they're standardizing on the Datadog platform. That's partly related to consolidation but also complements contract term. And particularly in larger customers in enterprise, they have been interested in the trade-off between volume and length as it relates to price. So if you're going to have Datadog as your platform, you have good sense of the amount of capacity needs. It makes a lot of economic sense to commit out longer.

Moving into gen AI. How does Datadog benefit from gen AI? Conversely, if AI is not a thing and we're not talking about it, not that I'm betting on it, does your business do fine?

Yes. Well, definitely. So I'll start with the core business first. So anytime there has -- so we are about the modernization of applications and the replatforming of applications and then delivering them in the cloud. So whatever technologies have been deployed, you could say containers, servers, et cetera, that's been a signal of a Datadog buyer. And so if LLMs are put in models and cause an acceleration of the investment cycle in replatforming or acceleration in the ability to develop new applications, that would complement Datadog. So it could be that. It could be anything. But anything that changes and stimulates modernization would -- has helped Datadog. Some other ways it's helping Datadog today, it's -- there's an end market of tools companies, which have been growing very fast. They account for about 4% of our ARR, and they're perfect customers for Datadog. They're modern application companies. We also have been injecting more LLM into our platform, whether it be with a chatbot or an LLM monitoring module. And we've been experimenting ourselves on how to get more efficient and more productive in using large language models internally.

We're going to get...

So we'll see what happens. But I would say if history repeats itself, if it is adopted in applications, it's been a helpful accelerant to date.

Got it. That was really well put. In recent quarters, you've seen after optimization ended, hyperscalers have been putting up, largely accelerating revenue growth rate. Your growth rate is good, but it's not seen acceleration. Maybe it's an unfair expectation. But it is what it is. So what, if anything, could be holding that marginal customer from spending more with Datadog?

We've had some, as I said, stabilization and some improvement of our net retention, but we still are in a cost-conscious environment. I would say in comparing it to the hyperscalers, we tended to have higher growth rates than the hyperscalers. But the hyperscalers have timing differences and potentially business differences. A good example is in order to deploy AI applications, you first have to invest in the infrastructure. And there have been significant beneficiaries, including the hyperscalers, NVIDIA, et cetera. And we tend to benefit a little bit later in that cycle once those are put in applications. So you may have timing differences in what's accelerating the hyperscalers' growth and Datadog. Another thing is they have much more broad product area within them. And we don't know what their growth has been in modern -- the modern cloud. So essentially, there's timing differences. I would say we are both driven by the same thing, which is that replatforming, development of modern apps and putting them in the cloud. But it has never and today doesn't match up from a timing perspective.

I remember when we met Oli some time ago, he said that we're not yet in the application phase of generative AI. We're still in the infrastructure build-out. And so that was a catalyst for us to actually form our thoughts on the report that we published a few days ago, where we say we're in the infrastructure build-out. Then comes platforms. Then come applications. So...

It brings us back to, remember, something we even said on our IPO road show that we are a follower from the deployment of infrastructure. That -- and then deploy the infrastructure. You get the new applications and then Datadog monetizes. So the -- I think as you talked about, we're not in that.

Later cycle, benefit cycle. Okay. That's good to know. Good to frame that. So AI growth -- is AI growth incremental to cloud growth? Or do you think there could be some substitution, meaning some cloud modernization projects didn't get the funding within -- they shifted to AI? Is that...

We've been asked that a lot and I don't know. We have seen this return, particularly in enterprises, to a more normal behavior of executing digital projects. And we -- if you look at our earnings script, a lot of new big customers, whether that would have been higher if there wasn't the AI out there, we don't know. But we haven't seen it disrupted to the extent that normal digital projects are not being executed. So that's what I have to say about that.

Yes. With gen AI contributing 4% of ending ARR in June, what are you hearing from your customers that are doing gen AI projects? Are they getting return on investment with their gen AI initiatives? You are benefiting from it, but are they benefiting?

Good question. So just to clarify, this 4% are not company that are putting -- so this is probably, again, what you would see is the first way. So it shows you a bunch of activity. We also have been investing in the platform. And we have, for instance, a metric we gave. Over 2,500 customers out of our 29,000 are using our integrations. But our thing is -- our belief is we're still early. And they don't even know yet about their return because most of what they're doing is testing, sandboxing, training. And we haven't seen those production workloads. So again, we're not them, but we would guess that it's too early to answer that question as a lot of experimentation going on right now.

These are models that are being trained, and Datadog is getting pulled into these projects as these model training -- simulations are extracting their total infrastructure, and you're there to provide that...

Well, I think we tend to be more in production. So after training, inference. So that would be when -- and that's what -- one of the parts of our LLM monitoring module. So I think there is a lot of training. I think there is a lot of experimentation. But so far, if you look at our workloads, there's not a lot put in production yet. And since these are customer-facing mission-critical applications, our guess would be that you would have to do a lot of testing and making sure that it worked well before you put it out there in a customer-facing application.

Oh, I see. The revenues you're seeing are more on the inferencing side.

We have very little -- we have very little revenues from production so that we have some use but not a lot in terms of revenues. It's still too early. For instance, the LLM modeling, this is typically what we do. We actually go beta but don't charge for it. We kind of learn from it. And so it's being used, but we don't have revenues from it yet.

Got it. Got it. Okay. Got it. So then this 4% of revenue, that is pertinent to what? I just want to clarify that.

Oh, sure. So that are companies that are providers of AI tools across the AI stack, and they are delivering an application or an API, and they're using Datadog for normal monitoring for the most part. Is the application working? Do they have enough infrastructure behind it? So that would be an observability of an AI software solution. That's what that is.

Got it. Makes sense. So we've talked about consolidation, recent tool consolidation in recent quarters, consolidation of your customers' tools. It looks like it's becoming a theme. Can you expand a little bit on this? What are the customer conversations like? And what are the longer-term implications for Datadog's business as consolidation trend continues?

Yes. Datadog is a platform. Our customers are telling us...

Because every company in this space are saying the beneficiary of consolidation. But clearly some benefit more than others.

I think we can prove it mathematically, but everyone should say if they feel it. It's really we're a platform sales. So our customers are telling us they want more and more functionality in the platform. We talked about it on the way in. A lot of customers, we didn't have some of these solutions. So if you have many point solutions, you have to run around when there's a problem, do your investigations. You don't have the correlation, et cetera. So the overall impetus and one of the big growth drivers of Datadog has been this multiproduct adoption, which we give as a metric in every quarter. And that's because customers view this as a product, the platform, and they want to see more and more data in it. So as we've had these products and they've matured and they've become best of breed, over time, we've been able, and this has been going on, to take market share from either cloud-native tools, open source in some cases and other solutions because our clients want to see all of this in one platform in order to do their jobs better. So this has been going on for some time. It's how we built the businesses that we said got to over $500 million in both APM and logs. Part of -- some of that is greenfield and some of that's consolidation. And we do see this as a major driver. What does it mean for Datadog? Datadog has growth vectors from new customers, from customers who are doing more and more digital products but also a strong growth vector in adding more products into the platform used by our clients and growing market share.

Got it. So we've got the core infrastructure business. You have APM and logs on top of that, 2 scaled businesses. You have synthetics and RUM. The last 2 of them have gotten to $100 million in milestones. Where do you see them going? Could they be the next $0.5 billion businesses?

We don't know. I mean it's -- we're very pleased that we've been able to get this kind of attach rate. Whether -- we don't know whether they could be. We have impressive penetration but not saturation, particularly in the things like APM, synthetic logs, database monitoring. So we think there's a lot of upside but aren't going to give a forecast of those. We basically report as we get to milestones but have not tended to give product forecast.

We hope that they mature into big milestones, which just naturally begs the question, core infrastructure. What are the growth opportunities still left in that core engine of the company? I'm just curious.

Well, I mean, that's very correlated to workloads. And if you look at research, 20%, 30% of applications are in the cloud. Enterprises themselves are, in many cases, very early. So the opportunity is, one, workloads and digital migration. Two would be other types of infrastructure, whether it be GPUs, containers, serverless, et cetera. And then on top of that, you have the concept of service management of a platform, which allows the customer to do more things. So that is sort of at the epicenter of the main driver, which is the modernization of the application stack and the deployment in the cloud.

Got it. Got it. Let's talk about APM and logs a little bit. What are you playing for in these 2 market segments, which are standalone? They could be separate companies almost, right?

Yes, exactly.

What's happening with your go-to-market competitive dynamics in those 2 segments and replacement opportunities that you might be seeing in those markets if you've been...

Definitely. Good question. So we didn't have those products 6 years ago. We've been successful in attaching but we're not saturated. So I think we're getting better and better of selling the platform and getting understanding what other solutions are out there and working with the client over time. So that's a pretty big opportunity. And that could translate. I know you have an additional question into more and more million dollar-type opportunities because that means we're taking more of the wallet in those companies. In addition, there are certain infrastructure things like in logs. We talked about Flex Logs. This is essentially making logs more flexible, of course, splitting out compute and storage. And that is an infrastructure element that will allow other products to benefit, for instance, the SIEM, the Cloud SIEM. We found we need to be able to slice those logs to make that product successful. So I think there's some core infrastructure things in these products that will help other products as well.

Got it. I remember watching the demo of your latest version of logs at your DASH conference. And I stood behind 3 customers who were techies. And they were asking all kinds of challenging questions of the demo person and said, run this scenario, that scenario and their jaws just went wide open. Later on, I asked them, so what do you find so good about this? Oh, my God, this is incredible.

Thank you for coming. That's great. I think we've done -- I think we've been really good at logs has to do with engineering the infrastructure, and there's all sorts of variability of clients in what they want to do in terms of storage, retention and all sorts of use cases. And so to be able to get to the point where with both Husky and Flex Logs that we're innovating the platform has really been a foundational element for the growth of the business.

And anything to add at all to the replacement opportunity since we've seen some consolidation happen in this space? Big networking company bought another big company. So what are you seeing in terms of deals where somebody is looking to replace old deployment, logs especially?

It's been going on for a while. So I think that's the case. And it happened in APM. It's happening in logs. It even happens in RUM and synthetics and others. So we've been investing and getting best in breed. And because of the platform, some of these other companies have lagged or had disruptive events, going private, getting into a large company. And we like that because as we continue to maintain the focus on investment, that allows both us, the business ourselves, but also customers get concerned because they've seen this before where the investment has not kept up in these situations. And we've been doing this for a while. And when other solutions have gotten bought by large providers, it's been an opportunity for us to further consolidate. We're optimistic here.

Got it. I know you don't disclose $1 million customers every quarter, but 2021 was 119 customers. 2022, we added about 101. 2022 is about 79. How important of a KPI is this? And if it is, where would you like for this to be in the near term and the long term as well?

So because we're land and expand, because most of those million-dollar customers were less than $1 million, they crossed to $100,000, $500,000. It's important. And when you look at our ability to cross-sell, get more of the platform sold, get more enterprises on, I think we do look at this as one of the metrics. It's a metric of health of cross-sell, of growth with customers. And we're not going to give a target, but we think there's a lot more opportunity in enterprises, which I think Oli has said many times are earlier in their development of digitalization.

Got it. So I had a chance to host a lot of companies in the last 1.5 days or so. And I've asked them about how their products are being positioned with generative AI. But I also have asked them what are they doing internally with generative AI. So are you doing any pilot projects? Any early results that kind of make you feel optimistic about how it plays out within the internal operations of the company?

What we're trying to do is we're trying to remove the barriers to adoption. So we're treating the large language models as a base part of the kit like you get Google suite and then look and see how it's being used. And a use case that surprised me to the upside is half of the use cases are in sales and marketing.

Really? Wow.

And what it is, is salespeople are able to investigate a company or a target or who works there or where they are, and where this has been databases in the past, they're able to do it in real time. And so that's been a really interesting use case. Marketing collateral has been an interesting use case and putting together case studies, et cetera. And I would say it is being used by development, but we probably aren't as far. They have a lot on their plates, so they're experimenting. But I wouldn't say it's dominant in terms of our software. So there's good case study. It's still very early. And some of the use cases have surprised immediately.

That's good to hear. Is this internally developed stuff that Datadog's engineers came up with or you're using third-party commercial software to achieve these gen AI...

We're using third-party software.

Anything that you can mention?

No.

Sorry. I understand.

We are. There is -- we are working within our own -- our LLM on -- it is called Toto. It was a blog about a large language model to work on the operation of the LLMs in an application. So we are doing it. We have a data science group working in the product itself. I'm talking about more of the overall broad use case in the company.

David, do you foresee this efficiency that you're uncovering to your pleasant surprise? At some point, does it become somewhat material to margins, efficiencies that you just didn't -- nobody anticipated 18 months ago, this could be at some point material?

I think our hope is since we have more projects and more territories to cover than we can digest in people that the most profound effect would be on productivity and that, that could translate into top line. So that's the kind of way we're thinking.

Top line.

How can we get more productivity, launch software faster, how salespeople operate faster. Reluctant to say whether there's another factor down the road, but I think that that's where we're thinking about it right now.

It certainly doesn't look like you're looking to replace humans with AI, and therefore, you're going to be hiring less with the result of this.

We think at least in this part of it, it will be help humans get more productive. Like we're not thinking out in that sci-fi way. But we see use cases where maybe it was more difficult to process all the marketing collateral in all the products we're launching. Maybe we can speed it up. Maybe we can do marketing collateral in lots of different languages faster. Maybe this will allow us to get out and communicate. So there are so many ways that you can define it, but I think it has to do mainly with helping the people become more productive.

And that was going to be my -- towards the final few questions here. The margin profile of the company is so unique that you have this amazing growth, at the same time, very strong operating margin. Can you share with us the formula for running such an efficient business? And maybe AI is going to contribute to even better efficiencies in the future. But what is the secret sauce here?

Well, I think the biggest birthright is the product and the platform. So the product in 2 ways is -- creates efficiency. One is that you can add additional functionality in a very efficient way. The architecture of the platform and the data infrastructure contributes to the velocity of product introduction. At the same time, it also makes it able to be used by clients without professional services and used by everybody, which helps with the sales velocity. That means that clients can adopt in a more frictionless way than other companies and has created that efficiency. That enables us to put it back in the product. So it's a virtuous cycle. Now I think that when you think about the combination, we definitely have a lot of projects and a lot of investment to make. What these 2 have enabled us to do is make those investments and develop the margin profile yet keep firmly fixed on the top line and innovation.

Got it. Time flies. In the minute that we have, any final thoughts? How do you look at potential acquisitions? And what are the criteria that you evaluate in regards to whether it's additive to the business? And also keeping in mind the investor expectation as well.

Yes, definitely. So it all comes off of our -- in a product-led company off our product road map. We have a number of areas of functionality. We understand that we might be able to accelerate that to the extent we can identify good teams who want to continue on with Datadog and the technology and infrastructure that we can integrate in the platform. So that's been the bread and butter. There may be -- it may -- we certainly look at a lot of things, but we have a pretty high bar given it has to have those things. And then we tend to be pretty disciplined buyers. And so that creates a bar, meaning that the acqui-hires tend to be the things we focused on. I think we could have a company that has more revenues. But the bar at a very large acquisition is really high because of those effects of the discipline, the platform, the people staying and not having to fix somebody else's thing up. So I think that's how we look at it.

Got it. Why don't we wrap it up? Let's give a round of applause to David. Thank you for staying with us.

Thank you.

Thank you once again. Enjoy the rest of the conference. We hope that you're better informed, better ready to take on the challenges after 3.5 days of investing your time with us. Thank you so much.