DDOG - Datadog Inc. (DDOG) UBS Global Technology Conference (Transcript)
2023-11-28 13:14:02 ET
Datadog, Inc. (DDOG)
The UBS Global Technology Conference
November 28, 2023 11:35 AM ET
Company Participants
David Obstler - CFO
Yuka Broderick - VP, IR & Strategic Finance
Conference Call Participants
Karl Keirstead - UBS
Presentation
Karl Keirstead
Okay, everybody, why don't we get started. I'm Karl Keirstead on the UBS software team. This is my first time on stage. So I'd love to just say how grateful I am for all of you to be here and how fun it is to inherit one of the world's greatest tech conferences. So that's been in addition to the fact that we ended up hiring 3 of the best members of the CS software team. Between them joining the team and inheriting this conference, the CS UBS merger has been admittedly fantastic for myself and the software team. So I couldn't be happier. And obviously, it would only be a success with high-profile corporates attending like Datadog. I think Datadog might have been top 5 in terms of most requested one-on-ones. You're doing something right.
David Obstler
And we're very honored to be your first and we'll always remember this.
Karl Keirstead
Good. Well, David, thank you. Yuka, thank you.
Question-and-Answer Session
Q - Karl Keirstead
Why don't we get started? Obviously, most in the audience evidently like your last print. Stocks had an amazing run. Obviously, I think it was your commentary about optimization trends in the environment at a minimum, stabilized, not getting a little bit better that acted as the catalyst David, could you just maybe at a high level, elaborate on the environment that you and the leadership team are seeing today? So a current snapshot.
David Obstler
Yes. About a year ago, we started to talk about the word optimization. And we talked at that point about it being concentrated in the larger spenders, cloud native that have ramped very rapidly, and some sectors that have been very affected. And in the quarter before this, we talked about seeing signs of stabilization of that weight, meaning those customers were getting to a level of commit and use that was consistent with their business, and they were acting in a way in terms of both spend and contract commitment that led us to believe that there was some stabilization. And we reported in Q3 that, that happened, that essentially that group of cohorts that was pressing most on our usage had stopped declining, had stabilized, has signed contracts and had, in fact, grew a little bit. And because the whole thing is a weighted average to have that weight alleviate was pretty important.
The rest of the customer base had continued to grow, but we saw that it had declined even though it was quite positive in its growth rate. And what we said last time was across the customer base, the previous speaker was talking about SMB, and I can talk about the different segments that it had gotten better. So we said we're not calling an end towards cost control or optimization. There's still a lot of risk factors out there. We can't tell the future, but there were some good tea leaves in the quarter that led us to say that the cost compression optimization had attenuated and there was more stabilization.
Karl Keirstead
And David, when you look back at that period that I guess, maybe 6, 8-week period in 2Q, where some of those all-in cohort really hit the break. When you look back, what's your perspective on maybe why they were able to almost in unison sort of snap out of that. Was there a particular segment that recovered? Was there an external catalyst do you think? Why the recovery or was it just generally they had gone deep enough in other cost optimizations that it was just a matter of time, and they all sort of came out of it together. What's the learning from what happened back in 2Q?
David Obstler
Yes. I mean, this is a process. So we're not smart enough for following each of the clients' activities to know what they're doing on a daily basis. We look at trends all the time. And I think what we saw is more of them got more intense. It takes a while to conduct the engineering projects to get efficiencies. And we said it had started in Q2 the previous year.
So what likely happened is we had a cohort that had gotten through it. Our gross retention stayed very high, meaning that our clients were not leaving data at all. They needed observability. That's one of the great things, the stickiness, but they had enough time to really work through their capacity planning, which is really what the cloud is. And it happened more altogether than we had thought, but it's also not something we can predict down to the 2 weeks.
Karl Keirstead
Yes. And when -- I think one question that I had out of all of this, and many in the audience may have as well, not that every usage-based vendor needs to go into that cost optimization in the same time frame and come out of it in the same time frame. But I'd love to ask for your perspectives on why some of the vendors seem to be going through it at different pace where it seems like Datadog and AWS are signaling, AWS, I think, more firmly yesterday actually at re:Invent that those pressures are moderating. Yet it seems at the other end of the spectrum, Microsoft and Google are saying relax. We're not seeing any moderation in that trend. We're merely lapping it, and hence, we're seeing easier comps. Why the difference do you think?
David Obstler
So a couple of different things, and there's been a lot of conversation. First of all, the data that's put out by the hyperscalers is not homogenous. So when you really -- and we don't know, we don't have that data. We don't have the segmentation on Azure into how much of it is delivery of the Microsoft Office products and other products. So it's pretty difficult to be as pointed as we all want it to be. It's not perfect. We also don't -- even though we're correlated long term, we're not correlated exactly on the timeframe. But I think overall, with software, we are more correlated to AWS simply because they're more correlated to modern applications, not lift and shift.
And so that's why AWS has been our largest partner. It's not -- it's really -- we're following the client demand. So I think it's a little more difficult to say. Another thing that I think might be the case in cloud and consumption-based software is it can move more quickly than seat-based. You -- basically, with AWS, with us, with the hyperscalers, you sign a commitment and we recognize revenues as that commitment is decremented. And so that can move, whereas if you're in a seat model or you're in a fixed 3-year commitment, it's more difficult for it to move as quickly both on the down and the up. So it could be that there's timing matters that are different in our model or consumption models to some of the other software models.
Karl Keirstead
Okay. That's helpful. And David, earlier on, you mentioned that you wouldn't mind elaborating a little bit on maybe some of the puts and takes by customer segment. The one I wanted to throw at you, if you could hit on this one, others perhaps, is around the small, midsized businesses because I think heading into the Datadog result a number of growth software companies had missed, disappointed. The spotlight was on SMB duress. And listening to your call, it felt like you were communicating that, that was not what you saw. And I'm just curious if you could talk about maybe some of the health by customer segment.
David Obstler
Yes, definitely. First of all, we're about 1/3, 1/3, 1/3 in enterprise, which is customers over 5,000 employees, mid-market. Frankly, Datadog is like a mid-market despite a $30 billion market cap because it's employee; and then small, which is under 1,000. And we essentially were, I would say, pretty unique in that we cover all of those different segments. So we're not specializing because of the platform and the way it's able to be adopted fairly frictionlessly, we can economically go to all those different segments.
And so what we saw was we saw all the segments relative to the peak have lower usage growth. But our smaller customers in SMB still because they're younger, they're getting going, we still had the highest of the usage growth, and there wasn't anything out of proportion between the 3 segments.
Now thinking about why, why could that be? Well, first of all, when we start small and we land and expand, are the ticket is not that large. And because these are companies that are modern software companies and are delivering products to clients through digital applications, the observability is a must-have. So when you're managing your costs, you are very likely and we read all about this to go into headcount, maybe real estate, maybe cloud costs, before you get to Datadog. So I think because this is land and expand, it hasn't gotten large enough in these companies to be something that is the place you'd focus on cost management.
Karl Keirstead
Makes sense. Okay. Let's keep going a little bit on some of the growth drivers. I'll share my personal view. I think generally, we're all a little bit over-indexed on the cost optimization side. It seems like that's all that's ever discussed vis-a-vis Datadog. But there are other very critical growth drivers, which organizations are migrating existing workloads to the cloud, the pace at which they are developing new apps. So I think we've -- you and you have done a great job addressing optimization, how about we hit on some of these other growth drivers? Are you seeing any change, David, in the last 3, 6 months on-prem the cloud migrations and new app development that are worth calling out?
David Obstler
Yes, it's a really good question. One thing about our business, which has been very high growth is, because it's consumption, it's not going to be necessarily straight line. And so the optimization is, let's say, the reindexing in certain customers of their -- either workloads or their use of the product. So I always think of this more, what’s the weighted average of this longer term? And that’s more consistent with the long term trend of moving applications to the cloud, growing digital businesses and the things we look at are things like -- again, these aren't our numbers, but Gartner and others, the low 20s in terms of workloads in the cloud versus on-premise, the constant development of technology, whether that be containers, microservices and LAI, which lead towards acceleration of the pace of software development.
We look at the number of customers we have versus the number -- the addressable market, looking at other software vendors and the hyperscalers and look at the fact that we're not that penetrated as sort of the underpinning. So the underpinning, as you said, is workloads to the cloud over time. And we feel really good about the very long-term trend of this. On top of that, you have -- so that is essentially both the growth of existing applications and the movement. And we saw that when you look at things like the stability of our new logos, and the initial projects in the period of cost control and that thing strong, that leads us to believe that we're still the priority projects and the trends of movement of applications to the cloud are intact. And yes, they might be more carefully managed, but that's not gone away, and we're very early stages on that.
Then other growth drivers that we see are, we started out more as a one product company. We didn't have infrastructure. We didn't have logs in APM. And we said on the last call just to give everyone a sense of how well the platform has resonated that both of those other products, which essentially 5 or 6 years ago, we really didn't have over $500 million. And so that's a driver, the driver to the platform and the demand and DevOps of clients to look at everything in a unified way and not to hop between different applications. That's been a driver for us in the product adoption.
It's been short term, probably 1/3 of the net retention. Longer term, even more, as you can tell, 500 million, you guys are all very smart. So 500 million plus 500 million is approximately the same as 1 billion. And so it's been a doubling by putting -- putting those products out there and we're not even fully penetrated in that.
In addition, as you talked about, we're a growth company, and we've been investing a lot in the platform. So what we're doing is looking at our clients and seeing how much more data and functionality can we put. And that's manifested itself in a number of things like network monitoring, digital experience monitoring, the security product line, the CI/CD. So that’s the matter of -- for the same customer base, trying to put more and more signals in the platform, which has been a growth driver and we are excited about that as a growth driver going forward.
Karl Keirstead
I want to ask you about a few of those, but maybe before, just to finish up on these net new migrations and new app development in the same way that Datadog started to see sequential improvement in that cost optimization front. Have you seen any improvement on the on-prem to cloud migration or new app development front? Or might that occur with a lag, perhaps as the economy gets a little better?
David Obstler
That has been more stable. When you look at the last 8 quarters, we've gotten about the same number of gross logos. We set some records in amount of ARR because the land size is going up. So that has improved as volatile. Why might that be? Probably it's because the mission critical, the things that are really at the top of the list are being executed. So we've had 2 very good quarters. Q2 and Q3 were both very good, and we gave a lot of examples of if the investors want to go back through the script, you can see a number of examples of fairly large and traditional industries and consolidations. So I would say that one hasn't been as volatile. It's been a good point of this whole thing. Still, as you might think, there's still cost management, and there's still likely the case that there's depression of this.
So having survived with such good new customers and workloads, we're optimistic that when the pressure eases, we could get to in their level.
Karl Keirstead
Makes sense. Yes. So let's hit on a little bit around the product comments you made, David. Around the APM and log side, especially. So it's great that you give us milestone sizing. You don't necessarily give us growth rates. But if I could sort of press you in an indirect way that if you were to look out, let's say, three years, which of those buckets would you expect to be growing the fastest over the next three years? The answer is probably the smaller ones because they've got a smaller base, but maybe between them, are you able to rank order them by likely a 3-year growth profile?
David Obstler
Yes, I think that -- and by the way, -- we try to give these metrics to give an idea, but we intentionally don't give revenues by product. Part of that is that we really have the platform and clients by commitment and they're able to use it as they see fit. So we don't control all of that. But I would say it's by order of our penetration and when we went in the market. So the underpinning is infrastructure, but because we've been getting a very good attach rate on both APM and logs, we've seen them be quite similar in terms of growth rate and get to this, maybe logs a little bit more because APM was there before, but roughly similar. And that has to do with the platform. That has to do with -- the fact that we arrived later so there were a number of customers that are using our infrastructure and are using another vendor, open source.
And the consolidation that we've all been talking about is largely coming from that. The fact that as we mature those products and knitted together in the platform, we are getting what is the natural state, which is clients wanting to be on the platform and adding on these other products.
Karl Keirstead
I do have a question about the consolidation, so why don't that up the rank. So we've seen in your broad industry around infrastructure management, let's say, Splunk folding their hand, New Relic folding their hand. Maybe that's the wrong term, maybe they got a great bid, so it was a terrific move. But you know what I mean, [logic] selling. Why do you think if this industry around infrastructure management is healthy, as I'm sure you consider it to be. Why have we had 3 vendors sell?
David Obstler
Yes. I think it probably has a lot to do with us.
Karl Keirstead
That might be the answer.
David Obstler
I think that we have -- this is all credit to the product strategy at the core of the company. So when Datadog platform, which was formed later than the others was created, it was created as this platform that was very flexible Datadog. Essentially, the whole thing is we're not an APM company, we're not a law company. We're a problem-solving company for a group of users in DevOps who need to see a lot of signals. So it was designed all credit to Ali and Alexi and Amit and others. They designed the platform that was so utile and ability to -- and able to be implemented very quickly with data input with lots of integrations and no professional services and price so that it can be used by everybody that sort of started to take the market because I'm in my sixth year and we were way down the list of sort of the Gartner and et cetera, observability. And now we're at the top.
So I think that has to do with it. It has to do with product-led growth. Also, there are some other things like Splunk is -- came from a very different business, basically security SAM. We don't work at Splunk, but probably a great business. We're not in it. They probably are really good at it, but they tried to get into reservability where the platform wasn't designed that way and others tried to catch up the platform. So it's also the other side of the same question like, are you going to be the king of security -- centralized security SIM? No, we're not. We have a very good market, and we're going for that. So some of these attempts were not their primary market.
Karl Keirstead
Got it. So let's sit on security while you're talking about it. How do you feel like that transition to DevSecOps that has gone? Is it at the pace that some of your prior product launches have tracked or perhaps a little bit slower because you're stepping into very new trains against a number of fairly formidable focused security firms.
David Obstler
Yes. So a couple of different things. One is when we entered and expanded the platform, those were established categories already. The DevOps revolution that happened. And we found that with our product strategy, we're able to get there really fast and DevSecOps is more mature. So security is purchased and used in different ways. And so therefore, there's the maturity of the market and the greenfield nature of the market.
The second thing I think is very important to everyone to understand is that when we're talking about security, the main businesses of Palo Alto, CrowdStrike, et cetera, those are mainly endpoint. We're not in endpoint. We're in cloud and app security. And we're architecting our platform like our other products, which is for ubiquitous real-time use and even when you think about Palo Alto, Prisma, et cetera, the market they're going for, although it's related is almost a different end market in the type of use. So in some ways, we're pioneering.
The good side of this is there isn't anyone else there really in established. It's not like APM logs where there were others there. The more sort of harder to define the exact pace is that it's a developing really early market. We went into it because we got a lot of feedback from customers in DevOps that there's a lot of value in using these signals. And it seems to fit very well with our real estate, our eyeballs, our platform, the data we have, but we're pursuing a little bit of a different strategy than the others are doing. And we'll see what happens. It could be that there might be a point where -- like in DevOps, you don't have to go to a central CTO to buy all development tools, et cetera, or it might be that it's a little more sticky there with the CISO, and we don't know yet.
Karl Keirstead
Okay. On the core observability business on the competitive front, David, which is the more formidable threat over the next 3 to 5 years? Is it the equivalent observability tools that the hyperscalers themselves begin to offer customers? Or is that at the other end of the spectrum, where you've got a number of younger firms may be coming from an open source route, which category do you view as maybe your biggest threat over a 3-year timeframe?
David Obstler
It's hard to tell. I would say that open source is always there, right? So I think the cloud tools are more or less inputs into our integrated product rather than a direct rival product. We're good partners, and they're really trying to maximize sales of their cloud. And they like the fact that monitoring can go along with it, so clients can feel good about adopting. So that -- and their -- there's a number of things, which over time have increasingly said that that's a friendship relationship. Open source is always going to be there. You're always going to have open source in some parts of it.
A number of companies have us and some open source. So that's always going to be there. And then you're going to have, in some ways, a little bit of a different approach like do I want real-time speed and ubiquity or do I want a more centralized tool and that gets into the architecture of some of the competitors. I would say in our use case, which is this ubiquitous real time, where meantime remediation has to go down, I would say that the main competitors have been -- the vendors we talked about, who I forget your words are no longer public companies or might not be public companies in the future and open source.
The smaller companies that have risen up and many of which have been acquired by -- when I say risen up, meaning they've gotten small amounts of revenue, generally have not been successful and are not major competitors in our market.
Karl Keirstead
Okay. Let's talk a little bit about AI, if we can. Obviously a hot topic. So the metric you've provided, I think, is 2%, 2.5% of ARR coming from AI start-ups that are direct as customers. But obviously, there are indirect beneficiaries, so to the extent, obviously, that increases the productivity of programmers. They build more apps. There's more observability needs. How do you -- how could you scope for us the size of that AI opportunity for Datadog?
David Obstler
So the first metric just to make sure everybody is that these are modern software companies that are generally providing tools as a wide variety of them. Some of the names are quite famous or infamous, others are deep in the infrastructure. And we've been servicing these for a while. I think essentially, we said from 2% to 2.5%. That's like $10 million to $15 million of ARR. So it's a nice, fast growing. But the bigger opportunity is the workloads of our customers and it would -- and it's really trying to understand what they're doing, and it's probably too early for them to understand what they're doing in deployment. So that is hard for us to say.
In other times, as I mentioned, when there's been technology advancements, which have made the whole thing more complex and more modern like microservices and Kubernetes. That's been a real friend of Datadog. And we're architecting the platform for that. We're building all the integrations. We're building remediation intelligence using models and everything. But because the clients haven't put these large language models in to a great degree in their client-facing applications, like I'm just making up, Salesforce. Has Salesforce been changed overnight in the Sales Cloud? No. So that has to happen before we can get give a better answer as well. We think it's a good opportunity. We think it's where the real money is. But we've been conservative and not sort of overpromising something that we don't see and have the data on.
Karl Keirstead
So let's get your perspective maybe I'd like to ask several on the stage over the next few days about where they think enterprises are in the getting AI into production. Right. It's almost an unfair question to you, I think, because you sell the observability suite, but you can't necessarily see how customers are using it. So it's a little unfair but despite that qualification, how would you describe where your customer base is in getting AI into production.
David Obstler
Well, I can tell you what we're doing, right? So that's what about us? So we have a platform. So we're investing significantly in building the integrations, we are -- and what that means is we look at correlations in the performance of applications, and we always have been one of the ones that organized that provides trends. Inserting large language models for us, we think for our utility, it's going to be able us to be faster and more intelligent in analyzing problems, some of which may be able to be auto remediated. Others where you're going to have really good suggestions. And we are working on that. And so we also have the BITS chatbot and everything like that. So I would say that -- our products are not -- they're in beta. We're working with clients or in private and things like that.
So if that's a good example, we've been working at it this year, and we're in construction mode and don't have it out to a lot of clients yet. Yes.
Karl Keirstead
Okay. Yes. My rule is you can't have a CFO on stage and not ask something about the margin structure. So I see we have 2 minutes. So let us take that one in. So you haven't given your margin outlook for next year. But are there any variables that you think we should keep in mind as we all try to figure that out? Or is there plans for any change in headcount hiring that might impact that? Are you thinking about the need to invest pretty aggressively in this product road map that you described in which case maybe we should be a little bit more tempered. Is there anything you can offer?
David Obstler
Yes. I think that we've been doing that all along. We've had no risk we did not go, I would say, overboard and then have to pull back. So we've been really good at sort of continuing to grow. There's a lot of opportunities. We have plans and we always have had to aggressively grow R&D and there's a lot of go-to-market. I would say that with the uncertainty, we got more prioritized, and we focus more on our efficiency and our optimization, and that resulted in the margins you saw around 24. And what we said was because of the uncertainty and because of the fact that we got better at our efficiency, we overshot a little bit. And we've been giving guidance over time sort of around 20 or 20 plus, et cetera. So that's kind of how we think about it. We want to maximize the long-term revenues while protecting a certain level of margin.
The reason we went out of our way to say 24 might have been an overshot is we don't want in the models our investors or analysts to say, "Wow, let's run that up going forward and sacrifice the investments to maximize the long-term growth and that's what we're doing.
Karl Keirstead
Okay. Great. We're out of time. Thank you, everybody. That's a great audience. Thank you, David and Yuka, for attending the event. Appreciate it.
David Obstler
Thank you very much, everybody. Have a good week. Good questions.
Karl Keirstead
Yes, thanks.
For further details see:
Datadog, Inc. (DDOG) UBS Global Technology Conference (Transcript)