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 three 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 five in terms of most requested one-on-ones. You're doing something right. We're very honored to be your first. That's good. We'll always remember this. Good. Well, David, thank you. Yuka, thank you. Why don't we get started? Obviously, most in the audience, evidently, liked your last print. Mm-hmm. Stocks had an amazing run. Obviously, I think it was your, your commentary about optimization trends in the environment, at a minimum, stabilizing. Mm-hmm ... if not getting a little bit better, that acted as the catalyst. Yeah. So David, could you just maybe, at a high level, elaborate on the environment that you and the leadership team are seeing, today? So current, current snapshot. Yeah. About a year ago, we started to talk about the O word, optimization. And we talked at that point about it being concentrated in the larger spenders, cloud natives, that it ramped very rapidly, and some sectors that had 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, you know, stopped declining, had stabilized, had signed contracts, and, in fact, the growth grew a little bit. And, because the whole thing's 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. David, when you look back at that period, that I guess maybe six, eight-week- Mm-hmm ... period in 2Q, where- Yeah ... some of those all-in cohort- Mm-hmm ... really, you know, hit the brake. Yeah. 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- Yeah ... was there a particular segment that recovered? Was there an external catalyst, do you think? Why the recovery? Yeah. Or was it just generally they had gone deep enough into their cost optimizations- Yeah ... 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- Yeah ... back in 2Q? 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. Mm-hmm. 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 Datadog. Mm-hmm. They, 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. Yeah. You know, it happened more all together than we had thought, but it's also not something we can predict down to the two weeks. Yeah. Yeah. I think one question that I had out of all of this, and many in the audience may have as well. Mm-hmm ... not that every usage-based vendor needs to go into that cost optimization- Mm-hmm ... in the same timeframe and come out of it- Yeah ... in the same timeframe. But I'd love to ask for your perspective on why some of the vendors seem to be going through it at different paces? Yeah. Where it seems like Datadog and AWS are signaling- Right ... AWS, I think more firmly yesterday, actually, at re:Invent, that- Yeah ... that those pressures are moderating. Mm-hmm. Yet it seems at the other end of the spectrum, Microsoft and Google- Yeah ... are saying, "relax, we're not seeing any moderation- Yeah ... in that trend. We're merely lapping it, and hence- Mm-hmm ... we're seeing easier comps." Why the difference, do you think? Yeah. 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 homogeneous. Mm. 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- Mm ... to be, you know, as pointed as we all want it to be. It's not perfect. We also don't, even though we're correlated long term, it's, we're not correlated, you know, 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. Yeah. That's why AWS has been our largest partner. It's not... It's really, we're following the client demand. 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. Mm-hmm. 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 three-year commitment, it's more difficult for it to move as quickly- Mm-hmm ... 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 out there. Okay, that's helpful. Yeah. 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. Mm-hmm. The one I wanted to- Yeah ... to throw at you, if you could hit on this one, Yeah ... others perhaps, is around, the small, mid-sized- Yeah ... businesses. Because I think heading into the Datadog results- Mm-hmm ... a number of, growth software companies had, Yeah ... missed, disappointed. The spotlight was on- Uh-huh ... SMB duress. Right. Listening to your call, it felt like you were communicating that that was not what you saw. Mm-hmm. I'm just curious if you could talk about maybe some of the health by customer segment? Yeah. Definitely, definitely. First of all, we're about one-third, one-third, one-third enterprise, which is customers over 5,000 employees. Mm-hmm. Mid-market. Frankly, Datadog is like a mid-market despite a $30 billion market cap, 'cause it's employee. And then small, which is under 1,000. And we essentially were, I would say, pretty unique in that we cover, you know, 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, still had the highest of the usage growth, and there wasn't anything out of proportion between the three segments. Now, thinking about why, why could that be? Well, first of all, when we start small and we land and expand, our... The ticket is not that large. Yeah. 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, you know, you get to Datadog. 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. Makes sense. Okay, let's keep going a little bit on some of the growth drivers. I'll share my personal view. Mm. I think generally we're all a little bit over-indexed on the cost optimization- Mm-hmm ... side. It seems like- Yeah ... that's all that's ever discussed- Yeah ... vis-à-vis Datadog. Mm-hmm. But there are other very critical growth drivers. Yeah. The pace at which organizations are migrating existing workloads to the cloud, the pace at which they're- Yeah ... developing new apps. Yeah. So I think we've, you've—you and Yuka have done- Yeah ... a great job addressing optimizations. Right. How about we hit on some of these other growth drivers? Are you seeing any change- Yeah ... David, in the last three, six months on- Mm ... on-prem the cloud migrations- Mm-hmm ... and new app development- Yeah ... that are worth calling out? Yeah. It's a really good question. So one thing about, you know, 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. Mm-hmm. The constant development of technology, whether that be containers, microservices, and now AI, 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, you know, we're not that penetrated, as sort of the underpinning. So the underpinning, as you said, is workloads to the cloud over time. Mm-hmm. We feel really good about the very long-term trend of this. Okay. On top of that, you have. So that, that is essentially both the growth of existing applications and the movement. Yeah. 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's staying 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. Yes, they might be more carefully managed, but that's not gone away, and we're very early stages on this. ... Okay. 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 five or six years ago we really didn't have, were over $500 million. Yeah. But that's a driver. The driver to the platform and the demand in 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 one-third of the net retention, longer term, even more. As you can tell, $500 million, you guys are all very smart, so $500 million + $500 million is approximately the same as $1 billion. And so it's been a doubling by putting those products out there. Yeah. 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. What we're doing is looking at our clients and seeing how much more data and functionality can we put? Yeah. That's manifested itself in a number of things, like Network Monitoring, Digital Experience Monitoring, the security product line- Yeah. the CI/CD. So that's a 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're excited about that as a growth driver going forward. Yeah. I want to ask you- Yeah ... about a few of those, but maybe before, just to finish up on these, these, net new migrations- Mm-hmm. and new app development. In the same way that Datadog started to see- Yeah ... sequential improvement- Mm-hmm in that cost optimization front- Mm-hmm 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? You know, that has been more stable. When you look at the last eight 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. Yeah. So that hasn't proved 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. Got it. So we've had two very good quarters. Q2 and Q3 were both very good, and we gave a lot of examples. 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. Yeah. It's been, you know, 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. Yeah. Having survived with such good new customers and workloads, we're optimistic that when the pressure eases, we could, you know, get to another level. Makes sense. Yeah. Let's hit on a little bit around the product comments you made- Yep ... David, around the, APM and log side- Yeah ... especially. So you, it's great that you give us milestone- Yeah ...sizing. You don't necessarily give us growth rates, but if I could- Right ... sort of press you in an indirect way. Yeah. That if you were to look out, let's say, three years- Yeah ... 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- Right ... smaller base. Yeah. Maybe between them, are you able to rank order them by a likely, three-year growth profile? Yeah, 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. Yeah. Part of that is that we really have the platform and clients by commitments, and they're able to use it as they see fit. So, we don't, you know, control all of that. But I would say it's by order of, you know, 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. Mm-hmm. The consolidation that we've all been talking about is largely coming from that, the fact that as we've matured those products and knitted together them 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. Mm-hmm. I do have a question about the consolidation- Yeah. So why not Yeah ... move that up the ranking? Mm-hmm. So we've seen in your broad industry around infrastructure management- Yeah ... let's say, Splunk folding their hands. Mm-hmm. New Relic folding their hands. Mm-hmm. Maybe that's the wrong term. Yeah. Maybe they got a great bid, so it was a- Yeah ... it was a terrific move, but you know what I mean. Yeah. Sumo Logic selling. Yeah. Why do you think, if this industry around infrastructure- Mm-hmm ... management is healthy, as I'm- Yeah ... sure you consider it to be, why have we had three vendors sell? Yeah. I think it probably has a lot to do with us. Okay. That might be the answer. I think that we have. This is all credit to the product strategy at the core of the company. Okay. 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 log company, we're a problem-solving company for a group of users in DevOps who need to see a lot of signals. Mm-hmm. So it was designed, all credit to Oli and Alexis and Amit and others. They designed the platform that was so useful and able, and able to be implemented very quickly with data input, with lots of integrations, and no professional services, and priced so that it can be used by everybody, that 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. Yeah. You know, so I think that has to do with it. It has to do with a product-led growth. Yeah. Also, there's some other things, like Splunk came from a very different business. Basically, Security SIEM. Sure. We don't work at Splunk, but probably a great business. We're not in it. They probably are, you know, really good at it, but they tried to get into observability. Yeah ... where the platform wasn't designed that way. Yeah. Others tried to catch up the platform. So it's also the other side of the same question, like: "Oh, are you gonna be the king of security-centralized security SIEM?" 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. Got it. Yeah. Let's hit on security while you're talking about it. Yeah. How do you feel like that transition- Yeah ... to DevSecOps- Yeah ... has gone? Is it at the pace that some of your prior product launches- Mm-hmm ... have tracked, or perhaps a little bit slower because you're- Yeah ... stepping into very new terrain- Right ... against a number of fairly formidable, focused- Yeah ... security firms? Yeah. So, a couple different things. One is when we entered and expanded the platform, those were established categories already. Yeah. The DevOps revolution had happened, and we found that with our product strategy, we're able to get there really fast. DevSecOps is more mature. 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 it's very important for 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. I see. So in some ways, we're pioneering. The good side of this is there isn't anyone else there really and established. It's not like APM and logs, where there were others there. The more sort of harder to divine the exact pace is that it's a developing, really early market. Mm-hmm. 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 some of the, than, you know, the others are doing. Okay. We'll see. You know, we'll see what happens. Mm-hmm. 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. Okay. Yeah. On the core observability business- Yeah ... on the competitive front, David- Mm-hmm ... which is the more formidable threat over the next three to five years? Is it the equivalent observability tools that the hyperscalers themselves begin to offer customers? Mm-hmm. Or is it at the other end of the spectrum, where you've got a number of younger firms maybe coming from an open-source route? Yeah. Which category do you view as maybe your biggest threat over a three-year timeframe? You know, it's hard to tell. I would say that Open source is always there, right? Mm-hmm. 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 the sales of their cloud. Yeah. They like the fact that monitoring can go along with it, so clients can feel good about adopting. So that and their switch, and there's a number of things which over time have increasingly said that that's, you know, that's a friendship relationship. Okay. Open source is always gonna be there. You're always gonna have open source in some parts of it. A number of companies have us and some open source, so that's always gonna be there. Okay. And then you're gonna 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 Mean Time remediation has to go down, I would say that, you know, 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. Okay. 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. Okay. Let's talk a little bit about AI- Yeah ... if we can. Obviously, a hot topic. Yeah. The metric you've provided, I think, is 2%-2.5% of ARR coming from AI startups that are direct as customers. But obviously, there are indirect- Yeah, yeah ... beneficiaries. So, to the extent, obviously, that AI increases the productivity- Yeah ... of programmers, they build more apps, there's more observability needs. Yeah. How, how could you scope for us the size of that AI opportunity? Yeah -for Datadog? So the first metric, just to make sure everybody, is that these are modern software companies that are generally providing tools. There's a wide variety of them. Some of the names are quite famous or infamous- Mm-hmm Others are deep in the infrastructure, and we've been servicing these for a while. I think, you know, essentially, we said from 2%-2.5%. Yeah. That's like $10 million-$15 million available. Yeah. So it's a nice, fast-growing, but the bigger opportunity is the workloads of our customers, and it's really trying to understand what they're doing. Mm-hmm. It's probably too early for them to understand what they're doing- Okay ... 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. Yeah. 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. Yeah. Is Salesforce been changed overnight in the- Yeah Sales Cloud? No. Yeah. That has to happen before we can give a better answer as to— Okay. We think it's a good opportunity. Yeah. We think it's where the real money is, but we've been conservative and not sort of over-promising something that we don't see- Yeah and have the data on. So let's get- Yeah your perspective, maybe. I'd like to ask several on the stage over the next few days. Yeah -about where they think- Yeah Enterprises are in Yeah the getting AI into production. Right. It's almost an unfair question to you, I think, because you sell the observability suite- Right But you can't necessarily see how customers are using it, so it's a little unfair. Yeah, exactly. But despite that qualification- Yeah How would you describe where your customer base is in getting AI into production? Well, I can tell you what we're doing, right? So that's- So what about us? Okay. Let's start there. So we have a platform. So we're investing significantly in building the integrations. We've always had machine learning and AI in our platform, and that, what that means is we look at correlations in the performance of applications, and we always have been the you know, one of the ones that organize them, provides trends. Inserting large language models for us, we think for our utility, it's gonna enable us to be faster and more intelligent in analyzing problems, some of which may be able to be auto-remediated, others, where you're gonna 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, you know, lot, lot... They're in beta, you know? Mm-hmm. We're working with clients, we're in private and things like that. Yeah. 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. Yeah. Okay. Yeah. My rule is you can't have a CFO on stage and not ask something about the margin structure. So, I see we have two minutes- Yeah So let me sneak that one in. So you haven't given your margin outlook. Yeah -for next year. Right. 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- Yeah -that might impact that? Are you thinking about the need to invest pretty aggressively in this product roadmap that- Mm-hmm You described, in which case, maybe we should be a little bit more tempered? Yeah. Is there anything you can offer? Yeah, I think that we've been doing that all along. Yeah. So we had, we've had no RIFs. We, we did not go, I would say, overboard and then have to pull back. Mm-hmm. 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 macro uncertainty, we got more prioritized, and we focused more on our efficiency and our optimization- Mm-hmm And that resulted in the margins you saw around 24%. And what we said was, because of the uncertainty and because of the fact we got better at our efficiency, we overshot a little bit. And we've been giving guidance over time, sort of, you know, around 20 or 20+, et cetera. Yeah. That's kinda how we think about it. Yeah. We wanna maximize the long-term revenues while protecting a certain level of margin. The reason we went out of our way to say 2024 might have been an overshoot, shoot, is we don't want in the models, our investors or analysts to say: "Wow, you know, let's run that up going forward- Right -and sacrifice the investments to, to maximize the long-term growth. Makes sense. That's what we're doing. Okay, great answer. Yeah. We're out of time. Thank you, everybody. That's a great audience. Thank you, David and Yuka, for attending the event. Appreciate it. Thank you very much, everybody. Have a good week. Good questions. Yeah. Thanks. That was a good interview, David. I like that.