Ready to do this? Let's do this. My name is Matt Hedberg. Thank you. By the way, I got to just pause for a second. You guys make this conference fantastic. I was talking to Matt Finnegan, who runs sales. You can't even hardly walk in the hallways. It's just great attendance. And so you guys make this all possible. So I really do appreciate everybody coming and supporting this conference and these great companies. So let's get into it. And by the way, this is a big room. I'm intentionally going to be leaving some time at the end. Save some of your questions for the end. There's a microphone that will pass here, and David will kind of squeeze in.
Thank you for having us.
Yeah. So David Obstler, CFO of Datadog. I'm sure you guys don't need much of an introduction. You guys are coming off of strong Q3 results. It seemed like a lot of stability in end markets. Can you just sort of double-click on what you saw during 3Q and kind of up until when you guided relative to maybe the first half of the year?
Yeah. So it was, I think, a continuation of the trends that we talked about in the previous quarters in that in our enterprise segment, we had sort of normal activity, meaning that larger enterprises were back to their digital projects and adopting solutions like Datadog. That resulted in a number of really large and nice deals that were featured on the earnings call and a continued improvement in the net retention rate there. And that's good news. That's sort of what we've been saying over the last two or three quarters. Continuation of an SMB of a stable environment, meaning it's continuing to grow. Nice net retentions, but not accelerating, likely due to the economic concern of the venture capital environment, etc. And we also said we continued our motions of cross-sell and development of a multi-product sale.
And evidence of that were the metrics we gave around the three pillars being now over $2.5 billion of ARR and quite a number of products getting past $10 million. So I would say it was a continuation of some of the trends that we've seen since what looks like the bottom in the historical time series of Q3 last year.
Q3 last year was sort of where you saw some of the peak headwinds. I know I've asked you this, I don't recall. What percentage of your business would you classify as SMB?
It's about 1/3 of our business, 1/3 to 35%. Our largest segment now is probably tied, maybe 35%-40% in enterprise and SMB with the rest being mid-market. Mid-market is 1,000 employees to 5,000 employees. Datadog was a mid-market up until very recently.
Yes. You said that sort of some of the cloud optimization trends that you were seeing last year sort of just stabilized. I'm curious. It feels like optimization is part of the new lingo. All companies should be optimizing all the time. What does that mean from an ROI go-to-market perspective? Has selling changed sort of in a post-COVID world where you're not an order taker, but you have to deliver ROI? Has that conversation changed relative to the optimization trend that probably continues?
In cloud applications, in cloud-related purchases, it's always been the case, so since we went public, we've said that our customers go through cycles of investment and deployment of applications and then optimization, and whether we're pointing it out has to do with whether it's become more or less intense, so I think that is more about cloud applications. Yes, in environments that are more careful, you have more focus on the cost side and potentially the trade-off between growth and optimization, but I think it's always been there and is going to be there in cloud applications.
Okay. I've talked to you about this and you and I were talking about this yesterday. You've been saying since IPO, billings isn't a great proxy. And reported billings wasn't great in Q3. You normalized and it seemed more sort of mid-20s, I believe you said, if you normalized for duration. Why shouldn't that be a concern for investors, especially when we think to next year, to next year's revenue growth?
Because billings will be higher or lower than revenues or ARR growth, which is what drives the business. Based on the timing of sending bills out that we don't manage for. I mean, we're not a company that is strapped for cash and is trying to, "I want this bill to go out this minute rather than that minute." And when we look at it over a weighted average basis, it goes back to revenue. So in a quarter, long term, it's correlated with revenues. And I think we said in the earnings that if you do the math, you'll see all these metrics go back to revenue growth. But in many periods, it's been higher and in many periods, it's been lower. And that has to do with the timing of billings going out versus the comparable period.
Don't forget, we don't have a renewal cycle like everything renews on this date. We have a commitment cycle, so it depends when the client is going to redo their commitment versus last year, and when it's been higher, we perform it back to revenues, and when it's slower, we perform it back to revenues, so I think we've been pretty fair in providing evidence.
Yeah. No, that's a great perspective. I think we all try to, I think, asking Yuka, what are some words of advice you give to investors? And I think the answer was people get too little detail in their modeling of Datadog and ultimately revenue is what matters most over sort of long duration period.
We're trying to help everybody. Hopefully, it's appreciated. We're basically saying, "Here, pro forma for this bill that went out at a different time." And there are slight duration differences. So we're trying to help everybody. If you look, I think, and if you look through all the scripts, you'll see that there's a lot of analysis in that or performing. It all gets back to this truth.
One of the real standouts from last quarter was AI Natives. Really, I believe it jumped from 4% of ARR to 6% this last quarter.
Yes.
Talk about what drove that, because that's a pretty remarkable uptake.
Yeah. So those are customers that are digital native or cloud native customers. They are not customers that are going through a replatforming. They're cloud native. And what really drove that was their demand cycle. And that's as more companies are investing in the infrastructure. You see that in the hyperscalers. You see that in the NVIDIAs, etc., that will increase their workloads. And we monetize based on workloads. So it's already a customer base that is a good fit because it's cloud native to Datadog. And we wouldn't have that acceleration within that customer base unless that customer base's workloads accelerate. We're following that.
Yeah. You also, I think you mentioned that it could cause some variability in second half numbers. Talk to us about why that dynamic could actually show up.
Yeah. So this is similar to what you saw before, which is that when customers ramp quite rapidly and they're above their commitment, they haven't repriced their contract to their volume. We basically price based on volume. And in some cases, based on term. So when they do that, and this applies to all customers, you generally see the ARR go down for a period of time and then go back up as they grow. And what we pointed out was there are some very strong rampers in here that should and will most likely adjust their contracts and redo their commitments. And that always creates volatility in the ARR revenues. It's just this is a small segment of our customer base, but you mentioned how rapidly it had increased. And so therefore, the effect may be more pronounced in the short term.
Just in terms of what some of these AI Natives are leveraging, I assume it's the whole platform, but is there anything that's unique with that cohort of customers versus cloud native sort of general Datadog population?
No. Those are best-of-breed platform. That's metrics, traces, and logs. That is essentially the same workloads that we are monetizing in our other customers.
I guess the other question that a lot of us are wrestling with when it comes to AI, and it's a broader AI question and maybe Datadog specific, but it does seem like customers are moving into more of the inferencing phase of generative AI. Are you seeing any customers shift monitoring towards more of that phase of generative AI, or is that hard to ascertain what they're actually?
Yeah, we have around 30,000 customers. And what we always do is we build out the integrations to allow them to access those AI tools companies. And we said we have 10% of the customer base, over 3,000, using the integrations, meaning they're sending us data. And we said we have hundreds in the low hundreds of customers using LLM monitoring, which would equate to around 1% of our customer base. So we're seeing some traction. And we monetize when applications are enabled, in this case, enabled by LLM and in production. So we're seeing green shoots, but again, it's a very, very small percentage of our customer base that are doing this.
That are doing it, yeah.
In production. Yeah, they're doing it. They're doing it, but where I'm talking about in production.
We're still early in that.
Still very early.
Are you seeing any outside of the AI natives? Are you, say, a bank like RBC, are you seeing usage of Datadog within sort of Gen AI usage in sort of the non-AI native cohort?
Yeah, that's what I just mentioned. Of the 30,000 customers, 3,000 of them are sending us data. This is not about the AI tools separating that out. This is about our NRRs or whatever. 3,000 of them sending us data. And then a few hundreds, low hundreds, are using the LLM monitoring module. That would mean those are companies that are not AI tool companies, but they have LLM data flowing in. So those would be. We're giving out two types of metrics. We're giving out metrics about a rapidly growing cloud native customer segment. We talked about the 6%. Now I'm speaking about our regular customers or non-AI native customers. And those are the two metrics that we gave out indicating activity, but again, not activity that is turning into monetization at a large scale yet.
Got it. Let's talk about pricing a little bit. You guys said you started out more as an SMB tool. You've obviously moved up to the large enterprise quite well. How do you think about pricing and the evolution of pricing longer term, especially when you talk about some of these customers spending big figures?
Yeah, I mean, we basically price very much the same way as the hyperscalers. We try to, which is on commits. And we basically give them access to a platform. And then they can choose what they commit to and whether they reserve or within the commitment use any of the products. And it's all in a calculator. And they can see it. It's very much the same way that it's the same way AWS prices. So that's the way we price. I think it's been a very significant factor in the evolution of Datadog in that we've priced based on use and value, not based on storage. I mean, there have been other models where there's been some concern, but we've tried to sort of lower, remove the impediments of adoption other than by consumption.
And we've basically, as clients have scaled, they've gotten a better price point, similar to what the hyperscalers do. We tend to be, for large enterprises, we tend to be a BU or a business activity type of spend. We tend not to be like Microsoft Office or their core cloud spend in that they're buying it centrally and using it across. So we tend to be based on use case and try to price so that it is completely correlated with the usage in that use case.
Got it. Yeah, and do you continue to scale the business? It's $10 billion more in revenue at some point. Is this sort of the pricing mechanism into the future?
Yeah. I mean, it's worked very well for us. We continue to evolve it. Over the years, like the last couple of years, we understand that clients sometimes need to ramp into their usage, often getting off of competitive products. So we try to give them, let's say, ramp pricing so that they can have periods where they're bringing that on and don't have to pay for that workload if they're not using it. But there's been evolutions, but the basic model has been consistent and has been a model that's helped us succeed with our clients.
Some non-core modules, like you said.
$2.5 billion for the.
For the three. Yeah. Staggering numbers.
When you look at the competitive landscape, obviously, logs is a big part of that growth algorithm. It's exploring this 1% right now. How do you think about that opportunity? Because it certainly seems like a lot of that could be up for.
Cloud logs or actually the use of logs to investigate the health of digital applications by reliability and production engineers. As you know, we said a few quarters ago that that business had gone over $500 million and has grown from there. That was business that Splunk had tried to get over the years. I would say over the last three or four years, we've been winning a good market share in that. We recently commented that for Cloud SIEM, and that would be SIEM, again, in cloud workloads that we feel good about where we've evolved the product and we see product momentum. That is not about the replacement of the centralized on-premise Splunk Cloud SIEM, but just similar to cloud logs competing around that cloud side of it and begun to have some success.
I think we get asked a lot about, is that going to change now that Splunk's been acquired? We don't know yet. It's too early. But when other companies have been acquired and put into the types of situations, we have tended to out-innovate and tended to get market share. So we'll see what happens there.
Splunk is one, and there's obviously some companies that have been acquired. Sumo is a private equity as well.
Yeah.
How often do you see Dynatrace competitively? Because they obviously do pretty well in the enterprise as well. How often competitively do they show up?
Yeah, I think Dynatrace, basically from their APM on-premise log, they're doing a good job. They don't have the complete observability platform functionality that we have. So buyers are smart. If it's for cloud workloads with some on-premise and they want the complete platform, the choice generally has been Datadog. And then if it tends to be more of an on-premise, either monitoring of on-premise workloads or an on-premise delivery, Dynatrace has a strong competitive position. In many cases, it's a choice of clients whether they want to port one way or the other or have both and have certain workloads monitored by Dynatrace and certain workloads monitored by Datadog.
When you look to the future, obviously, TAM expansion and innovation of Ollie and the team has been incredible. Where are we in that evolution of product? I don't know how many you technically have today, 20 23. I mean, what do the next five, 10, 15 look like?
Yeah, I think we have, I think, talked about that in our Investor Day, and we sketched out the service management so I think some really interesting ones that we talked about are we have tended to be a company that is a real-time analytics company, but in terms of workflow, once you identify a problem, how do you create the workflow to root it and resolve it? We really have, and that's with service management. I think Ollie talked a lot about On Call, and that means essentially not just paging, but also the whole operational fabric of getting through the documentation and routing of a problem. This is something that we have early signs compliance that they really want to pull us in this direction. We've been investing.
We don't even have a GA product or a product that's out there and I would say sold to everybody yet. We have a lot of clients who are asking for commitments on that. So that could be a vector of growth. We have the security that we talked a little bit about, which is there are three products in that. We talked about the Cloud SIEM, but there's container and infrastructure security and application security. And in terms of DevOps environments, we've begun to see traction in attaching them to APM and to infrastructure. That could be a growth factor.
We have a product that's in beta and working on in business analytics that we think has a high degree of correlation to RUM. And what that means is that that allows you to see more of how the application is affecting client buying behavior.
And we do that in terms of how the application is working and the mobile device, say, but this was business analytics. So these would be other vectors of potential growth. And then there's the whole AI set of things, whether it be Bits or the LLM observability. So we think the reason why we're rapidly innovating is that there are two things. One is our clients want more and more functionality in a single platform. And the world's changing very fast. And so that's the two reasons why we're continuing to be a product-driven company and innovate quickly.
You have $3.2 billion in cash. I believe that's the number roughly. How do you think about broad capital allocation? You really haven't relied heavily on large-scale M&A. Does that change in the future? If that cash balance continues to build as the company continues to scale? And I guess, what would be a material acquisition for you guys?
Yeah. I mean, I think that when we look at our market cap and our amount of cash, we do not feel we're over-capitalized. We feel that it's an appropriate capitalization and gives us flexibility. And so far, we've tended to invest more in technology teams that have maybe monetized their opportunity a little bit, but not a lot. And so, what could be a larger acquisition, we look at it. I don't know if these things will happen, but it could be something that is on our product roadmap because everything we do is about the product roadmap and the product vision, but has gotten farther ahead in monetizing the opportunity. And therefore, we would have to pay a higher price for it. We would only do it if the R&D team was of quality, would stay, the product architecture is compatible with our platform.
Very importantly, we got strong signals that we could retain the people that were important on the technology side.
Yeah. Yeah. Yeah. There was some news earlier this summer. I think Ollie kind of said, I don't know if it's the call or investors, that you're not looking at anything super material right now. We're big right now.
That's what he said. Yeah.
In terms of balancing growth and profitability, you guys have been really a standout in delivering really both. As we think towards the future and maybe the economic environment starts to improve a little bit, how do you think about then that balance of growth and margins?
I think we've always risk managed this well, maybe too well, when there were more risk factors, but we are really always thinking about this as a long and large opportunity and always looking for ways to invest in the product and the go-to-market, so I think we said, I think it was third quarter last year, that things were not as bad as they could have been on the top line, and we had been really good operators in optimization of everything, and so we got ahead of ourselves on margin, and we were now next going to go and work on investing against this opportunity at a higher rate, and it takes a while, and I think we're doing that.
I think we talked about on the call having some evidence of success of ramping headcount capacity significantly in the second half of this year and then into next year. It takes a while to get the people in and get them ramped. So I think that's a good example of how we are trying to push the investment forward, which hopefully on a weighted average basis, as we have shown, will be quite disciplined, but won't be in a straight line, meaning the revenue growth and the expense growth won't match in every period. On a weighted average basis over the long period, they will. But we're probably looking at trying to expand our investment, as you said, to market better, winning in the market, big opportunity, investing in both the go-to-market organization and in R&D.
Is there a framework which you think about long-term margin expansion, like sort of as a rule of thumb?
Yeah, we gave a long-term goal of 25%+ . So we understand that it's important to the investor world what happens in these micro time periods, but we're really sketching out the company to be a compounding grower with a 25%+ margin over the long term. And as I said, that'll result in periods where the investment exceeds the top line growth rate and periods where it might not. As you know, because we're a consumption model, we've always said we have to plan our investments in a certain way. And then the top line, because of consumption, could change versus that. But that's how kind of we think about it. Long-term opportunity, emphasis on growth, compounding top line growth rate in a strong economic model. And we gave the guidance of 25%+ even.
When we think towards next year, obviously, it's too early to guide to 2025. But what would you say is, for us as investors, sort of a framework? Are there things that you think about like, well, I would consider growth in logs or I would consider AI natives? What are some of the key tenets that we should think about when we consider 2025?
We consider two things. We consider the white space, which is the customers that are out there that we don't have or don't have in a material way. We plan bottoms up. So we go and we look at the accounts. And then the growth, product-led growth in the customers that we have. So far, we've had new logos, growth contribution being around 25% and existing 75%. And then we know against that that there's two things we can control. One is the ramp throughput capacity and doing that in an intelligent way. That's a broad term, but underneath of that are many, many micro decisions, geographical decisions, and then the product investment. So we really think that way.
Yeah. I'm going to pause here for a second. Is there any questions about any of those? Yeah. You can bring your mic up if it's not here.
How much exposure do you have to government spending? And can you help us give a little color on this DOGE threat we are all facing?
Sorry. The DOGE, Government Efficiency.
Yeah. Right. The Department of Government Efficiency.
So for better or worse, we have a very small government business. We came to it later. You have to invest in a separate instance and develop a team. So we've seen it as an opportunity. We really, really have a very small government business. So to the extent that there is no more government, then.
Whoa, whoa.
Yeah, I don't know. But let's just say it wouldn't affect us as much as other people. We don't have a very large government business. It could hurt our ability to grow into that, invest into that. But I don't think we're not very exposed to government today.
You do a great deal of business with small businesses. How do you characterize the strength of the SMB as an industry from your point of view?
Yeah. We tend to be, I would say, more towards the M, the middle. So we don't really sell to mom and pops. The reason why you have to have a product that's you have to be buying cloud. So you have to be an AWS or GCP customer, that kind of thing. So essentially, I would say that we have the larger side of that. And we have strong growth in both customers and net retention, but not where it was in the bubble. So I would say it's been stable and good, but it hasn't been re-accelerating back to the glory days.
Maybe to wrap it up, we're sitting here next November, looking to crystal ball. What are you going to be most positively surprised with? Is it, gee, the enterprise business is re-accelerated because the economy is in a better place? AI Natives are 10%. I mean, what do you think some of the surprises could be next year?
I think something we're all watching is the effect of AI in production environments since we're so nascent in that. So things are moving fast and if that's the case, that could accelerate the replatforming because essentially most of the applications out there are legacy applications. Most of the infrastructure, as you well know, is legacy. So those things that accelerate the migration and replatforming are positive, so that's something I think we really are watching out for. I think some of the new products that we have talked about that are exciting, like On Call or Cloud SIEM, I'm very anxious to see what happens with them over the next year and see if they become meaningful parts of the revenue.
Then, yes, the enterprise buying cycles and enterprises all over the world, what they are doing with their cloud migration and how that's moving, which is a little bit related to AI, is another thing that we watch very carefully.
On that enterprise piece, what do you think if we're on the 4Q call and you say enterprise spending is in? What are some of the KPIs you're looking for that would give you confidence and say, "This is actually getting better"?
Oh, yeah. It would be the net retention, the composition of net retention between existing products and new products, the number of new logos in enterprise, the geographical distribution of them, which could inform us on where to place our investments as we go forward, and that would all go back to the revenue, the enterprise revenue, and their growth rate within Datadog.
Yeah. Excellent. Well, we're out of time, unfortunately.
Thanks a lot.
Thanks from all of us at RBC, David and Yuka, and thanks for coming, everybody.
Thanks for having us.