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

Thanks for your patience. The elevator was a little bit of a sardine in a can situation, but we've made it. Day 1 of Citi's Global Tech Conference. I'm Fatima Boolani. I jointly head up the software research team here, and I'm very, very excited to host the CFO of Datadog, David Obstler. Thank you so much for being here.

Thanks for having us. I appreciate it.

So we're going to jump right into the discussion. Datadog has had a tremendous and significant amount of product expansion and innovation just in the past year, but certainly when I look over the arc of time since the IPO. I think a great place to start for all of us would be if you can give us an update on the platform expansion in this time frame and where you currently stand in terms of your product suite and just the respective markets that you're now touching versus 2, 3, 5 years ago?

Yes, the majority of the growth has been expanding the observability platform, started with infrastructure, and we've given over time some information on the effectiveness of this expansion. A few quarters ago, we said that the infrastructure product had passed the $1 billion, and both APM and logs had were over $500 million. We essentially have had a very good cycle of adoption of those products as we've developed those products. We further gave information last time that the next products that it passed $100 million, which were the digital experience products, which were RUM and Synthetics are further evidence. And I think a bit of a couple of quarters ago, we talked about how products that we've invented in the last 3 or 4 years, I believe, have gone over $200 million. So we -- you're right. We have been innovating in the platform, in the observability platform. Now why? Because our clients want to see as many data points and have as much automation of the workflow and analysis of problems as they can get. Problems can be everywhere. So they've had a strong desire to adopt the platform and Datadog has been relentlessly innovating. In addition to the core observability, we've shift lefted -- shift left, not lefted -- shift left to look at software creation before it goes into production. We've also developed security suite to again have security signals. We have recently talked about service management, which means handling more of the workflows within the platform whether it be on call cases or, et cetera. So these are some of the areas that we've expanded into. They are aimed at the same DevOps audience and the feedback and what we do is, it comes from them that they want additional functionality in our platform.

A lot of great pillar based threads that I'm definitely going to want to pull on and go into more detail. But before we do that, the last, I would say, 12 to 18 months, have been very much a roller coaster for investors who have gotten much more acclimated to the whole notion in concept of consumption and usage-based modalities of deploying and consuming enterprise software, right? So that's been relatively new and you've -- like I said, been on a roller cost of swords as it relates to how organization ramped -- organizations ramped and then maybe rationalize in periods of maybe economic volatility and uncertainty, et cetera. And I know you have very close and front seat to that. So can you take us through the arc of the usage growth trends you've observed over time across different customers, maybe products, maybe verticals if that's a very specific angle that kind of changes the narrative and where has culminated in that? Where are we today as it relates to normalized consumption pattern in the aggregate for the business?

Yes, yes. Good question. So a couple of things. One is we sell on a consumption model with commitments. So essentially, it's not pure. You can -- we don't get any revenues. We don't use it. You commit to it, and then you get credits and use it along the way. And we tend to sell, I would say, not in terms of what the usage might be 2 or 3 years out, but what the usage is at that moment. So in terms of a couple of things that some of our competitors or other companies that face is we haven't had the issue of shelfware or pretty much unused. But what happened was in the COVID period, we had a rapid expansion of businesses, cloud native businesses and consumption. We also had a capital markets that was encouraging companies to spend and emphasize growth not profitability. So particularly in cloud natives, we had, I would say, overspending. And whereas normally, there would be optimization all the time, there was culminating, I think, peaking in the second quarter of last year the rationalization of that over consumption from the post-COVID period. And if you look at the time series, the most intense part of that was in the second and third quarter of last year. What we said, it was across the customer base, but mainly focused in large significantly ramped cloud natives. Those were the ones that first optimize. Because it's not a seat model and because it -- you can modulate this quickly, we probably saw that sooner than other software companies, but we also saw the recovery of that potentially sooner. Because if you look at the time series, once they did optimize by the fourth quarter of last year, we saw that cohort stabilize and start to grow again. And that's where we are today. We've seen essentially resumption of more normal activities, growth of those cohorts that were weighing on our top line growth but not at the same level as they had in the bubble. Further information across the customer base. For a while, it was fairly pro rata in enterprise, mid-market and SMB, and it was more correlated with cloud nativity and ramp of spend and that lasted in that period of optimization. But what we said in the last call is that we saw enterprise break out to have a higher net retention, meaning there was less optimization or more return to normal growth in enterprise, which for us are companies with more than 5,000 employees, whereas SMB had stayed stable, still growing unlike some of what other people have said, but not growing at the same rate. And we also said it may be [ too little ] or that may not be enough quarters, but sort of we wanted to update everybody on a divergence in that trend of net retention between enterprise and SMB. Enterprise has been traditional industry has been behind, probably didn't overspend this much and is acting more normal in the time series since the optimization trend.

David, I wanted to go on a little bit of attention, specifically from the vertical lens. Some of the discussions and certainly conversations I've had is, hey, the generative AI cohort has been a very exciting area for you in terms of -- I think you've quantified it about 4% of your ARR is coming from, I guess, from a B2B perspective, these generative AI companies that are using the entire Datadog footprint to monitor their own environments, right? So I think there's maybe a degree of debate happening as to, hey, some of what you saw with the cloud native in terms of how hot and heavy they got with the spend and having to rationalize. Is there potential for that situation to transpire with kind of the generative AI because it's so new, it's so embryonic, it's growing so quickly. So how would you maybe debunk some of those considerations or concerns? Or rather, can you draw any parallels between the behaviors of the two such that 8 quarters from now, we're not potentially talking about, hey, generative AI companies are now in a rationalization mode?

Yes. I mean we have many sectors, and this one, we've said is 4%. So we -- certainly, they're ramping quite rapidly. There is a lot of usage. You could well have an optimization, but we've had optimizations in end markets that have been a lot larger than this. In this cycle or even, let's say, at the beginning of COVID, we had in hospitality and airlines and things like that. So yes, there could be. But secondly, it's not that larger part of our business.

Understood. And zooming out your exposure to more of the VC-backed companies who are naturally more cloud native, how should we internalize that in terms of aggregate exposure? And from what I'm gathering from your observations and what you're seeing in customer behavior, we're definitely past sort of the open-ended period of ramping spending and ratcheting it back. So maybe there was a degree of stabilization and even the way the cloud natives are spending, but how would you articulate -- is that a correct observation in that, hey, the cloud natives are just better at managing their own businesses now? And to the extent some of that risk doesn't creep back in because you do have exposure to...

Great question. So we've been pretty much 1/3, 1/3, 1/3, enterprise, traditional enterprise. It's a little more than that. It might be around 40% mid-market and then SMB, where SMB, yes, it is -- has a high intensity of cloud native, venture backed cloud natives. So we've been highly diversified and increasingly diversified. Everything from companies like plumbing supply companies, you would never expect to have a digital presence but have warehouses and have to distribute. So that would be traditional industry all the way to what you read about in cloud natives. And yes, I do think that that when you look at the cloud native sector and the SMB, what you see is they are growing and they -- but they're spending in a more responsible way. That shouldn't surprise anybody. The whole capital markets and availability of capital in that period was encouraging, spending growth without looking at profitability or maybe even exits. So yes, we do see that behavior, but our SMB, which happens to be probably more like M because in order to have a cloud presence and have hosting and all, you're not a 50-person mom and pop. So I think this is probably more towards the mid-market. It has, like we said, been solid but hasn't accelerated to the degree that the enterprise has. So yes, I think they're spending patterns, and it's probably very healthy long term for those companies.

I appreciate that color. Thank you. Let's go back to kind of the product footprint and how the platform is being deployed and scaled within your customers. So I want to start with what is your largest, most dominant kind of most mature solution area, which is infrastructure monitoring, right? That's your climb to fame. $1 billion-plus ARR footprint. What observations can you share to give investors the confidence that, hey, this is still a growth engine for us, even though it's our biggest? And why there is still significant runway for the infrastructure monitoring set of solutions to continue to have very strong growth?

Yes. It's -- that's the part that's most correlated to cloud workloads, and to cloud workloads of modern applications. The percentage of workloads that are in the cloud are still -- I don't know if it's 20, 30, but still pretty small. There are many enterprise -- many entities like enterprises that are very early in their journey. The hyperscalers and Datadog and infrastructure have experienced long-term and sustainable growth for many, many years. And we think we're still early in the transformation of applications into the cloud. In addition, we could have accelerants like AI where you have to replatform. You can't leave the old applications. So all of these drivers in the technology base are helpful to the replatforming of modern applications, and that's what Datadog does. So we think we're still very early in the long-term and very, very sustainable trend of putting applications into the cloud, modern applications.

You all have a very close relationship, and Oli feels very strongly about all the work you're doing around open telemetry, right? So the whole concept of just make it easier for enterprises to firehose and ingest for data. So you just -- you see more, you know more, you can do more, right? I think functionally, that's the thesis, right? So when you think about open telemetry, how do you kind of navigate some of what might be less monetization opportunity for you because you are democratizing the ability to gather more data? I mean, how do you kind of straddle that openness to being open telemetry but also your -- you've got commercial software intellectual property, how do you navigate that?

Good question. Well, to be the platform that you do DevOps on and that you standardize your workflow is a really, really important birthright. And essentially, in order to do that, you have to be open to getting everything in, and we really don't care. Once data is in there, there are so many ways we can monetize it, and there are so -- and it becomes easier and easy. That's what we've been doing, building the platform. So it's part of a strategy, which, again, back to Oli and his philosophy has been really successful of being the ubiquitous platform that everybody can use. And if you do that, you have to be open to get data in from everywhere. We found that opened so many vectors of growth because once you get the data in there, you can do so many things on it. So that's the strategy.

I want to move to the APM side of the house. Still a relatively new product area for you, I think, introduced maybe 3 years ago, in earnest 4 years ago, in earnest.

Five years ago -- 5 or 6 years but probably it took us a couple of years of bill to get up to product parity and beyond.

So a lot of interesting things happening on the application performance monitoring side, very natural kind of -- the swim was very natural for you. I want to tackle this with a number of different angles, right? So first and foremost, the consolidation generally we've seen in the observability arena. As a general matter, how has some of that shake up in the market structure impacted your business and/or the types of conversations you're having with customers in terms of what type of more visible opportunity that brings for you from a wallet share perspective?

Definitely, definitely. There -- our -- the clients have evidence wanting to have everything in a platform. We didn't have APM. Over 50% of our deals over $1 million involve consolidation. And a lot of the consolidation involves APM and logs. So it's a huge opportunity. We're not that penetrated in APM. So we've -- how we reacted to this, we've gotten better of value selling, selling the platform, organizing and understanding when's the right time to consolidate. At the same time, we benefited from a benign or weakening competitor set. And as we have out invested, I think our R&D is more than everybody else combined, we have been able to, as you said, it's a smooth track and it is, and we've been able to consolidate this. So that's been a significant vector, not only APM, but all the things that hang off of it in applications. We announced on our last call that we had both RUM and Synthetics, this is digital experience, or how the application works out to the mobile device for the -- or website that those pass the $100 million. So there are many more things that hang off of this that we've been able to bolt on as we've consolidated this APM business onto our platform.

So two-part around this. In the heat of the pandemic, I think APM and logs tended to see the most upward elasticity, right, in terms of usage, and that's probably been where you've seen...

And probably infra and logs.

Infra and logs.

But two different reasons. One, logs has -- logs can be -- you don't have the instrument. You can start -- you can just run more logs. So it's about the settings. And infra because there wasn't as discipline of a hygiene of managing hosts. So I would say those are the ones that -- APM was a little more cushioned in that area.

Okay. That's very helpful. So outside of just more rational buying behavior, what have you done institutionally or the go-to-market organization has done institutionally to just be mindful of not being caught off guard in terms of erratic customer spending behaviors, right? You clearly have learned through the pandemic. You have a lot more data points. So what have you done intrinsically to just have more visibility that the higher elasticity and infra and logs, some of that you could try to have more under your control, so it's not as erratic?

Definitely. We try to increase transparency by instrumenting the platform, so clients can see their usage patterns. We have a couple of different departments, including customer success, and solutions engineering, the technical account management that are available for customers to help them regulate this. If they've done something that is not intentional or they didn't think through it, we basically say, we're not going to charge you for a spike this has been going on for some time, but even more important than this. Let's work together to get this calibrated. We've -- in selling, we've become more flexible in our drawdowns and commits. So essentially, you can use all the different products, there's different SKUs. And we have technical, we have both a paid and a not paid service, it depends on the intensity on helping you use the platform in the right way. I think this has been very important in the evidence being that our gross retention has stayed very stable and very strong throughout this and part of it is because of this way of dealing with clients.

So you brought up gross retention, so I have to bring up net retention. Just from a net retention rate standpoint, a number of different vectors here that contribute to the outcome or the output of that KPI. You were punching very much above your weight class. You're still kind of upper watermark for all the enterprise software net retention rate levels in the aggregate, but we are below the historical high watermark, right? So can you talk to some of the key drivers that can really fuel a re-expansion and an improvement here? And what is the balance that you're trying to strike between increasing usage of existing products, flagship products that everybody knows and loves you for, but also driving more multiproduct capabilities?

Definitely. Yes. So about 30% -- 25% to 30% of our net retention, this will be the growth of a customer in 1 year. So this doesn't talk about the ramping of new products, it comes from new products. And so what we do is we essentially try to educate the customer on the platform that's in the customer success enterprise selling, so that they know products are there. There's also marketing efforts and product marketing efforts around those products. There are metrics that we provide that are all pretty much up in terms of the multiproduct adoption. And we have incentives in a number of departments on their compensation and their plans that are based on cross-selling. So all that is part of a package in trying to get more adoption. Now why is that important? The more the clients using pieces of the platform, the stickier they are. The more they're standardizing, the more they're going to stay with us and the more it's going to grow over time. So it's all part of the strategy of how to deal with clients at Datadog.

Part and parcel to this is your engagement with customers more and more on a multiyear basis, right? We can see that in some of the backlog metrics you disclosed. I know you don't manage the business to that. But as you think about driving this type of multiproduct behavior, as you drive incentives to push those outcomes, the observed trend of customers doing more multiyear deals, how is the go-to-market organization evolving, maturing to kind of those types of engagements? And relatedly, how is the philosophy around ELA, EAA, bundling type conversations? How are those evolving? Because I know historically, you've taken a very firm stance that, hey, we don't want to just offer a buffet. We actually want to see what customers are using, so we can extract value and understand where we want to invest. So I get that ethos, but large companies are large companies, they're used to doing business in a certain way. So how are you combating some of those demands from larger enterprises?

Well, I mentioned that we -- for larger enterprises, we generally sell on credits or commitments. It tends to be flexible within that and that they can use the parts of the platform. We do believe in the transparency, so they can see what parts of the platform they're using. So we have maintained a SKU-based strategy. So those are things that we think are valuable and allow the client to calibrate their use. We've been on a journey over the last several years, but it continues in selling -- combining bottoms-up selling with selling higher and higher up in the organization and navigating vendor management to try to identify what the use cases are in projects and the other vendors that are in there and try to work with the client to get to a long-term -- a commitment that they will themselves trade off between the flexibility on the short term and long term. I know I manage vendor management. And I know that if I -- I'm going to stick with the vendor. I know I'm not replacing them. I can -- there's a lot of flexibility in a multiyear deal. I can always buy along the way and many do, but it helps to manage the price point. We do it in a disciplined way. The unit pricing hasn't really gone down, but we try to balance this out. And it's worked for clients in that they are able to both commit and have more transparency in the amount of usage with us, and we get them on a longer-term commitment. So it's been a win-win largely pulled by clients.

On this trend around upsizing transactions and deals, we've seen a lot of momentum there. Can you frame for us a vanilla infrastructure monitoring customer? When they do, their eyes light up and they see the value from APM. And then you've gone further, see the value from synthetics. Can you talk a little bit about quantitatively how that influences the scope of the initial commit? And then -- these are maybe some of your more established areas, how does the conversation change even more when you talk about security, when you talk about cloud service management? So anything around uplift or contractual size or commit size improvements as more and more goodies get added to the commit?

We're still land and expand. So we still often don't take the full bucket upfront. So I think 80% of our customers land with two or more. So generally, the motion is you land with infrastructure and either logs or APM. We then identify the most obvious ones you're talking about. There are other things that tend to get bolted on really, really pretty easily. I think we talked about the digital experience synthetics, RUM, database, lots of different sources. So all of that, as I would say, continued to be fairly frictionless. Of course, we educate them, we eliminate them, we put it out there. And I would say that, that all is within the observability suite. It could well be that service management, we're talking about it, which is really how workflow gets done in the platform, are on call where you're able to manage cases in paging and all, it's too early, but they may also be frictionless as part of the platform. And then when you go to security, you have a less traditional use case. We're selling to DevSecOps, which is still in -- still emerging as a use case. The more progressive cloud natives are doing it, but you also have a different buying center in a CISO. So that's proven to be a little bit more of a challenge. What we told everybody is as we're building out the product suite and getting to product parity, we will look at the pace of DevSecOps development and see if we're going to be able to do what we did in DevOps, which is attached and be a winner potentially alongside the more traditional security companies or we have to adjust our go-to-market. And we haven't been at that point yet, but we may have to adjust our go-to-market or our marketing to address this other use case. It really will depend on how this gets bought, whether it gets used decentrally but bought centrally, whether it gets adopted ground-up or whether it really gets used centrally. And that's still an evolution, which is different than what's happened in DevOps so far.

Where are you performing from a security standpoint relative to your initial expectations? I think there's still some tinkering going on with the portfolio for the product and go-to-market standpoint, you just alluded to that. So where is it performing at current levels and where do you want it to be?

Good question. Yes. We, basically -- in terms of the number of customers that use it, which I think we said is over 6,000, that meets our expectations, et cetera. But I think what has been harder is the use cases and the price point. So I think what we are still doing is building out the functionality. This happened in APM too, where it took us a number of years. So we think there's a very large opportunity. I mean we have -- even though we have 6,000 customers and quite a bit of revenues, we haven't said what that number is, but it's a good business, but we think it has so much more potential. So I think that it's for the reasons of the build-out, not having the security team there, making some aqua hires and then the buying patterns, perhaps it's been slower than the other products, but it's also evidence of some of the things that we've seen in things like APM.

What are the types of -- I don't know if ARPU is the right way to think about the DevSecOps opportunity? I think not too long ago, you decided to institute more bundles, right?

Right. Yes.

So that's been one approach to kind of really galvanize the momentum in DevSecOps. Any feedback on what the reception from customers have been to this? Because the persona is adjacent, but not exactly the person that you're saying the portfolio to. So any lessons they've learned in the data points you've been building via the sale of DevSecOps bundles?

Exactly, exactly. Yes. It's working in that the bundling exactly right. We went to a bundle because we felt that in this...

[indiscernible] bundle for us were...

We have infrastructure and app bundle. And it basically means you're buying the APM plus the app security or you're buying the infrastructure plus cloud security or host security. So given what you said, we are sort of thinking that this could lower the barriers to adoption to this adjacent group, and it's working. It's too soon, but it's growing quite rapidly. We're having hundreds of customers adopt it, which is good, but we're still only up -- I guess, we're 2 quarters in -- 2 or 3 quarters in. So we think it works. We think we've learned a little bit about how to lower the friction in going to market, and we're optimistic here.

You're pursuing all of these growth objectives, multivector, multiproduct kind of tweak in, jiggering the go-to-market as you go along, specifically in these areas and cloud service management, you talked about really ripe opportunity to monetize workflows for organizations that are already captive inside Watchdog, right? Just from a financial model perspective, it doesn't seem like you've missed a beat on delivering incremental profitability while still pursuing these growth objectives, right? So what's the sauce here? Why are the unit economics so high? And I mean the other angle I want to ask you about, as you've said this earlier, you do have your sales folks and your sales organization. They're moving upstream selling to enterprises. Those are harder to sell to, longer sales cycles. And yet, there's still not -- there has not been a significantly deleterious impact on your margins. So what's the secret?

I think the secret gets back to the product and the architecture. The fact that it is efficient to add functionality because it's all based on a common data architecture. And so unlike other companies which have had struggled to get everything together, we've been pretty efficient. And because of the way it's architected, and we talked about frictionless adoption, it's a very efficient way to sell, which has helped us manage the sales price point. So I think it all goes back to that, and then it's a decision with that type of efficiency, how do we reinvest? Where do we reinvest back in, in order to have the economic model that we've shown? And so far, there's been -- that's been efficient enough and the economies there have been to reinvest and still improve margins. And I think we've been making many of the types of investments that you're alluding to. Enterprise salespeople, sales engineers, channel and still fitting that in, in the model and being a leader in R&D. So I think it all goes back to the fact that it's a mousetrap that is efficient in terms of the ability to add on to it and have clients adopt it.

Is there anything you can share with us on, hey, each incremental product, 50%, 60%, 70% of it drops to the bottom line? Again, maybe trying to quantify a little bit of that efficiency of just the core product architecture that allows that frictionless adoption. So any sort of quantitative finer points you can...

I think you see the gross margins, which, as we've been adding products have been stable to up. That's not possible unless you were engineering gross margins for incremental products that are near where the overall is. So again, it goes back to the engineering of the platform. Now I think we've been hoping this because we've been good engineers when you think about flex logs or how we are reengineering the platform all the time. Half of the investment in R&D is platform investment. And so we're reengineering the architecture of the platform to relentlessly make this the case. It isn't that products -- they haven't had been like this or this. We can engineer and price a product to have a gross margin that's similar to our company gross margin.

David, I want to end the conversation on how you're thinking about the journey to $5 billion in ARR in revenue from here. What's the force ranking of the things that you're focusing on to get you to that -- into those revenue levels? And how does M&A factor into that strategy?

Yes. Good point. I think Oli has said, we feel we can get to $5 billion or beyond with the growth trends and our product suite and observability and we have optionality above that in shift left, in security, in -- I think we've talked a little bit about business analytics. So there are other things. So essentially, we feel that the growth factor is back to only 20% or 30% of applications in the cloud can do that. Now M&A for us, largely has at the very center as when we have a product road map and our limitation has been the getting enough engineers in and getting them going on it. So we've been looking for engineering teams that can accelerate the product road map. It's -- so far, it's been pretty much -- you look at this box, you say, here's a good group of engineers. They seem to have the way of thinking that we can integrate in. And so that's been the core. Now we are open, as we've talked to, to looking at M&A that might be larger, but the bar gets higher because we're a platform company, the people have to want to stay and we are pretty disciplined on price in acquisitions. So we're willing to look at things. But the core of what we're looking at is that aqua hire or the technology. And I think we said on the earnings call that we're not looking today at anything that's material.

David, I want to thank you for a fantastic discussion. We covered a lot of ground.

We did.

I appreciate the insight. Surprise when you're having fun.

Yes. Thanks a lot. Thanks for inviting us. Thank you, everybody.