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

Good afternoon, everyone, and welcome to day 1 of the Morgan Stanley TMT conference. This afternoon, we begin with another great session. Really pleased to have, for the second year in a row, Olivier Pomel, Chief Executive Officer of Datadog. Olivier, thank you for joining us.

Thank you for having me.

So we're looking forward to having an insightful discussion. Before I get there, let me just quickly go through some disclosures. For important disclosures, please see the Morgan Stanley research disclosure website at www.morganstanley.com/researchdisclosures. If you have any questions, please reach out to your Morgan Stanley sales representative.

And with that, the way I wanted to begin the conversation is sort of a look-back into 2020, Olivier. And just sort of from your perspective, how has the pandemic sort of changed your view of market opportunity. So looking forward, but then also looking back, what were sort of some of the puts and takes from last year in terms of how the pandemic impacted the business?

Yes. So I mean I won't be the first to tell you it was a very, very special year in many ways. Overall, I've been super impressed by our team. I think everybody had a standard performance in what was a very difficult year. We saw our team stand up and show up for the occasion and change the way we did things and really deliver. We had a pretty good year by every single metric. From an operational perspective, we actually had to make a lot changes in the way we go-to market, and we used to generate a lot of leads by going to in-person events and getting in front of our customers. And that's something we had to redo, and we've been pretty successful at it. So it's been an interesting year for that reason. From a market opportunity perspective, really, what we saw this year is that the cloud emerged as a clear winner and a destination for the future. We saw that the companies that were in the cloud, through the pandemic, were able to scale up, scale down, retool, adapt very quickly while the companies that were not in the cloud were not able to do that. And I'm not just talking about the folks that have an incredible increase in traffic. So we have customers in [indiscernible] that obviously had their operations skyrocket in a scale of 10x in a matter of weeks. But we also had a number of customers that just had to organize their operations, move things from 1 place to another, shifting applications, and really has been able to that pretty well in the cloud. So that's been a great validation. And we've seen, more recently, companies that are extremely negatively impacted by the pandemic, engage in digital transformation and the cloud migration the way they hadn't before. So I think it's promising for the future.

And as you look to 2021, Olivier, were there any headwinds in 2020 that were -- that if I could turn into accelerators or tailwinds going into next year, whether it's better customer acquisition, better product expansion, what's sort of your base case expectation of how 2021 plays out versus 2020?

Well, there's definitely a few things that are -- that makes to go better, right? I mean the -- obviously, on the customer side, as the world emerges from this pandemic, there's going to be more opportunities to do -- to commit to the future and move with more certainty and be a little bit less conservative. We will see many more go-to-market opportunities and modalities is opening up again. I mean, I was talking about in-person events is just one of them. But these are things we didn't have in 2020 that we expect to get back at some point in 2021, maybe. And we also expect as a business that we might see less volatility month-to-month in terms of what customers adopt and expand in their cloud environment. So it's -- what we see in 2020 is as the effects of the pandemic have rippled through the economy and through the world, we've seen a bit more volatility from customer behavior that, in the past, would have been equalized, I would say, across the customer base and everybody's situations. So all of that is going to be good things. I think it's fair to say also it's going to be a transition year, and also there might be some headwinds in 2021. I mean, for example, some expenses that we didn't have in 2020 ever to come back. We will start travel again or do things we haven't done before. It's also, I think, healthy and prudent to assume that there might be a drop in productivity at some point in the year. I mean if you think of ourselves, I mean, just myself and both also our employees, we can't wait spending more time with our families and traveling and doing things. So if everything goes right and we emerge from the pandemic, I think later in the year, it's likely, we'll see a little drop in productivity at some point, and that's healthy.

Yes. Such a great point. It would be nice to be human again after living a pretty crazy lifestyle for the past year. It's such a great point. I wanted to lead the conversation a little bit take a step back and just sort of talk about product engineering and product development at Datadog. I think it's one of the key things of the story that most impresses me about the company. I look back to 2016, you were a single-product company. You had mostly in infrastructure monitoring. And then today, you have 9 to 10, I think, paid products. If you take us back to that time, what was the sort of thesis on how the market would evolve and sort of got you on this journey to become a platform and a multiproduct story? And maybe related to that, what signals do you look for? And where do you find these signals to sort of guide product engineering at Datadog to tell you where the market's going, where the puck is going.

Yes. When we started the company all the way back in 2010, the outstanding point was not to build a better monitoring product, it was to bring different teams together, bring dev and ops at the time. And really, to be able to have a platform that would act as a layer to integrate all the various data streams they had and present it across teams and get the maximum number of eyeballs on it. And still, for our G&A or what's behind companies is to really reconcile, bring as many parts of the addition to the party as possible. And so we always -- we're building it to add more than just one product. In the very early days, it was actually very difficult to finance the company because nobody understood this integrated platform approach. It was hard to understand who it was going to face against the pure-play vendors, infrastructure monitoring APM, among other things. And we ourselves had to make a choice early on to focus with the very first product on infrastructure monitoring use cases and focus on companies that move into the cloud with that, which as we've seen, has been successful today. So in terms of who we get the signals to keep developing and add more products, the first thing is we spent a ton of time with customers, and we focus -- before we actually bring solution, we focus on understanding their problem for them. And that goes back again to when we started the company, it was difficult to get it funded. We had many competitors funded with [indiscernible], and so we're very scared of getting a problem wrong, and spent all that time with customers, making sure we actually understood that problem before we build a solution to it. What we saw when we talked to our customers early on was that they had tons of tooling, but it was not very broadly used. Even deployments that were successful were typically used by a handful of 10 power users and everybody else will be scared of them. When we started shipping our products, we saw our customers engage very broadly with our platform. We saw every single engineer, every single of ops person, every single developer actually engage with our platform every day. And that's what gave us the -- in addition, basically, to keep developing and keep adding to that platform.

That's great context. As we think about the progress on that multi-product, which I wanted to step through that in just a bit. And then maybe we start with core infrastructure monitoring, and I guess the context here is about 70% of the base is running 2 or more products. I think over 25% of the base is running 4 or more. So a lot of traction on that front. But in terms of core infrastructure monitoring, how much runway is left in terms of that opportunity? Is there any way to sort of frame out the opportunity for infrastructure monitoring in terms of greater cloud adoption, movement to Kubernetes and microservices? How do you think about that?

So we think the very first product we ever brought to market in infrastructure monitoring is still extremely early. The product migration itself is early. And our product, our infrastructure product is actually growing faster than the cloud providers even though their revenue include more than just infrastructure. So the one way to think about it is that, I mean, it is somewhat mature like it is understood, but it is mature in the way the CRM market was mature in, I don't know, 2007, 2008, meaning it's still the beginning of what's a very long transition, and there's a lot more market ahead of us.

Yes. And actually, just to dovetail on that point, you mentioned Salesforce, and we've talked in the past on how Salesforce is one of the companies that you sort of look up to. Why is that? Like what about the Salesforce story provides a good analog for what you're trying to do with Datadog?

Well, I think they're one of the dominant platforms in the enterprise, right? So when you think of the various job functions, the front office runs pretty much on Salesforce. That's the platform that everybody is using. And any tooling that is not made directly by Salesforce run on their platform. So I think that's an equivalent of what we want to be for the developers, the operations people and all the people who care about what the applications do and do for the business. So we think there's room for another platform like that at that level in the enterprise. We think we are in the early days of that. We think it's going to start with observability. Observability is going to be a big part of it. And we're building towards that.

That's great. And so I wanted to continue the product level discussion, and I want to hit on 3 things. I want to talk about -- there's so much to talk about in terms of product, but a couple of areas I wanted to focus on was APM logging, and then I want to talk a little bit about security. But starting with APM, my question is, what's the role of APM in a modern cloud environment versus when it was primarily a point solution market. Like what's the key distinction today in terms of how people use APM, how the users use APM versus maybe 5, 6 years ago?

It's actually a very different landscape. The landscape, especially in cloud environments, is a lot more dynamic and more complex. So it used to be that an application was in like 3 tiers. You had a application, then you had a database on the back end of it, and you had a front end of the application. And basically, all you could do is sit on the web application, and you see all of it and understand everything that's going on. Today, most of the applications, especially in cloud environments, are going to have multiple tiers, multiple services, multiple micro services, in some cases. And anytime you process a customer transaction, in your transaction, you have to go through 5, 10, 15, 20, 30 services, and each of them are going to have multiple tiers in turn. So it becomes extremely complicated to understand what's going on, what's happening into those applications. The other thing that's happening is you have many more teams that are actually participating into that. It used to just be where you have the business application team that's going to look at what's happening for the APM. Today, you're going to have people that span development operations just because those teams are somewhat conflated in cloud environments. And then the last part is there's a bigger footprint for these applications. They're becoming the main way for companies of whole sizes to innovate and to interact with their customers, and so there's many more of them that matter to many more users. And so when you combine all of that, like it's a very different game. Like it's a lot more complex, a lot more participants, a lot more [indiscernible] higher impact. And so we see products -- functionality for APMs will have changed a lot over the past, I would say, 5 to 10 years.

And as a follow-up to the APM question, I think you entered that space in 2017. How have you -- how has the team been satisfied internally about the progress of APM? Is that sort of the adoption curve been faster than you expected at a higher level of attach rate than you sort of imagined in 2017? Or is it sort of in line with what you guys were thinking back then?

So we're thrilled with the APM product. It's one of those products that, initially, in 2016, 2017, when we first launched it, had a bit of a slower start. For one thing, you needed to get some maturity and collect -- support all the languages our customers had. But also, it's a product that's a little bit higher friction because you need to pass the application to include it as opposed to just deploying it on each server, which was [indiscernible] product. But it's been a very, very steady grower, and it's a product with a very long fuse. So it's interesting because we see it growing for a very, very long time. And we see also the percentage of applications that are -- I think, one of our customers that are being covered with APM, growing over time. I mean, there's this famous quote from Gartner that gets used -- I think I only used it once today, but it's become a ceremonial quote almost, and it says 5% of application use APM. And that number is already a lot higher from what we can see, and it's growing over time. And it's not going to stop growing. I think it's going to convert to being a very substantial fraction of all of the applications, of all of the infrastructure with APM. So there's a very long fuse on it. The other thing that we are very satisfied with on our product is that even though most of our customers or most of our deployments are in net new greenfield environments in the cloud, we do also see some displacements once in a while. And we are successfully displacing pure-play APM vendors in the enterprise, which is a sign that our product is fully mature. And on its own right, if we so wished, we could lead with APM as well.

Yes. That's very interesting. And that way, in terms of like the adoption curve and the attach rate. I mean the infrastructure modeling always had a relatively high attach rate. But see that APM start that convergence, that could definitely be a runway for growth ahead. As we move into the conversation on log management, and I guess, the context of this conversation is for stepping into these products one by one. But of course, it's a little bit silly, right, because there's a platform play going on. But in the context of Datadog and log analytics and log management, it's a well-established market, there's some clear incumbents. How is the Datadog log management story differentiated in the market? What are you trying to solve for that may be different than what a Splunk or Elastic or some of the other players are going after?

Yes. So you've touched on the first point, which is that it's only one part of the platform. And we see logs as one of the many signals customers are going to use to understand what's going on in the applications. So we don't think of it as being a product in isolation. Now if you think of the broader positioning, we're fully cloud-native, cloud-first. We're not starting by going after all of the existing corporate IT applications on-prem. That's not what we're going after. We're going after product applications in the cloud. And so the focus is a little bit different to start with. The other, I would say, main differentiator with the competition is that we built a product for maximum usage across team members and across teams as opposed to other products that were built to be used by just a few specialists and then everybody else would be scared and they wouldn't touch that. So when you look at that, fully cloud-native, fully integrated across infrastructure, APM and logs and built for maximum usage across teams, that's the value offering of our product that I don't think anybody else offers today.

Great. And so the last product area that I want to touch on before we move to some of the topics around go-to-market and maybe a little bit on pricing, is security and security of Datadog, but security more generally. Maybe to kick off the conversation is how do you see the opportunity, Olivier, as companies try to move to a cloud operating model. Is this sort of similar to like 2017, where you started to see that convergence between the devs teams and the ops teams, and this is just an evolution of that? What is Datadog's role going to be in terms of security as companies move to the sort of cloud operating model?

You understand it well. I think the -- this is exactly the same kind of motion we've seen when we started the company. When we started the company, we had problems with dev and ops not speaking to each other. And writing and running provisions was very hard to operationalize for that reason. And today, we see the same thing with security. The security teams are not on speaking terms with the development teams and the operations team in the organization. Even at Datadog, it was difficult, initially, to get the security teams and the dev teams and the ops teams on the same page. So there's a real problem there. It's a problem we're good at solving. It's part of who we are as a company, and that's -- I think it's a very logical extension for us. The other part that is very interesting to us is that it is really hard to instrument environments for security, especially in cloud environments, where the complexity is shooting up. And so who we can help there is that environments, so applications, infrastructure and everything in between, is already fully instrumented in Datadog, whether that's APM traces, infrastructure management, [indiscernible] logs, whatever the end users are doing, what the developers are doing. Everything's instrumented at Datadog. And so we can use this information readily in our security product without our customers having to capture it twice. So this is what's specifically interesting to us. And obviously, the market we're going after is production applications in cloud environments, which is also very much of a greenfield market, very similarly to what we've done initially with infrastructure monitoring.

I think it's important that this is a greenfield opportunity. I think a lot of the questions I get is that, is this just another sort of playbook on Splunk or Elastic? But I think the focus on cloud applications is a little bit of a different paradigm. And I wanted to see if you can pretend to do my job for a couple of minutes, which is what's going on in terms of the broader market? Because you're starting to see some of the security players, Zscaler, CrowdStrike, come into the infrastructure monitoring side. And so just sort of your view on what's happening from like the modern security guys? And then, I guess, the ultimate question is what's the matter and where they come from? It's similar to the APM versus infrastructure monitoring debate. Are the security guys well positioned to do this unification that this platform play you're doing? Or is it -- do you need to have -- come from more of the data angle from the infrastructure monitoring and the APM side? How do you sort of frame that debate and sort of make the argument for Datadog to do well on this opportunity?

Yes. So at first, it's hard to comment on what others might be doing. I would say the focus on some of the companies you mentioned is still very much on the corporate side of things, not the production applications. So not the things that are going to be manipulated by developers, basically, which is where we are. So I would say if there's a collision course there, it's probably quite a bit in the future as opposed to being where we're going today. Now if you want to think in terms of what's the high ground strategically for security in the cloud, we think the high ground is, whatever has the highest surface of contact with the customer, which, in our case, means the amount of infrastructure and data points that you interact will be there from the customers' environment, but also, the number of users you have inside the customer teams. Are you -- is your product basically something that is being manipulated by 5 or 10 people in small security operation team? Or is it something that reaches 1,000 engineers or 5,000 engineers? And I think, obviously, our take is we want to have the maximum level of usage, and that's how we become more valuable in the long run, and our customers drive more value in the long run. But we see a lot more building we need to do to get there.

Great. And then the last question to wrap up the conversation on security is just the go-to-market sales side. What's the latest thinking on the team about how to approach this market from a security sales perspective?

Yes. So like them, we're very early in the life cycle of the products. We're still in plan and build mode. And so we don't lead with the sales team for our security product. We let our existing customers sell, select and adopt them. And at some point in the future, when we've reached the level of functional coverage and product maturity we want, we will actually start actively selling and pushing that to other customers. At that time, it's possible we'll have to specialize some of the sales force to sell to security. It's also possible we won't have. I would say the main determining factor there is how much of the company is moving to the cloud. We'll adopt DevSecOps versus remaining with same kind of security practices they have today. If they add up DevSecOps, we can probably keep using some of the same bottom and go-to-market motion we're using across our existing products today. If not, that's fine. We probably will have to have a bit more of a top line motion in our security. So we don't know yet, and we don't expect to have to figure that out for a number of quarters.

Understood. The next sort of topic that I get asked quite often, particularly in the last year, is just around pricing. There's been a lot of vendors that have changed and altered their approach to pricing. And it's a similar question to like the product discussion. Maybe just to start off with, if you could just sort of describe the Datadog philosophy to pricing because I think it's a very specific creative logic with respect to your pricing. Maybe describe that. And then for those that -- for those customers who they feel like they get bill shock, or they feel like they don't have enough visibility into their costs because they're using Datadog in a significant way, what sort of flexibility are you providing those customers? Or just addressing some of those concerns around cost management and visibility. But first start with the philosophy in terms of Datadog pricing.

Yes. On the terms of [indiscernible], there's really 3 tenets to our pricing. The first one is we price in a way that [indiscernible] let's us land and expand. So we can start very small with a customer, and we can go with them. And we go through then having a larger cloud footprint or using more of our products, more SKUs. That's number one. Number two is we're not a price disrupter. We actually -- our goal is not to be followed in somewhere at the bottom of the deal. Our goal is to actually have a conversation about pricing at scale with our customers, and we get signals from that. That's how we know we do something valuable. And the virtual cycle there is when you ask a customer, for a lot of money for a product, they're going to say, okay, that's fine. I can tell you, but you need to solve these problems for me. And if you solve them, then you can actually get more money and it's a virtual cycle. It's much better than if you're a pure price disrupter, the question is always, what can you take out so it can be cheaper, and that's not what we want. That's the second point. The third point is, we have to be able to align price with value. And for that, the only way to do that is to keep the customer in control and give them the levers so they we can keep that alignment. And to give you more color on that, so everything that relates to observability, you can generate absurd amounts of data in logs from any application. Amounts of data that are going to be completely de-correlated from the business value these logs are going to bring. And there's no way to deal with that through pricing. I mean, you could say, well, I mean, why don't you cut your prices in half? I mean, yes, maybe you cut your prices in half, but the availability on the underlying data in the goes across like 2 or 3 others of magnitude. So you're not going to solve that with pricing. The only way to solve it is to give customers the levers so they have the feedback loops and controls over that. And that's what we're doing. In the end, we keep pushing to give customers more levers. And that's -- we actually made an acquisition recently of a company called Timber.io to actually give them more levers to control their data volumes and filters and sample and do everything they need to do so it aligns with business value. At the end of the day, though, we have a very high leverage spend. To caricature -- I think these numbers vary by customers, obviously, but we're about, let's call it, less than 1/10 of our customers' cloud bill and less than 1/50 of their -- the cost of their R&D. And so we have a very high leverage spend, and that's the value we bring to our customers.

That's great. Fantastic. Another topic I wanted to hit sort of partners and go-to-market. You just had some pretty important strategic partnerships. Azure as well, but most recently with GCP. Can you discuss what the partnership is looking to do from a joint product perspective? And on the Azure side, I don't think the product's been officially GA-ed yet. And so when should we expect that? And what sort of -- why has it taken to get to market at -- on this many months after the announcement?

Yes. I mean, the first thing I'd say is all of this is a continuation of what we've done with cloud providers in the past, right? So we -- it's more of an evolution than a revolution. We work with all 3 major providers. We have agreements with all 3 [indiscernible] customers, buy to their cloud providers and use their commitments to these cloud providers to buy Datadog, and we work with the go-to-market teams in all cases as well. So it's an evolution of that. On the Azure one, we're especially excited because there's a direct integration with the Azure console. And the technology is there. I mean, it's in preview right now, so it's not fully GA. There still a few things we need to clear out before it can be GA, but we expect that to happen in the first half of the year. So it's too early to tell. Not all of these are going to work. And we're never going to be done with those. I mean, we're going to keep adding and refining those partnerships and spend more time with the cloud providers because that's what our customers need if they want us to. And so we'll keep working on that.

And you're not going to answer this question directly because most executives wouldn't. But in terms of like the revenue contributions from the 3 other cloud providers, would it fall in line with kind of the market share between the top 3 in terms of contribution to Datadog's revenue? Like if you look at the applications in the cloud that you're monitoring and instrumenting, would market share be a good proxy for the contribution in terms of AWS, Azure and GCP for Datadog?

Well, we haven't shared the exact numbers for that, really, but the -- one way to think about it is we're still more weighted towards what was usual -- historically, our customer base, which all started in AWS and then the other providers came in. Initially, I would say a lot of the users on Azure were mostly purely Microsoft stack and a lot of straight hosting of Microsoft products on Azure and not general-purpose computing, which is what the Azure is becoming today. So overall, over time, we're getting more and more usage across those. And we -- our growth looks a lot like the ongoing market share of these providers.

Great. Well, unfortunately, 30 minutes went like that. Obviously, had more questions, but I think we hit on a number of topics, and that will be the teaser until next year's conference. But Olivier, as always, thank you for joining us at the Morgan Stanley TMT Conference. Best of luck in 2021. And thank you to everyone in the audience who've been joining us. Really appreciate it. Thank you.

Thank you.