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

Thanks, everyone, for joining us. My name is Sterling Auty, I'm the software technology analyst here at JPMorgan. Very happy to have with us Olivier Pomel, who is Co-Founder and CEO of Datadog. Olivier, thanks for joining us.

Thank you for inviting me.

All right. Maybe to get started, for those that are not as familiar with Datadog, can you give us a bit of a brief overview exactly what it is that you offer and where do you fit in the overall software landscape?

Sure. Yes. And by the way, before I start, you might hear some cries and screams because there's a baby in the next room. It's naptime, and there may be some protests. So back to Datadog. So what we do is we do monitoring and more broadly observability for development and operations teams. And we do it more specifically for companies that are in public and private clouds. So the good news here is that just about everybody is busy moving from legacy IT to public and private cloud today. So what we do is we help them do that. We help them turn into digital businesses. And we've been public since September, a bit more than 11,000 customers that range from individual developers to the largest enterprises basically.

So we were fortunate enough to have Bill McDermott on just a little bit ago, and he was talking about just the pace of digital transformations that we're seeing. And with that comes a ton of new application deployments. And I'm kind of curious, where does Datadog fit in terms of capitalizing on that digital transformation.

So when you think of it, there's a -- in terms of what it relates to us, there are 2 things that are happening. So the first one is there's a big move to online for a lot of customer interactions. And that's been particularly true. We can discuss that with COVID more recently. But also more broadly, companies of all sizes are turning to software to operate their business but also to compete and to differentiate. So what results there is a true explosion of complexity really as their infrastructure, their data volumes and their engineering teams all expand. And so what we help those customers do is we help them tame that complexity, and we help them become more agile, more innovative, so they can not only maintain the quality of service and also whatever they're offering to their customers, but also to react quicker and shipping products faster. And one really good example of agility is all of the scaling and the new services that came online with COVID more recently.

To help the investors, again, that are a little bit newer to the story, you've got 3 buckets. You have infrastructure, APM and logging. What is the tip of the spear that gets you into an account. And what are you experiencing today in terms of the uptake across your solution set?

So today, we typically enter through infrastructure monitoring, which is something that every single customer needs -- absolutely needs to have when you move to the cloud. So when you start a cloud environment, you basically -- you open up a new building, and you need lighting on all floors, and that's what -- so you certainly need to have it. So we're typically our customers' first call after the cloud provider who provide infrastructure monitoring. And then we typically are going to land maybe 1 or 2 other products with that. We think we communicated 75% of our new logos land with 2 or more products. And these other 2 products are going to be our APM solution for monitoring the application itself and -- or log management solution to look at a lot of data. So we typically enter with that. We typically enter very small, too. I mean, we start in the cloud window -- we start with our customers when they start on the cloud, which is usually small. And then we're going to grow with them as they deploy more workflows into the cloud.

I think the number one question that I get around infrastructure is, "Wait a minute. So if a company is putting a workload in AWS why do they need Datadog to monitor? Can I just use the inherent tools that AWS has or Azure has?"

So first thing to say, most customers don't see themselves as being users of only one service provider forever, right? As a starting point, they already have an on-prem infrastructure and then they're adding a cloud provider to that. So they all need some level of visibility that is going to cut across all of those various service providers. So that's one of the things we do structurally. Now in addition to that, every single cloud provider provides some level of monitoring data inside of their platforms, but it is typically only a small part of the data is being generated by those platforms. So they typically look at the components to build themselves, but not at anything else. And so what we do is, our job is to integrate all that into one place and to assemble the full picture for our customers in the way that it's completely agnostic to the vendors.

That makes sense. And there's -- the term, I think, is probably overview -- overused is this quote of Gartner saying now only 5% of applications are instrumented for monitoring. On the infrastructure side, how would you look at what portions of kind of the just core workloads that are in the cloud are instrumented or monitored with tools?

That's the thing. And by the way, I think that Gartner quote, maybe it will be on my EBITDA for annually. It will follow me my whole life. But the big difference there is when it comes to infrastructure monitoring, absolutely everything has to be monitored because that's what tells you that the infrastructure is there to start with, which you need to know. So if you look at the old world or the legacy world, there's been infrastructure monitoring ever since there has been infrastructure to monitor. And in the old world, it's the market that's already assigned. Companies are using the IBM products or Tivoli, they're using the old BMC product, what used to be CA, HP Openview, that sort of stuff. And it's super, super sticky. Once you deploy on a -- build so much of it over time, that is really hard to replace, which is why this industry has been really uncontested for a while. What really opens the path to us is the replatforming that is taking place as companies are migrating to the cloud. And that none of those products work anymore in the cloud environment, and everybody adopts a new stack from scratch. So that's where we come in. So 100% of the infrastructure needs to be monitored in the cloud.

But if that's the case, listen, you're growing quickly, and you've got good scale, but looking at the dollars being spent in the cloud across the AWS, Azure, GCP, Oracle Cloud, et cetera, it would seem like if they're paying independent vendors, either yourself or somebody else will be much, much larger. So what are those companies using today to monitor? And how do you kind of motivate them to switch on to the Datadog platform?

So I'd say, we're growing faster than the market, right? We're growing much faster than the overall cloud market. And mostly, what we compete with is a temptation for some companies to do it themselves or to glue together some open source to make it work. That, by the way, is the difference between the accounts in which we land small and the ones in which we're going to land bigger, is that the ones that where we land bigger are the ones that try to do it themselves first. And then after a couple of years of getting to some scale with that realize that it's not going to be a workable long-term solution or it's draining way too much work [indiscernible] resource from them and they switch to a product like ours.

On the infrastructure side, is there any other independent competitor that you watch? You mentioned kind of companies doing it themselves. You've got the inherent tools in the cloud platforms. But are there any other independent vendors that have a cloud infrastructure monitoring tool such as yours?

Well, there are many companies that said they're doing that, they're going to do that. But really, when we sell, it's not typically not a competitive situation. Like it's a greenfield, new opportunities, net new workloads, companies that are starting their migration. So by and large, really what we see is, do it yourself, open source and us. And then they always have some existing technology on-prem, like they all have a BMC somewhere or a Tivoli somewhere. And then they're going to have -- like, they typically have an APM and the log management solution on-prem too that they're going to have a strong follow and they're going to have NetDynamics or Dynatrace for the APM, but that's separate from what they're starting to do in the current environment when they deploy us.

So the one other questions that we got from the audience is just that when you look at the areas where you do compete against a Dynatrace or Splunk or a New Relic even, what are those competitive dynamics like?

So we're a very different breed. So we didn't start Datadog to build a better mousetrap, to be a better infrastructure monitor, to be a better APM or any of that. The reason we started it was to bring together teams and systems that used to be separate. The back story behind the company is that my co-owner and I used to run different teams, I used to run development and he used to run operations and the teams didn't really speak to each other. And so the same point was, hey, let's bring those teams together. So what we ended up building was really a platform whose full purpose was to bring together different data streams and make it as relevant as possible to different teams. So what we got out of that was a product that was meant to be deployed everywhere when we're starting cloud environments because we started -- as I said, we are customers' first phone call when they need to instrument their infrastructure. So we get deployed everywhere. The other thing we've got out of that is that we get used by every single engineer on those teams because we actually built the product to be relevant to as many people as possible. We built it to be as easy to adopt as possible. We do it for the professional services. So when you combine those 2 things, the platform that is fully integrated, that is deployed everywhere that is used by every single engineer every day, it gives us a lot of surface of contact to solve as big a problem as possible in all that. So we started then adding into that platform elements of what used to be different categories before. So infrastructure monitoring, APM, logs, and there's a few others now. But really, we see that as being one big unified category. It didn't make sense to have these separate, it's like selling left shoes and right shoes from separate vendors, right? It doesn't make any sense. So we started doing that. And that's how what makes us different from those other companies.

Let's switch over and talk a little bit about COVID in light of -- there's a perception from investors that, hey, all these hot new tech start-ups are built on Datadog, and that's going to be a significant exposure. And if a lot of them come under pressure from COVID, then what's that going to do to your growth in revenue. So maybe to put it in a simpler term, what is your SMB exposure?

Yes. So when you look at our business, we about have the same amount of business in large enterprise, which is over 5,000 employees. In mid-market, which is 1,000 to 5,000 employees and in SMBs, which is less than 1,000 employees. And then if you further slice in SMB, we have less than 15% of our customers that are on the -- the companies that have less than 100 employees. And then if you even zoom on that, it's actually many, many different companies, it's thousands of companies. And the exposure there, I would say, is there's less of it to COVID because these are not typically companies that are in the travel industry or event industries. It tends to be more software and actually companies that might be more beneficiaries of the crisis. So overall, we think the exposure is low. And another way to look at it is if you forget about the company sizes, and if you just think of the sector of our customers, we have about 10% of our customers -- or I should say, 10% of our revenue that comes from customers that are very negatively impacted by COVID. But it's counterbalanced by about 10% of our revenue that is very positively impacted by it. So we have, on one hand, airlines and hotels that are really suffering right now. But on the other hand, you have the companies that deal with streaming, and video games, and deliveries, and collaboration and all those are scaling up like crazy. So today, those are basically kind of balancing each other.

On the complete opposite end of the spectrum, investors also worry about large complex CIO level deals that might be easy to push down the road and delay. What are you experiencing on that end of the spectrum?

So there, our go-to-market interestingly is not very different, whether you're a small business, mid-market company or a large enterprise. It's largely bottom up. Even in large enterprises, when we're going to start conversation at the CIO or CTO level, the first thing that happens after that is we get kicked down to the teams, and the teams are the ones that adopt us and test the product and deploy. And by the time we come back to the CIOs with a large number or a large quote, 1 of 2 things has happened. Either, we've already deployed and pretty much everywhere we need to be. And we just need to go ahead, and that's a very signification conversion to be had, or in many situation, we're -- they're already paying for us already at a smaller scale, and it's just an expansion that we're considering. So from that perspective, I wouldn't say we're extremely impacted by COVID. We can say right now, in Q2, the pipeline we have is fairly consistent with what we had in previous quarters. So there's no -- nothing drastic. But at the same time, we don't really know what's going to happen. We have less visibility into it. It's new to everyone. We don't know if maybe some of those deals will take longer, there's going to be more scrutiny on some of those things. So compared to previous quarters, we do have less visibility because [ of the situation we see ].

What are the typical sales cycles that you [Audio Gap]

So it's typically a couple of months in the mid-market and below and a couple of quarters in large enterprises. But obviously, these things vary depending on the enterprises. I mean we have customers that are very large enterprises, but very forward thinking that moves very fast. And we have other customers that are also large enterprises that are government entities and need everything to be planned upfront and RFP and everything else. So we have 2 extremes there. But overall, our go-to-market is very rapid. The reason for that is we don't actually sell the cloud transformation we're a critical part of it, but we hop on it as it gets into motion. We typically start showing value for customers when they start feeling some pain in those environments, which means they have something going. So we don't need to be the ones evangelizing for years before deals get going.

That makes sense. Just as a reminder to the audience, if you would like to ask a question, just go ahead and click on the Q&A button at the bottom or on the side and enter your question there and I'll incorporate it into our session. Olivier, given the technology that you have and where you're actually deployed, you have a unique view. I'm kind of curious, in light of everything that's happened with COVID, this massive shift to work from home. We're seeing Internet volumes explode. How are the systems actually holding up out there?

So it's actually an incredible success story when you think of it. We've had a number of customers that had to scale 5x, 10x, 20x almost overnight. And they've been able to do that. When you look at the -- what happened at the end of March and early April, everything that was physical was under extremely -- extreme strain that we couldn't find toilet paper anywhere. But at the same time, everybody's kids started watching Disney+ all at the same time and Disney+ remains up. So it's a great success story for the cloud, it shows that it can scale up. It shows that it can also push the worry of operating data centers to others, and there's real value in that in terms of crisis. But it also shows that you can react quickly like for companies that had to orient their operations, that had to ship new products very quickly, so they could continue serving their customers to ones that were digital-native and that also had large footprints on the clouds capable to adapt very quickly to others struggle to do that.

So what about deployments, though? Obviously, you're 100% cloud, are there any changes in deployment time lines because of COVID?

So we don't know yet. And that's one of the questions. We -- while we have a very good pipeline right now, we don't know if some of the new projects will take longer to get started because of the current situation. We have some anecdotal evidence that companies that are extremely affected by COVID are starting to plan for transition. We had some inbounds from, I'll tell you, amusement parks, for example, that are figuring out what they need to transform and what investments they need to make. So it's a good indication that, at least in the midterm and long term, we think it's going to be an accelerator or an amplifier of the cloud migration and the digital transformation. In the short term, there's more uncertainty because these companies have so many different priorities to juggle as they try to survive. That's basically what can happen.

So we have a couple of questions from the audience that I do want to put in here. The first one is asking what are the benefits to the customer if they purchase infrastructure, APM, logging, basically the entire suite from single vendor?

So I would tell you, we think it's actually -- those 3 products, it's actually one category. It should have been one category all along. Right now, really, people are buying left shoes and right shoes from different vendors. And it doesn't make any sense. So when you think of the use cases, all the use cases actually cut across the different products. Like when you have an issue, it doesn't stop at the [ bannery ] between the application and the infrastructure or it doesn't manifest itself just in logs, but not in the application itself. It actually has to cut across those. The users are the same in the end and the way these are deployed is also the same. I think these things we deploy everywhere the same way. So really, it's one big category.

The next one is kind of a follow-up on the competitive landscape. How would you characterize the competitive landscape today versus 18 months ago? Who's doing a better job competing against you? And maybe where are you picking up incremental share?

So interestingly enough, the competitive landscape hasn't changed all that much. We still see the same dynamics when we're being adopted by customers and when we're scaling with customers. I think the biggest change is that it's become very obvious that having everything as a unified platform makes sense. So we see a lot more talk about it from just about everybody in the industry. I would say, though, that it's very difficult for other players in the industry to transform from whatever they are into a unified platform. They all have initiatives to that effect, but the bulk of their customers, the bulk of their revenue and the bulk of their functionality ease on their old product that is not integrated, that is still mostly on-prem for most vendors, and that's what their customers are using pay for today. So it's very, very hard actually to -- for them to transform.

So you mentioned a lot of times you're getting pulled in once cloud deployment actually happens. But is there a trigger point or a milestone that a customer gets to where the engineering team goes, "Ah. Now we need to have that infrastructure monitoring." And who exactly is the point person that drives the purchasing for it?

Yes. So it's very interesting because we usually come in when there's a little bit of pain. And when companies start moving into the cloud, like they typically have their DevOps teams or sys admin teams set up their first cloud environments. And the moment they start running a few applications, a few real applications in there, they realize they have no idea what's going on. They have no visibility on those. The complexity is completely -- they're drowning in it because they have all these instances and containers coming up and down all the time and they can't keep track of that. So the people who bring us in are typically those sys admins or DevOps teams that are going to feel the pain first. So what they do then is they sign up for Datadog. They deploy it. They set up all of our out-of-the-box integration so that they integrate with all of their off-the-shelf components in their stack. And what happens then is they invite the developers because the reason you move to the cloud is to dis-intermediate IT and to actually have developers, self-serve and build applications themselves directly with infrastructure. So they invite the developers. At that point, we're fully integrated with infrastructure. And then what the developers do is, this time, they're integrating us with their own custom application. They can do it by manually instrumenting the application and putting -- and submitting metrics and meta data and advanced circuit from Datadog or they can set up our APM product which trace automatically and map the application automatically. But basically, the developers are the ones who are going to take the next step. And what's interesting there is that the developers outnumber the DevOps team or sys admin teams by a factor of 10:1 or 20:1. So typically, they are not the ones that bring us in initially. They're not the ones that put the first credit card into the system or sign the first PO, but they end up being the vast majority of our users in there.

That makes a lot of sense. And one of the follow-up question is, okay, what do the price points look like for the various solutions that you have? And what kind of uplift do you get when a customer starts to add the additional model?

So we charge per usage and per volume, basically. So the price we charge is commensurate to the size of our customers' infrastructure, which can vary over time, right, because they are cloud environments. So basically, to simplify, customers are buying a certain number of -- they're paying for a certain number of instance hour in the cloud environment, say, for example. And the way we price is, we do it so that we want to be deployed everywhere and used by everyone. Meaning, the unit price is going to be lower than what you'd expect from one of the legacy vendors. But when you extend it to the full infrastructure, including development and production and across all of the applications and across all of the teams, we typically are going to end up with an equal or larger wallet share than what you'd expect from a legacy vendor. We shared, I think, in the past, the ASPs for the various types of customers we have. But today, our ASP in large enterprises, I think, is north of $200,000 a year. And our ASP in mid-market is around, I think, $160,000, $170,000 a year, and those are growing very quickly. Our customers that are, at some form of scale, in large enterprises, are paying us millions of dollars a year, and these are still growing. So we see, basically, a long runway of growing with our customers as they move more and more of their work closer to the cloud.

Based on the comments that you made around it makes more sense that this is one category, not the 3 buckets. What does that mean in terms of 3 to 5 years down the road? Is there room for multiple monitoring vendors?

So look, we don't think it's necessarily a winner-takes-all category, right, because these are all separate products that separate customers are going to use. We do think it's a very, very, very large market, and we do think that we are very well-positioned to be in the leadership position in that market. And the reason for that is not only the breadth of the offering and our ability to add to it. I mean, we've added those categories I mentioned before, but we also started entering more recently security, for example, which is related, but not the same. But also, the breadth of our customer base. We serve both small businesses and very large enterprises. And what's very interesting there is that the small business aspect of it and even the individuals that use our product for free, these give us a network effect when it comes to integrating with absolutely everything in the world and getting as much data as possible. And keeping the product as simple as possible so that we can be as efficient as possible in the large enterprises as well.

You talked about companies doing it themselves. One of the audience questions is open source is very active around the idea of observability. How often are you competing, or how do you compete against some of the open source projects that are out there, the Jaegers, the ELKs, et cetera?

Yes. So it's a good question. Look, this is the biggest competition, right? This is -- companies might be tempted to build it themselves. For that, open source is only a small part of it. The analogy I would give you there is that, even though it's a troubled industry. So I'm sorry to maybe [ add you to ] them. But the open source is basically the engine. Like they are building engines, they're great at building engines. But what we're building is the airline. And because our customers, what they care about at the end of the day is getting from point A to point B. And the open source bits are only providing a small part of it. So if customers are interested in building all of that, that's fine. Maybe there'll be a great [ caveat ] for open source and look, Elastic is doing well as a company and they largely appeal to the crowd that wants to build it themselves. We think, though, in the grand scheme of things, this will be a niche in that most customers want their problem to be taken care of for them. They just want to go from point A to point B. They don't want to build it. They don't want to own it.

Do you see the COVID environment actually being a catalyst for RUM adoption? And now that you're out there in GA, how easy is it to deploy RUM?

So RUM is very easy to deploy, right, it's a -- you add a few lines to your JavaScript, in your web pages and that's it. You're up and running with RUM. I would say it's still fairly early for that product. So we are happy with the uptake we see. We've barely started charging for it in Q1. So we -- it's not really yet material, but I'm expecting we'll talk more about it in the future. COVID, in general, is an accelerator for everything that relates to monitoring your interaction with your customers. I think it's pretty clear. In the short term, again, it's not completely clear what's going to happen because there's so many different factors that can come in. But if you look at the midterm or long term, it's very clear to us that it's going to accelerate or amplify the trends of digital transformation and cloud migration.

That was going to be my next question, is just are you actually seeing accelerated cloud workload deployment in March and April?

Yes. Look, we had some customers that had to scale up like crazy, right? And so we've seen some of that. I think what it is kind of balanced by is that in the short term, all companies, big and small, are being very careful about cash. In the face of uncertainty, every single company, and you've seen the capital markets, many companies are raising debt and trying to make sure that they're well-equipped should the crisis last and be difficult for them. So we -- again, short term, not completely sure exactly what's going to happen. We see -- the numbers we see in usage, things are looking okay, but we're still very early in that. So we'll see what happens.

With the launch of your security product, where do you see that fitting in that cybersecurity landscape? And are there products that it will directly compete against?

So for security, our approach is it's fairly similar to the rest of the product. The starting point is, it's not, "Hey, there's this great category with this other product out there, let's go compete with them." Because most of the interest in categories and security, there's already a bunch of companies doing that. You don't want to do exactly something they do. What's interesting to us, though, is that the problem we saw with dev team and ops team not talking to each other actually exists with security. It might even be worse. In many companies, the security team, the dev teams and the ops teams are not on speaking terms. And so the thing we can do there is we need to get those teams together, get them to look at the same data, and really help our customers operationalize security. So that's the focus. In addition to that, there is a great market that is still not attributed to security, which is production applications in cloud environments. And if you think of the main -- like the most interesting category integrated to us, which is SIEM, the bulk of it tends to be on-prem corporate IT, and they sort of reassign that. People are using Splunk, they're using Curata, they're using a bunch of other things for that. But when you look at SIEM in production applications in cloud environments, there's mostly nothing, and we see that in our customer base. So the same opportunity we had with infrastructure monitoring, we see it in cloud environments with security, while also playing to the same strength of bringing different teams together. And with the added advantage of already having all of the signals flowing through our system. We know everything about the infrastructure, everything about the application, everything about the end users. We have all the logs. So we have basically all the data we need to pass judgment and operationalize security.

So there's a whole segment of security that's geared towards kind of source code analysis to find security holes. Some, you send the actual source code, some you send the compiled binaries. They run it through analysis and find the issues. Given that you've got so much root cause analysis and instrumentation built in for APM, is that a natural extension?

So our focus initially is more on looking at the run-time data and what's actually happening on the live application and live infrastructure. What are users doing with it? What are they trying to do with it? What traffic are we seeing on the network? What can we tell from those patterns? Do we see signatures of attacks? Do we see -- these are the kind of things that we're focusing on for now. But look, there's many other -- like this -- we're very early in the investment cycle for the security products, and I expect there'll be many more things we do over time.

Yes, that makes sense. There's a follow-up question from the audience on RUM. What would be a hypothetical ASP for an enterprise customer for RUM?

Well, it really depends on it. Because RUM, really depends on the number of users you're going to track and who often use the application. So it can vary widely depending on the use case. But look, I mean, overall ASPs for large enterprises, we start -- we land typically in 5 figures, almost 6 figures a year. And within a couple of years, we're in the millions, and then we keep going there. So we can assume RUM to be a part of that.

It feels like each year you have an area of focus in terms of technology. Last year, for example, it felt like logging was a big area of focus for 2019. What's the area of focus for 2020?

Well, the big new product is security. And that's something that we're investing in, and it's going to take us some time, really, to fully deliver on that. We have some really exciting first results. I think we have some of the customers that took part to the beta of the product that already became paying customers for it, which is very exciting. But it's going to take us more time to make this product broadly relevant to all kinds -- all sorts of enterprises, not just specific kinds of customers. We -- but we also invested in our existing products. So we have -- we're still early in the cycle for infrastructure management product, even though it was our first product. We're still heavily investing in it. In fact, the majority of our resources is going there. We have heavy investments to be made in APM, heavy investments to be made in logs. These are early in their life cycle, and we have a long road map for all of those. So we're investing across the board, which is why we're not slowing down the hiring. Despite the COVID situation, we're hiring as fast as we can both on the engineering side and on the sales side.

All right. Fantastic. With that, Olivier, thank you so much for joining us today. Really appreciate it. Stay safe and stay healthy.

All right. Thank you.