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

Okay. Good afternoon, and thanks for joining us. My name is Tyler Radke. I co-head the U.S. software research sector here at Citi, and welcome to day 1 of our technology conference. We have the CEO of Datadog, Olivier Pomel, if I pronounced that right. And thank you for joining us today.

And I thought we could just start off with a high-level discussion. Just talk about the strong Q2 results that you saw, a big reacceleration in the business. Maybe just at a high level, like was this driven by something that you think is happening in the industry? Or was this Datadog-specific?

Yes. So first, thank you for having me. So there's a couple of things that were tied to what's been really a few very good quarters and especially the reacceleration you noted in the last quarter. The first thing to note there is that we do see overall the cloud migration and the digital transformation picking up where they were before the pandemic, pretty much. And in many ways, the rhythm at which companies are moving more close to the cloud, starting new workers to and expanding their footprint in the cloud is coming back to where it was before the dynamic, and that's definitely one of the big drivers of our business. So that's a driver of growth. And you see it also in the growth of the platform [indiscernible]. Another aspect that is, I would say, more specific to us and what we've been investing in is that we see that some of our new products are coming to -- are getting to critical mass and are contributing significantly to the business while growing extremely fast. So we said on our earnings call that our APM suite combined with our log management product add up to more than $400 million in ARR and are also seeing hyper growth mode. So just about doubling on more. So this is what gets us really to solve a bigger and bigger problem for our customers and increase the wallet share we get, and that explains why we're growing significantly faster than the cloud providers, which those numbers. Overall, I would say, we're still largely capacity-constrained, not demand-constrained. So what our focus is, is on really increasing the size of the engineering teams and the go-to-market teams so we can build more products, solve a bigger problem for our customers and also beyond more conversations that more customers basically serve all of the different geographies and customer segments.

Got you. Got you. And I'm curious from a customer demand perspective, I think a lot of investors were -- remember the second quarter last year, where you did see some kind of unusual customer behavior just around rationalization of spending. But how would you say that the customer demand has evolved since then? I mean, to your point, we have seen kind of some reacceleration in both AWS and the Azure numbers. What are you seeing out there from just a customer spending perspective?

Yes. So what we saw in that Q2 last year, so that's more than a year ago now. What we saw was that customers that were at scale in clouds, that consumed a lot of clouds, tried to serve as much as they could on what they hadn't spent yet, which is our next cloud deal basically. And so they try to save as much as they could on their AWS bill, on their Azure deal. And then that flowed down to us because the way our customers increase their consumption of Datadog is by increasing the consumption of cloud services. So we saw that happen. More recently though, we've seen that trend return. So we've seen customers revert back to growing in the cloud, moving more workloads. Throughout the pandemic, actually, we saw a lot of new customers that started their cloud migration. So we're seeing hotel chains and amusement parks and airlines, companies that were really under serious stress during the pandemic that felt the urgency of starting their digital transformation and cloud migration. So we saw them come in. They didn't have any volume though because by definition, when you start in cloud environment, you start small, like you have to start the workloads and then you have to grow them over time. What we see today after more than a year of that is that some of those customers that started their transition during the pandemic are reaching some scale. And these are starting to drive some real numbers. So overall, the demand situation is very similar to what we saw before the pandemic with the added benefit of having a lot more companies having started their transition during the pandemic.

I see. Okay. So Oli, one question that we often get asked, I would say, for Datadog, Elastic Dynatrace, Splunk, right, there's a number of players in the space around observability. And so I guess just competitively, what is Datadog's biggest source of differentiation versus kind of the traditional observability players as well as the hyperscale cloud vendors that have their own monitoring solutions?

Yes. So the first thing I've said to start is that I always talk about competition a lot in investor conversations and almost never in between. And for us, the situation is the same as it's always been. Like it's a space where there's a ton of opportunity, and there's a lot of players at different levels. But when it comes to what defines our success, what drives our success, it has very little to do with what happens competitively. In fact, we don't even have a competition slide in our Board decks. This is not what we think about. The drivers of success is how much of the product can we solve for our customers and how much can we build capacity to get in front of them. Competition comes second to that. In terms of the differentiation, what we do different is that we get deployed everywhere at our customers. Like we really cover their environment end-to-end. We get used by every single engineer there, which is also [indiscernible] not what you find with the typical observability players. So we get maximum of [indiscernible] contact with those customers, which lets us over time, solve a bigger and bigger slice of the problem and external product footprint, which we've proven through the growth of our new products and where we've been able to attach those to add on initially. So what we start to do in the end is solve the end-to-end problem and connect everything together for our customers, which is not what the rest of the industry has been able to do, hasn't been the case of the point solutions that were there before in the legacy world, and it hasn't so -- it hasn't been the case and like it won't be the case of the cloud providers and [indiscernible] more deliver of the building blocks.

Got you. Got you. Okay. And I guess just in terms of the market, right, I mean I think it's wise not to overly focus on competition because clearly, the execution has been really strong. But just how do you think this plays out? I mean, on one hand, it is a very large market opportunity. I mean most -- tens of billions of dollars, clearly, is kind of what we've seen from a TAM perspective. But how do you think this plays out with the number of vendors out there? Do you think there will be consolidation? Do you hear that from a customer perspective that there's a desire to consolidate the number of observability tools that they have under management?

Yes. So from a customer perspective, what's very clear is customers don't want to be integrators. They want -- they don't want to be -- to manage like 127 different tools that have to connect to each other and that have to be -- so that's definitely not something they want. And we are the beneficiaries of that because we do see that as our customers use more mobile products, they want to rationalize and basically fold more of what they were using before into what we can do for them. That's part of the vision we sell to them, that's part of what they're already buying to. So we definitely see that playing out in the market from a customer perspective. From a competitive perspective, we do think that it's a very large opportunity. There's room for many companies being successful, but there's also room for, I would say, Salesforce-style winner for the broader category, and that's what we're after. And we think that getting there depends on us executing on our road map to basically deliver as much product as possible, solve as much of the problem space as possible for our customers and make sure that we scale the organization in time to get in front of them.

Got you. Got you. I wanted to ask you about how you think about pricing longer time in the market. And I think one thing that has been pretty apparent from the beginning of Datadog is you have been -- had a very transparent pricing model that I think a lot of other industry players have kind of had to respond to. At the same time, I think pricing, to some extent, is tied towards a measure of infrastructure consumption, which clearly with the pricing declines that we see at -- for kind of core infrastructure services at AWS and Azure over the years, I think some investors do have a concern that maybe there's pricing pressure longer term in this market. I'd be curious how you think about that and how you kind of are thinking about going to customers and proving that value longer term.

Yes. So there's a few things. So in terms of pricing philosophy, first of all, so we want to be -- we're not here to do price disruption. We lack the cycle where we charge for products and customers saves a lot of money and they have to demand value. And in return, we have to deliver the value and they give us a sort of clarity on that, and it's a virtuous cycle because it gets us to develop better and better products for them. It's better than the opposite, which is trying to figure out what you can take out of your products so they're even cheaper. So we lack that. Philosophically also, we think it's very important for our pricing to put the customer in charge and give them -- like put them in control of and give them the tools they need to align what they pay with the value they get. And that's why you see we have a number of different SKUs that we actually unbundle the product as much as possible, so customers can pick and choose for which part of their data and which part of their infrastructure they use, which part of our products and make sure that they get the full value for that. Now in terms of the broader question for what happen to prices over time, what's pretty clear is that the footprint and the impact of applications and infrastructure is only growing over time. It's becoming a bigger and bigger part of every single company. I mean obvious that for cloud-native companies and digital businesses, it's a big part of it, but for every single other business becoming additional business over time. So it's becoming a bigger and bigger part of what they do. At the same time, data volumes can explode in a way that way outstrips the top line growth of any of our customers, right? So what we believe in is that we'll be getting a larger and larger wallet share from our customers, while at the same time, the pricing of the data point of the gigabytes that we get are going to decrease over time. So one way to think about it is maybe in 5 years we charge 10x less per gigabyte, but customers will send us 100x more, and that's the way it would play out. That's pretty much the way it's played out in software in general, too. That's what we're seeing.

Great. Great. So going to kind of talking about the products a little bit deeper. So obviously, observability is a big buzzword and encompasses a big part of what Datadog does. I thought it would be helpful just to kind of frame for investors where your various product SKUs across monitoring, APM, logs, kind of where they are in terms of the state of maturity today and how they fit together. And specifically, just trying to understand if you see customers still kind of buying point products in this market if they're looking to you to kind of deliver that full end-to-end suite.

Yes. The very quick overview is that the bulk of our -- what drives the revenue today is observability. And there, we started with infrastructure monitoring, which was our first product, obviously still -- we still think it's early in its life cycle, still kind of going heavy investment and growing very quickly and being what most of our customers land with when they start using Datadog, which is at the center of what they do. It's a great basis to start because by definition, it has to connect to everything our customers use. It's there to tell you it's working well, so it needs to see it. On top of that, we lay our APM suite and our log management product. So these are the ones that, I said earlier, over [ $4 ] million in ARR and growing very quickly. So these products are reaching maturity now. They're, I would say, at initial scale but growing very quickly as we still have a very large opportunity for them. They're still way underpenetrated even among our existing infrastructure customers. So they have still a long road map for that. We have a number of smaller products. We have a release that fits in between APM, logs and infrastructure, I would say that rounds up the observability experience, so things around UX monitoring, [indiscernible] monitoring, for example, synthetic testing, things around network performance monitoring, other new product, we just announced them and database monitoring. So those tend to be maybe smaller SKUs, but it really round up the platform and that in aggregate actually drive quite a bit of growth for us. And then if we set aside observability, we started also entering another very large category we see as being a potential TAM multiplier, which is security, in particular, cloud security. And what we target with that is security for applications in the cloud as opposed to corporate IT. And we will have to bring there, and we can talk about the product some more if you're interested, but we'll have to bring -- to bring there is the fact that we bring together DevOps and security in a way that we think is a [ missile ] for the future. So security is still small for us. We think it's going to be a [ long ] deal, but it's part of our story moving forward. So when you add up all of these, we basically have a whole collection of products that range from very mature and drive the bulk of our revenue while growing very fast to a number of products that are -- that don't really have a ton impact on a top line yet, but are growing extremely fast in our [indiscernible] some form of initial cadence.

Right. Right. Okay. Yes. I would love to touch a little bit more on security. I think you recently acquired a company called Sqreen that was kind of focused on application security. But just given how -- obviously, how important that market is, but it is a market that's hard to get right, and I'm sure you've been super thoughtful about investing there. I mean, how should we think about the ability for, on one hand, what is a huge market, but on the other, it is kind of a newer market for you? Like how should we think about the ability for security to scale in terms of ramping into a meaningful portion of revenue.

Yes. So as I was saying, I think it's going to be a long build for us because there's a lot to do. It's also a market that I would say is more nascent than the cloud observability market. If I were to drive parallels, the market for cloud security bringing together developers, operations and security folks today is very similar to the market for observability bringing together developers and operations 10 years ago, which is when we started that. We see all the same trends. We see those teams fighting instead of working well together. We see the greenfield and the multiplicity of point solutions and open source projects and things that try to solve only a part of the problem for part of the team. So we see all that. The situation is very similar. With specific security is that there's a lot of functionality to build for these products to be minimally useful and broadly applicable across different customers. And in the case of security, there's also a number of point solutions that we compete with as we start selling into these markets specifically.

Got you. Got you. Okay. Curious, one theme in the industry is just around the idea of AIOps or AI-driven operations. I guess what does this mean for Datadog? And what are some of the things you're doing from a product investment perspective to put more intelligence, embedded intelligence of your...

Yes. So the way we see AIOps is we don't see it as a category so much as we see the capability of the platform as a whole. So we don't actually have a SKU or product that we call AIOps. We also are very wary of leading with AI and disappointing, which has been the MO for every single software company in the history of the world where you promise something fantastic with AI and then everybody realizes that with [indiscernible] does best is being a kitchen timer, which is at least how I'm using it. So we -- the way we do it is -- we actually [indiscernible] the system in a way that is publicly multi-tenant and completely SaaS-based in such a way that gives us anonymized access to all of our customers' data for infrastructure, application and things like that. We use that to train algorithms, train our models. And we're in a position since we don't sell SKU specifically that says we're [indiscernible] everything was AI, we're in a position to actually pick and choose the use cases for which all data has been too low us to give excellent high accuracy insights and answers. And over time, basically, we keep delivering more and more of those to our customers. So that's something that we're investing early on, and the product that we -- the name we put on that was Watchdog. That's what we call that feature in our product. And this is something that permeates our product and something that we see also investing in for the years to come.

Right, right. Yes. I guess related to that, I mean, I think the pie in the sky dream of observability is to be able to automatically detect and kind of instantly remediate outages. I guess, from your perspective and what seems to be taking like a very practical approach to this market, I mean, how far away are we from that vision? I mean do you think it's ever going to be achievable? I would just be curious to kind of get your perspective on it.

Well, it's a race against complexity. I think that as we keep adding more smarts and automating more, developers keep building faster and with more abstraction layers and more modules and more things that drive the complexity of basically that make the problem harder to solve. And so -- we don't think that we'll catch up anytime soon because that complexity train is running so fast and is only accelerating at this point. But what it means is -- If you don't actually rely on some of that insight and automation and some of those smarts, you're going to be enabled as a human to keep up with all that complexity and so [indiscernible] that's what owes you to keep driving faster and doing more. So we think that risk will keep on going. We think that's why we'll have to keep on investing for a very long time and building those marks and reducing the complexity for our business.

Right. Right. So kind of back to the discussion earlier around the TAM, I mean, I think one thing that is unique about Datadog is the range of companies you sell to from Fortune 500 to kind of digital native start-ups. I guess how do you think longer term about the potential size of your customer base? I mean, I think today, you're around 30,000 logos. I mean, do you see potential for that to be 100,000 logos, 500,000 logos? I mean, what's kind of the upper limit in terms of the type of customer that you could serve?

Yes. So in unique -- customers, we have in the mid-teens, I think, and we have the exact number in the last communication -- tens of thousands. And we -- the way we see it is -- so right now, we have a very broad user base because we have a number of communities using us and not paying, a number of individuals, like we have a free tier also. I would say even the lower end of our paying customers doesn't pay us a lot and represents a very small fraction of the revenue. The reason why we're serving that broad customer base today is not -- for the revenue rates, for the work in [indiscernible] gives us when it comes to integrating with everything under the sun and keeping the product very simple. And that's why organizationally we'll invest in that and we have those customers. Moving forward, though, as we keep adding more products, the lower end of our customers becomes more and more investable and can actually become a very significant growth driver. So that's something we keep in mind for the future. In the short term, though, our priority is still investing as much as we can in the mid-market and the enterprise to make sure that we fully solve that problem, and we were into every single conversation in that part of -- in those segments basically. But yes, as we grow, look, we think we're nowhere near done in acquiring the customers we need to acquire. We definitely think we will be in the hundreds of thousands, if not more, of customers.

Okay. Okay. And in terms of the focus of the -- it sounds like mid-market and enterprises where you're deploying them as resources. Would you kind of size that as maybe the global 15,000 or 10,000? Just kind of curious where you see the biggest opportunity?

Yes. So for us, the way we [indiscernible] this mid-market is above 1,000 employees, and enterprise is above 5,000 employees, right? And for now, this is the part of our business that's the most investable because you're -- order the ASPs and the revenues we generate from that. But as I said earlier, as we keep adding more products and as the ASPs go up across the segments, it becomes easier and easier to invest into the lower segments from a go-to-market perspective.

Right. Right. And I guess what are you seeing on -- the enterprise side, I think has been certainly slower to move to the public cloud. And I think historically, Datadog kind of captured some of the momentum at the mid-market and maybe commercial segment in the market. But what are you seeing among enterprise customers do you find that -- is that kind of what's been driving this inflectioning in growth that we've seen? Or just help us understand how that -- how those conversations with the enterprise customers may have changed.

Yes. So the bulk of the opportunities we see today are enterprise or mid-market or -- because there's many more traditional enterprises than there are brand new cloud native companies. And so that's where we see the bulk of the growth and the bulk of the activity, or I would say the bulk of the new demand, though because all of our segments are growing fast, like the mix actually remains pretty similar for that ad. We see great growth in SMB, we have growth in market, we have growth in enterprise. So even though we're scaling the teams as fast as we can and the productivity is great on the enterprise, the mix remains more or less similar for us. I should say that the conversion with the enterprise are going well. And initially, when we started the company, none of the enterprises were in a public cloud. So the only companies we're selling to are basically the cloud natives. Then the mid-market came in, then the enterprises started coming in. Today, it's very clear for every single enterprise that they need to transform majorly and they need to be in the clouds. [indiscernible] being in the clouds goes through being the public customers. And we see that conversion repeat again and again and again and again. I think the main difference between enterprise is that not everybody is as far along in that transformation. So when companies are new, when enterprise are new in there, they're going to expand a lot of effort, but they're not going to have a lot of workloads. As they get further into it, the workloads flow, which is when they start becoming significant customers for the cloud providers and [indiscernible] response.

Right. Right. So on the enterprise side, and I guess this is a question more broadly, but how should investors think about kind of the mix of Datadog workloads, if you will, that are on-premise versus the public cloud? I mean I think Datadog kind of gets the reputation for being mainly exposed to that public cloud side. But obviously, for the enterprise, it's a much more kind of complicated hybrid on-prem world. So how should we think about kind of your relative exposure to on-prem, hybrid or public cloud?

Yes, it's interesting because for the first few years as a company, like I would say, first 5 years of the company, we were about 50-50 between public clouds and cloud workloads but running on private infrastructure. I would say more recently, we've seen the public clouds grow faster, mostly because -- many of the TAMs have private cloud and private clouds didn't work so well. And so you saw a lot of enterprises refocus on public clouds instead. I think we might see a little bit of a buckling and [indiscernible] now where we actually see also some private workloads come back. We also see something that is new for us, and we've seen that in the past in the last year or 2 where customers start asking us to monitor their legacy on-premise infrastructure, which is not cloudy. Like when I talked about private infrastructure before it was cloud development, cloud deployments, but on private infrastructure. In this case, we're talking about legacy infrastructure and on-prem, which is something that is not a major part of our business today but might be in a few years as we get far along in the cloud transformation that a number of enterprises are getting to the point where they want to unify everything under one. Again, not something that's happening much today, but we start seeing the first signs of it.

Yes. It's interesting. So I guess just a follow-up on that. I mean, what do you think is driving the need for more either private cloud monitoring or even legacy monitoring? Is it tied to some of these broader digitization initiatives just around putting more instruments on applications, maybe even like homegrown legacy application? Just help us understand kind of that catalyst to do it.

I think it's a question of rationalization and, I would say, internal user experience. When those customers have started their cloud migration like 3, 4, 5 years ago are getting to the point where they have 30%, 40%, 50% of their workloads in the cloud. They have 80% or 90% of their mind share in the clouds. And so the question is, okay, so how do I avoid completely [indiscernible] switching to something else, just so I can keep an eye on what was going on before? Or how can I minimize the chance of failure as I migrate or rebuild this application from legacy into a new environment? So that's when there's interest in actually cutting across and bringing everything under one roof. So that's not something we trigger like we actually don't go up to our customers and say, "Hey, why don't you bring all that stuff in." [indiscernible] conversation we're having with them as they try to simplify their lives and bring everything under one.

Right, right. In terms of one question I wanted to ask was just around the public sector. And I think, obviously, a lot of focus from investors this time of year with the federal fiscal year-end. But just kind of talk to us where Datadog is here. I think we're still kind of awaiting FedRAMP certification. But maybe just help us understand kind of the opportunity and kind of progress you're seeing in that market.

Yes. So right now, we're still just building for this. So we have -- we're waiting for FedRAMP. The good news is there's nothing for us to do. The bad news, there's nothing for us to do yet there, so we're waiting for it to come through. The -- on the go-to-market side, we've been building the team, and it's brand new for us. So we don't know yet exactly what to expect from it. But for us, it's upside when you look at our forecast and everything else. But when we look at some of our peers in the market, for some of them, it's a very big part of their business. So we think we can make a very successful business there, too. So we're building towards that, but it's still early.

Right. Right. And we've got a couple of investor questions come in. The first one was just on your sales buildup. And I know you've talked for several quarters now about being kind of capacity-constrained versus demand-constrained. But just curious, as you've kind of made the sales investments, I guess, number one, how are you expecting them to kind of ramp? And the follow-up question that was asked also is like how to think about that combined with kind of reopening of economy and the intersection of those 2 things being a positive tailwind for the business.

Yes. So I mean, obviously, we like the trends of going back to where the digital transformation and cloud migration worked before the pandemic. I'd say today, from a go-to-market perspective, we're still -- there's still a few things we're doing very well before the pandemic that we're not doing. So getting in front of customers at events, for example, in-person events is not something that we're doing at this point. We're doing different things online, but still there are more avenues for us to generate growth. So we expect that when that comes back, this can be a positive to us. In terms of getting the teams, I mean, look, we've been hiring quite at a fast clip basically during the pandemic. We haven't slowed down at all, which I think serves us well today because we've been able to bring a lot of capacity online. And we're very pleased with the productivity of the teams. I mean if you look at the [indiscernible] payback, for example, on our numbers, it's extremely low right now. It's actually a little bit too low. We don't want to be prisoners to these very, very high levels of productivity as we need to -- as a focus to scale the [indiscernible].

Got you. Got you. Okay. Another question that came in, I think, kind of back to the earlier conversation around the enterprise and what they're doing in terms of legacy infrastructure monitoring. Who are you winning against and/or replacing and monitoring in on-prem legacy infrastructure?

I mean legacy can be anything. It can be -- sometimes the customers are going to retire their old APMs. In many situations, they're going to have infrastructure monitors for their legacy environments, like the VMCs of the world and things like that. So in many cases, can be open source, it can be a really a collection of anything. And it doesn't really matter to us so much because in those situations, what determines the -- whether we're going to go after those specific workloads isn't so much what we're replacing, it's what pieces of technology we need to integrate with to get that data. And are these things that we want to integrate with right now? Or do we want to wait until the end [indiscernible] this effort?

Right. Right. Okay. Makes sense. One question just on the go-to-market. I think you had some changes in terms of go-to-market leadership earlier this year, and a new COO, Adam Blitzer, joined a few months ago. Do you feel like the go-to-market management team has kind of settled? Any potential changes in terms of structure or focus that you anticipate throughout the rest of the year?

Well, the main change is we're growing. So we're very happy with the way the team is working today. I touched on that a bit earlier, but productivity is extremely high. We've been growing the team quite a bit, but what we need to do is we need to grow it a lot more. And we need to develop new segments as we talk about government, for example. So there's plenty of work to start building new things, but the things we have today that are working are working very, very well.

Okay. Okay. And then a question came in just around the Microsoft partnership. So you announced a strategic partnership last -- about a year ago. I think Datadog being able to be -- kind of access directly natively through Azure. Just an update here, you're seeing kind of benefits of this partnership come through?

We're seeing some benefits. Not everything is there yet. We're still -- there's still some parts of it that are not online. The technical parts of it online are showing great promise to us. So it's a very good way for customers to start using our product without [indiscernible]. It's also a very good validation for our Microsoft-centric customers that we are first class citizen in the Azure ecosystem, and it allows them to invest in [indiscernible]. So that's very right.

Right. And then it's kind of a financial question. I promise it's not too deep in the weeds, but just as you think about the drivers of growth going forward, I think you've largely had a pretty consistent balance between new customers versus expansions. I think this past quarter, if I'm not mistaken, it ticked up a little bit more in terms of percentage on the expansion side. Like how do you think about that mix going forward? Do you think it should stay relatively stable or perhaps more on the expansion side?

Well, I think we're still very early so there's many more customers we need to get. On the other hand, the part of our business that's -- that can drive the more quarter-to-quarter is how quickly our existing customers are expanding because they obviously drive more volume than the brand new customers. So it's harder to predict on a quarter-to-quarter basis basically.

Right. Right. And I guess going back to your question -- or your comment earlier around just being kind of capacity constrained. I mean, how would you just kind of characterize the hiring environment and kind of what are you doing competitively to make sure that you're staying ahead of what seems like a pretty challenging hiring environment across the board?

Yes. The job market I think for both go-to-market and engineering is as tight as I ever seen it. So it's definitely a lot to work with. I would say it doesn't change our outlook. It doesn't change our plans or anything else. What it does change though is the amount of effort we have to expend to grow and hire. And so we do while investing in that, and we're growing our recruiting teams -- and we're adapting to basically what candidates are looking for in positions.

Yes. Great. So I know we have about a minute left, but I just kind of wanted to, open-ended, if there was anything you wanted to cover that we didn't hit and maybe just leave us with kind of what you're focused on for the next few months.

I mean, for us, the focus is on scaling the product team and scaling the go-to-market team. There's -- the equation is simple for us. We're early in a very, very big market. We need to be in all the conversations, which means we need excess capacity and marketing capacity. And then we also need to make sure that we solve a bigger and bigger problem for our customers, which means investments in the product team. So if we can get those 2 things right, we should be fine. So that's what I spend my time on.

Okay. Well, with that, thank you so much, Olivier, for joining us today, and thanks, everyone, for the great questions and for participating in this session.

All right. Thank you very much.