DigitalOcean Holdings, Inc. (DOCN) Earnings Call Transcript & Summary

Good morning. Thank you for joining us on day 3 of the Goldman Sachs Communacopia and Technology Conference. I'm Gabriela Borges. I cover the emerging software vertical here at GS. And I'm delighted to have on stage with me Yancey Spruill, CEO of DigitalOcean; and Matt Steinfort, CFO. Thank you for joining us.

Thank you for having us. Good morning.

Yancey, I want to start off with the news from the last 2 weeks on the leadership transition at DigitalOcean. We've spoken about how there's never really a right time to hand over the reins. Would love to get your perspective on how you're thinking about the timing of the transition? And why now is the right time for a fresh perspective at DigitalOcean?

Yes. I think as I reflect on where we've come from, where we are and where we're going, there's never a good time in a company until they get to $2 billion. You could look at a chart and say, well, you came at [ $200 ] and 10 years later, it was $2 billion. What was the right time in there. And -- so let's just say that in terms of timing. We transformed the company. When I reflect on where we were 4 years ago in terms of -- more than triple the company in terms of revenue, every venture has been transformed. 50% of every incremental revenue dollar has gone to incremental free cash flow. So free cash flow margin has gone from negative 25% to mid-20s-percent. EBITDA margins in high teens to the low 40. Employee attrition down, engagements up. We've added -- we've gone from a pure Infrastructure-as-a-service company to having Platform-as-a-service, software capabilities, managed hosting like Kubernetes and our databases, Kubernetes and now an AI platform as well as managed hosting. So we have a more complete platform. And we have a great balance sheet and a good team. And so I think when we look at the next several years, which this decision was all about was the future, this is a good time from a fresh perspective for me to get some rest. And I think for the team to move forward for the next push. Because I think in 4 years from now, someone will be talking about different metrics that are also up into the right. And the job from all of us is to give the keys to somebody in a lot better shape than you took it. And I think that's the case here.

You referenced transformation a number of times. Certainly, the company has hit some tremendous milestones during your tenure. What are the 1 or 2 big strategic questions that you have been spending time on over the last year and that you think will be important for the next year to some time?

Yes. I think the big -- our differentiation is pretty clear, simplicity, our support model, our documentation, the fact of how we help customers remove barriers to them getting their ideas on the Internet and starting a business is real. I think the question now becomes, as we serve ever larger customers who have growing needs, how do we maintain the simplicity of the experience, while also meeting their growing needs. And I think that is the central question for us because as long as we can do that, this is a massive market, $100 billion annual spend, very fragmented. We have a good position in that, but we will need to keep maintaining that differentiation of simplicity and the experience. And I think as we do that, there's just an limitless sort of growth opportunity.

I'll touch on this idea of simplicity of experience in the context of some of the more sophisticated use cases that customers are asking you to solve? What is the 1 or 2 biggest technical challenges that you think customers are looking to DigitalOcean to calls in? And how is that shaping the way that your R&D team is spending the time?

Yes. So it's amazing to talk to our customers, which I encourage all of you to do because it's -- you'll hear it live from them. You don't have to hear it from us. And I was having dinner last night with a customer here, rapidly growing business in the marketing automation world. And the amount of effort that we removed from their system with our serverless products, that allows their team to focus on innovation velocity for their customers is massive. And I was joking with the sales guy last night, I was like, wow, we're really underpriced because they're saving a lot of money, real dollars that they don't have to spend building and maintaining an application that's fueling their business we call our platform. And so I think that's what they want from us. They want -- and I think a lot of times people conflate the word simplicity with sort of low functionality below a threshold in terms of quality, it's not. We're frequently ranked as the fourth cloud behind the 3 hyperscalers. So the capability is substantial. But what they're looking for is, I wouldn't use the word sophistication, they're just looking for us to take burden off their plate so they can have more development of velocity, more velocity to focus on their customers and serving customers and growing their business. And that's what's one of the most unique things about DigitalOcean as you talk to your customers and in 5 minutes, they say back to you what your mission is, which is to simplify the cloud so we can focus on building software that changes the world. And that's what we do. And you hear it if -- I heard it last night over pizza, 5 minutes from here, last night, and it's amazing that, that's what we do. And that is a super power as it relates to the competitive landscape out there.

I want to bring it back to this idea of R&D priorities, which is, if I think about what DigitalOcean has excelled at, that simplicity, taking the more complex functionality that 1 of the big 3 hyperscalers would typically provide and making it accessible. How do you balance the nuance between being a fast follower and also being innovative so that -- especially with what's happening with artificial intelligence, you're capturing more sophisticated cases as well. How does that push and pull work within your R&D?

Yes. I think it's important to understand what we -- we're not a pure technology company, we're a services company. And so the innovation for us is simple -- the experience that we create on the platform by making a lot of complexity, which the cloud is extremely complex. We talk about it, it's a buzzword. It's a very complex interweave of systems and processes that allows people to share in a completely different way than has ever occurred before. And so I think for us, the priorities are going to come down to enabling small businesses to get big -- to become big businesses in a nutshell. We've talked about it in terms of storage. We talked about it in terms of security. But if you sum it down to at a much higher altitude, what does that really mean? It means as companies go from a 2-person team to a 10-person, 100-person, 1,000-person team as their customers go from 1 to 5,000, their needs evolve. And what they want to run it on DigitalOcean, but they want the consistency of the experience and the barriers to remove at each level of scale. So the priorities are really around SMB enablement, I would say. And storage is within that, our security tools, some of the other aspects of integrating the functionality. So if you want to manage hosting or you want to do it yourself that experience, we're integrating that. So a lot of it is enabling small businesses. I'd say the second big one is obviously AI. We made a big investment in a platform and a cloud infrastructure platform that enables people to create, whether it's language models or inference models or run a business leveraging AI, that's called paper space. And so that's obviously a big priority for us as well.

I'll stay on Paperspace, which is -- and Matt, this is the strong and I think you as well. What are the guidelines for the framework that you're thinking about when deciding how much to invest behind the Paperspace?

Yes. So we think, as Yancey said, it's a tremendous growth opportunity, and we were able to get a small company that had market traction and some real good customer base. But it was underfunded and hadn't been taking advantage of the opportunity, which is a great way to buy a company. And so we're pretty excited about that. We've already seen more demand at that business than we had expected in our deal model. We already doubled the amount of capacity that we were planning to invest in 2023. We said it's not going to be a material impact on revenue this year. It will be kind of sub-$5 million, but we think it could contribute up to 3% of growth next year. So I'd say we're being cautious and optimistic about the investment. So we will put additional capital to work if we see the growth, but we have clear milestones we need to see in terms of customer traction and the build of the funnel before we would do that. I don't think you'll see us going out there like some of the other competitors that are in the pure kind of bare metal market. The size of those comments, I think will be more moderate than that, but we certainly see a big opportunity.

You said up to 3%. I think it's at least 3%, is what we said, and it's probably more than that. The only thing I would add is, I think the part of that caution is, as we're learning and unpacking double-clicking, we've supported the inference market. There have been businesses who run on AI inference for years on our platform. I think the new news -- and we didn't have a -- the Paperspace gives us a platform to enable people to create models more effectively with a higher capacity GPU. I think what we're really trying to discern is, what is the sort of run rate opportunity or long-term opportunity for language model building versus the run rate opportunity for the inference and for what do you do with that? Once you have a model, you have to run it on an infrastructure platform. And I think that's a nuance the market doesn't yet know and we're still figuring it out. But we're going to invest aggressively, as Matt said, but I think the question is, how do those 2 align in terms of what our longer-term vision is to -- so I think we'll be cautious in terms of commenting on longer term until that is sorted out.

That's a really interesting observation on the nuance between language model training and the inference use cases that are already being done on the DigitalOcean's platform. As you did your due diligence on Paperspace and as you talk to your customers today, would love if you could educate us a little bit on the differences between those 2 use cases. And you mentioned it's still early days, what are some of the factors that will influence whether large language model building becomes a bigger opportunity for you versus the inference type.

Well, I think the language model is sort of -- you have to build a model, and these are large complex data sets that required a lot of compute intensity over a very short period of time. And then the question is, what's the maintenance run rate? And where does that actually run? It doesn't probably need to run on H100, right? There's a mad dash for the -- and so we're trying to figure that out. But there's endless amounts of use cases. It's an open netted growth opportunity because these models will transform sort of almost every discrete aspect of work, life, et cetera. But is that a big onetime event. And then what's the -- and then there will be -- there's a maintenance, you have to tune it, et cetera, on an ongoing basis, which plays into our infrastructure platform. And as does inference, which is just running a model on inference needs core compute and et cetera. And so I think the diligence proved that the capabilities on their platform can support a lot of the applications, use cases. And that's why as we've been putting more wood behind the fire in the last 2 months since we've owned it, as Matt mentioned, the pipeline has more than doubled in terms of the revenue outlook from when we first bought it. So -- but that's what our diligence is focused on -- can the scale -- it's a lot of small companies have interesting ideas, but you put 10x on it and they break down. And fortunately, that is not what we saw here, which is what we got incredibly excited about.

I think it leads nicely to the bigger question on the business model between balancing growth and profitability. And there's been so much progress on free cash flow margin, EBIT margin for DigitalOcean over the last 3 years. How do you think about the risk that perhaps the balance could have been more towards growth given the opportunity that you see ahead of you? And is there a scenario where you think that balance shifts more significantly towards growth and away from margins over the next 4 years? And maybe that's a question for that as well given that you're staying on it.

I'd say that over the long term, one of the benefits of the work that Yancey has done over the years is we have a tremendous free cash flow margin. So we have -- we did 27% in the second quarter in the core DO business, and we headed towards 30%. That affords us the ability to invest without materially backing off the free cash flow margin. And so while we said Paperspace is going to cost us, call it, 300 basis points, we were still able to maintain our guide for the full year at 21% to 22%. And so we like to think that we can manage through that balance, still generating very strong free cash flow margins while still investing in growth. And when you see big opportunities like we did with Paperspace, and you have the ability to accelerate dramatically accelerate the growth so that we can invest a little bit behind that to take advantage of it without sacrificing kind of the principle of -- we need to be running a healthy business that generates material cash flows in any market environment, which is really the thesis that we've been operating under.

Matt, as you spend time with the P&L and as you've gotten familiar with the business model, how do you think about the amount of progress in heavy lifting has already been done on the cost structure versus what's still to come. What are your observations on? Are there still meaningful categories where you can move the needle in terms of margin improvement?

Absolutely. And one of the things that most excited me about joining Yancey and the company is it feels like there's a tremendous amount of opportunity to drive continued operating leverage. And so there's a couple of areas where we would see potential. One is, we announced in February of this year that we were reducing some headcount. But at the same time, and probably more importantly, that we were shifting some of our resources to global capability centers in Pakistan and India and Mexico. We have the ability to continue to lean into that model, and that's a long-term structural improvement in the cost structure. That's not a onetime, we're cutting costs, trying to accelerate EBITDA. That's a change in the way that we operate. And there's still a lot of room for us to shift in that direction. The other area where I'm in particular, very excited coming from the telecom world is, there's a lot of opportunity on the gross margin side. The data center infrastructure, the way we buy bandwidth, the way we peer, this is a company that was built -- it's more of a cloud, it's more of a software orientation. And there's a lot of scale benefits that I think that we can get in terms of how we evaluate and structure our network. And I've owned the data center business as part of Zayo that we sold. We were a cloud provider and sold to the cloud providers. I feel like we've got a pretty good line of sight to continued operating leverage there.

So you mentioned the strategic optionality that you'll have at Zayo. And so I'll use this as an opportunity to ask about the strategic optionality with DigitalOcean, given how differentiated the assets. How do you think about DigitalOcean future as a stand-alone entity versus perhaps partnering or merging with another technology provider?

Well, I think our job is to operate the company as best we can and manage balanced growth and profitability. We've never -- we've always had a bias to grow faster than more margin, but it had to be in balance. We just philosophically think company should make money. And as we said, we just accelerated margins this year from -- into 2023 from 2024, given the uncertainty. We were here a year ago at this conference and people are like, what's your run rate growth rate what do you think the sustainable growth rate, nobody had an answer a year ago. I'm not sure people still know what the sustainable growth rate is in this different economic environment. And so the right thing to do from a fiduciary is we operate that at a different margin profile. That's what we've done. And yet, we still invested in new products, new teams, new capabilities and longer-term growth levers. So you could do both. We can walk as you come, we're doing that. That's an age-old question that will be cited over time. But this is a stand-alone business with a strong balance sheet, a strong product set, a strong brand, strong differentiation that can go on for a long time in a large market opportunity.

Just to bait around normalized and sustainable growth rates I think, is particularly relevant given the both secular tailwinds and the cyclical factors that are outside of your control is in the medium term. And so Matt, maybe a little bit on how the near term the modern environment is evolving? And a little bit on your perspective on how do you think about -- when you look at the growth that DigitalOcean has experienced over the last 3 years, how do you think about what piece of that is secular or structural versus cyclical?

Yes. That's a great question. As Yancey said, it's unclear how to tease that out of the market to say how much was COVID-related kind of demand tailwinds and what's the new normal going to be. So as we think about it, we think about what can we count on? Like what are the building blocks that we feel like we can rely on as we think forward. . And fortunately, we have a self-serve funnel that generates new customers every year. It's been incredibly consistent over the years, regardless of whether there was a pandemic or whether there was a recession or there was war, we add new customers at a relatively stable clip. And that will give us about 5% growth, just did that. We get another 2% from that phenomenon. If you think about the new customers I added last year still kind of contribute the ones that haven't hit a full year into the second year, so we get another 2% from that self-serve funnel from the prior year sales. So you're 7%. Then you've got our cloud waste business, which is growing at 45% as of the second quarter, which is an incredibly strong business, that will generate 3% to 4% growth for us on a consistent basis. And I can tell you, Cloudways, it's the 1 business that Yancey and I have seen that wasn't impacted at all by the slowdown. In fact, it's growing faster now than it was when we acquired it in September of last year. So that's 3 to 4 points. And then as we said earlier, Paperspace should be 3% at minimum, that puts you at 13% to 14% growth is kind of a floor. But the core of our business, the cohort, the big meaty part of our customer base, that's really what the swing is at that point. And over the past 6 months and longer, we've been getting less revenue year-over-year from that cohort. So it's been a drag. We'll probably exit the year at high 90s in terms of NDR. So that's where we get to the kind of 12-ish percent from a baseline growth. Now if we can get the cohort above 100, now you're building on top of that baseline of 13% to 14%. And a lot of the investment that we're making now in new products, the go-to-market motions around sales and partnerships, those are all additive to that base. And that base is really just -- what can we wake up and we'll be confident what happen every day. we feel like that's the right floor. So whether the new normal is at that floor or whether the new normal is something we hope higher than that, we'll see. But we're planning our business to be able to generate cash flow and thrive in that kind of environment.

And so as you think about your 2024 planning assumptions, I really appreciate the way that you laid out the building blocks like that. Are you all thinking through scenarios where, okay, what is the core high 90s business, what if macro eases, you end up with something north of, let's say, 15% versus what a macro remains a headwind than you sub-15%. Are those are types of scenarios you are looking here?

Exactly scenarios. And then you layer in Paperspace, how big could that be? How much capital would that require? Those are the exact scenarios that we're looking at. And we're paying a ton of attention to the leading indicators around churn and contraction and expansion to make sure that we have the best view possible of where that NDR is going to end because that's the big lever, the big difference maker for [ 2023 ].

What are the leading indicators telling you about the factors outside of your control, mainly the macro?

Yes. So churn, we've said this, and it's continued to play out this way. Churn has been stable all year, a little bit elevated in the beginning of the year. It's stable, and it's stable at historically good, reasonable levels. So it's not elevated. Contraction is elevated. It was elevated by about, call it, 400 basis points versus the prior year at its peak, but it's stabilized. So it's, in fact, both July and August, it got a little bit better, but it's still elevated. . So contraction is stable, maybe improving a little bit, but still at an elevated level. And what we said on the earnings call was that the metric we hadn't seen yet bottom was expansion. Expansion continued to drop from the high 30s a year ago to the low 20s, but it was dropping at a decelerating rate. August was the first month it was flat. And so it's still well below 20 points -- 15, 20 points below what it was last year, but it started to flatten. So we're in the same as what -- I think you're hearing from hyperscalers, bottoming is an appropriate word. Have we hit the bottom, hard to tell. I like August results, looks like bottoming, but we're going to be very cautious as to when we think that that's going to kind of return to a positive trajectory.

Do you have any observations or data that helps you compare and contrast the DigitalOcean installed base with some of the hyperscalers? And what I mean by that is when you have AWS commenting that they believe they're out of trough in their infrastructure business and you look at the data in your business, where do you think your customer base leads or lags the broader infrastructure software or the broader macroeconomic economy. I'm curious if you have observations on that.

Well, I don't know that they said anything different than we're saying. You use different words, but I don't know that the meeting is different. And I think the open question is, and this has been a question for 18 months as this happened is, does SMB behave differently than enterprise. And the markets believe yes, worse, going into it. And we saw it early because of the consumption-based business. So months ahead of when people were seeing -- when you have license revenue early in the year, if there's a slowdown, you don't necessarily -- didn't see it until the renewals start which we all saw this time last year. We saw it in April 1 of last year as the slowdown started. And I think what we're seeing now is we have a window into the slowdown slowing, bottoming. And then the question is, what's the bounce back look like? And is that going to be different in SMB? I don't think we know the answer to that. What we do know is that we are a very efficient way for people to run a business on the cloud. And we have been seeing -- we have a concerted effort to go into our cohort and People are very cost sensitive. I think the elevated contraction is related to people being a lot more diligent about their growth in cloud spend because they've seen significant slowdown in their top line. And that's been playing to our benefit of getting other workloads over to us. So the question from bigger competitors. I think the real question is going to be, is the behavior different. We're certainly slowing in the same sort of we slowed down rapidly a year ago, it seemed very -- we were a little bit early, and everyone saw it is uncertain, and it seems like we're coming out of it. And the question is, is the bounce different, and that gets to what's run rate, growth rate on the other side of this, and I don't think we know. I do know that when we talk to our customers, despite the fact that they may be growing slightly slower, they're very optimistic. And most of our customers tend to be bootstrapped. And so they run to make sure they have money at the end of the month to pay rent. And so I think that plays into the contraction dynamic here a little bit more. But they're very optimistic. So they're still seeing growth. And even though it's not 40% growth as it was 18 months ago, it's still over 20%. They're expanding, which is 5, 6x GDP, right? So that's still higher growth than broadly seen over the economy even as we're troughing or bottoming in our language. So -- but to go forward, we just don't know. And I think we're going to be -- you'll see much more caution about -- talking about that given the dynamics.

Understood. I had to pause for a moment and go to the audience questions from the audience. I get up on the go-to-market and specifically on the efforts that you have made to build out a direct sales force with [critically] carrying it. Matt, maybe for you, give us an update on how that's going, and it can be a little bit of a new motion for companies that have historically had a really strong product like growth, really strong health service. It's a little bit of a different animal. Would love to hear an update there on how you're approaching it?

Yes. There's -- with Aaqib, who is the CEO of Cloudway is joining us and becoming the Chief Revenue Officer. We've put a big focus on trying to build out those other go-to-market motions, the direct sales, which would be primarily aimed at targeting customers that have already have a size that we can then port onto our platform from some of the bigger cloud providers and then a partner motion, where we're leveraging partners to reach segments of the industry where we don't typically have -- the self-serve funnel doesn't resonate. . I'd say both of those are still very early innings. We spend a significant amount of time, focused on how do we improve the onboarding process and identify customers early in our self-serve funnel that have the potential to grow. So there's an aspect of sales in that. It's more of an inside sales, more of a nurturing motion than an outbound kind of cold calling motion. And I say we're getting really good traction there. I'd say the aspiration when we talked to this at the beginning about our guidance range for 2023, it was only going to drive about $5 million of incremental revenue for the year. And we're a little bit behind that, but not materially behind, and I think that it's going to be a much bigger impact in 2024 and beyond as we make the changes that we need to make. As you said, it's a very different motion for the organization. So it's going to take us a bit to get that settled.

What are your sales leaders sharing with you in terms of observations on what's been perhaps easier than they expected and what's being more challenging?

Yes. The easier than expected is customers see the value of our platform. They understand the simplicity. They want the higher level of support and they get the economics. A lot of times, we say that the cost difference can be like 50%. So it's a material story that the salespeople have to communicate with those customers. But what's harder is it gets back to the R&D question that you posed to Yancey, SMB ready when you need to be a bigger company on our platform, you need maybe some different storage capabilities that you didn't need if you were a smaller company and you might need some access control capabilities that are a bit more complicated than what we have provided. . When you have a smaller customer that starts on your platform and they grow, well, they live with what you have and you get -- they grow with you. When you have someone you're trying to convince to switch they may say all of that sounds great, but until you get this particular storage capability, I can't move, it's too risky. And so that's the kind of part of the reason why the sales motion takes a little bit because the feedback that we get, which is really promising, it tells us we're on the right track. We've got to go then go back to development and say, "Okay, we've got to iron out a couple of these edges so that we can smooth the process for converting customers to our platform.

It leads a little bit to the contribution of upsell and cross-sell, more sophisticated functionality, some of the more sophisticated products you have in your portfolio and how that contributes to your growth algorithm? How do you think about your pricing strategy holistically? How much do you think it can contribute to that 15-point kind of growth algorithm? And is there a risk that you perhaps see more compression in the more commodity products and you're on -- you have to ensure that you're able to capture more value, higher up to be able to offset that?

I think the opportunity to use packaging and pricing, not pricing like a list price increase, but pricing a specific set of value capabilities to address a specific value proposition. It's a huge potential lever for us and one that we've not historically taken a tremendous amount of advantage of. So there's a lot of opportunity for us. . You've seen some of the announcements we've made around like the premium and dedicated droplet, which is a bandwidth optimized droplet. Those are the kind of things where, as Yancey said, with the customer he spoke with last night, we're saving them a material amount of money by bundling capabilities in such a way that meets the use case that they have. They're willing to pay for that and they're willing to pay more for that. We've historically had the simplicity of the droplet. It's a fixed amount of compute, storage and bandwidth. And if people wanted more of 1 of those 3, they bought more droplets. And so we've got customers that would buy 2,000 droplets, but it's really because they wanted a bunch of bandwidth. And so they didn't use the compute and they didn't use the storage. They just use the bandwidth and a subset that's a pricing opportunity where you could say I could charge them more and dedicate less kind of capacity to them, and that's a huge lever for us.

The other thing I'd add to that is when you look at how a customer grows when they go from being a learner builder to scaler, a big chunk of that growth at least historically, has come from just organic growth. That's a business that had $100,000 of revenue last year. It had $300,000 of revenue this year, $1 million next year, right? This is massive growth rates. And so that's been a big part of ARPU growth. It's just there are small businesses that grow much faster. And where we've been able to turbocharge that, we've kind of tripled ARPU over the last 4 years, is by product mix. And the more extensive the share of wallet that we can capture, whether it's storage, security, databases, Kubernetes, AI, manage whatever it is, as they grow, every incremental customer they can't take on, we can support more of the workloads that they need, that accelerates what's already a naturally organic growing business. And so I think that speaks to the packaging opportunity versus just the explicit pricing and also what's the right product and how do we bundle those together. And that's an enormous opportunity as we look at it.

The last question I'll ask you is when you think about your conversations with Cloudways and Paperspace, and when you think about the future M&A strategy of DigitalOcean, what are the 1 or 2 questions that you think are most important to spend time with the companies that you're potentially thinking of acquiring, just helping whether it's good or bad for DigitalOcean.

I don't understand. Not sure I understand.

When do you meet with founders in the pipeline for your corporate development and thinking about M&A, what are the 1 or 2 questions that you like to ask determine whether it's good for DigitalOcean?

Well, I think the first question is what is our mission? And does this help us simplify, because adding this to the platform help us simplify cloud in a differentiated way. We're adding capability. Can we do it with simplicity, whether it's day 1 or day 180, but does that fit? Does that mesh? Will that support the way we have to engage with customers? How they would ultimately use the product? Is there something about it that would make it too unwilly to integrate? I think that's really first and foremost because if we just go buy technology that actually can't be integrated in a way that preserves the experience. I think that would be -- that's the -- what we said earlier, that is the ultimate challenge for the business. How do you grow capability over time and maintain this highly differentiated experience that's centered around simplicity.

Excellent. Let's leave it there. Thank you both for your time. We appreciate it.

Thank you. Appreciate it.