Fastly, Inc. (FSLY) Earnings Call Transcript & Summary

All right, everyone. I am Sanjit Singh. I'm the infrastructure software analyst at Morgan Stanley. We are pleased to host our next session with the management team of Fastly. We have Joshua Bixby, CEO; and we also have Ron Kisling, Chief Financial Officer of Fastly. Looking forward to diving in and getting the update from the Fastly management team. Before I do, let me just quickly go through some disclosures. For important disclosures, please read the Morgan Stanley research official website at www.morganstanley.com/researchdisclosures. If you have any questions, please reach out to your Morgan Stanley sales representatives. So thank you, Joshua and Ron, for joining us at the Morgan Stanley TMT conference. I think it's the first time you and I have -- we've had a session together. So I'm really looking forward to get an update on the story.

And maybe just sort of start there in terms of talking about this edge compute platform and where we sort of stand in the evolution of that from a high level. Can you talk about, from a developer-focused heritage, starting with core delivery and performance and moving on into Secure@Edge, where are we in that curve, if you will, as we stand in early 2022?

Sure. Well, it's an honor to be here. Thank you. It's nice to see everyone in person. And this conference itself marks an important moment for us because it was the first thing we canceled as a company and -- 2 years ago and then we went into complete lockdown, and it feels like we're emerging today. So it's really nice to be here. Yes. I think on the evolution, the last 2 years have compressed a lot. But if you rewind 10, 11 years ago, Fastly started with a very simple premise, which is that developers needed to look beyond the database, web and application servers where they had only been able to work their whole lives, right? So if you think of a developer for 20 years, they've been put on almost blinders and they've been told, you can work here. And then the IT department or the marketing team owned everything ultimately beyond that. Because in many cases, the content delivery network was owned by the marketing team. So fundamentally, what Fastly brought to the table at the time was, no, these edges that are sitting all around the world close to users are something you the developer can write code on and you can have an impact. You can make your sites faster, you can make them more secure, you can make them more scalable. So fundamentally, we started out with this notion. Now we started out with the notion that, that would be for delivery, and it would be in a certain language, VCL. The reason we chose VCL was because it was a language that would allow us to make sure we could do this safely at the edge. In the past, everyone had said it was impossible to create safety. We knew that we could create safety through syntax sort of language. And so if you look at the evolution, what's happened is developers have been doing this. If you look at the vanguard of the Internet, we are lucky to have many of them as our customer. They have set the precedent for what this looks like. So their code is at the edge, it's running their websites. Two things came out of that as we evolved. The first thing was we now need to secure these applications at the edge. So we saw the emergence of security being absolutely important and critical to the delivery side. So those came together. We also saw customers come to us and say, VCL is great, but I want to do more things, and I want to do it in more languages. And so that's where you see the next generation of compute. It's not the first generation. We've been allowing people to write code for years. It's that people want to write effectively more dangerous things. And by writing more dangerous things, we need to protect them and everyone else. And that's where the next iteration of compute is. So where are we on the journey? We started 11 years ago by being a place that developers land and write coded edges, we're still that. We are now at a place where we're letting them write more code in more ways and more languages and to tie together security. And we're well on that path for the vanguard of the Internet. Every time we see these evolutions, we see that the vanguard of the Internet, those who set the templates and the primitives for the future, they adopt first, and everyone else adopts after. We are still in that point where we are seeing the largest, most sophisticated organizations adopt this, and then we will see the floodgates open and the rest of the world will.

Great. And so if we think about the opportunity position that you just laid out, where we've been, where we are today and where we're going. A couple of quarters ago, the team put out some 2025 targets, talking, I think it was about $1 billion in revenue. Can you talk about the decision to introduce those targets? What gave the team the confidence that you could effectively reach those milestones, just given where -- how the business has performed over the last 18 months, it's been a bit of an up and down ride. And so what was sort of the thinking behind laying that out there as a milestone for the team to hit?

Sure. So Fastly didn't get to the place we're at by setting low expectations. So I think I would start by saying we came to this place because we thought we could change the world. When you're -- if you use a football analogy, it's really easy to predict you're going to get to the Super Bowl when you're in the conference championship. And that's one way to motivate your team. Actually, the most -- the best way to motivate your team is in preseason, when you look and you say, how can I get to where I'm going to go. That is the point that you say, we're going to win the Super Bowl, folks. Come join us, you can win it, too. And so I think we're looking at it at a moment where after some over the last 18 months, it was time for us to reset both internally and externally and tell folks where we're going to go. And we're not going to do that when we're already at that point. We're going to do that when it's time to really make sure everyone knows who we are and what we're trying to achieve. And I think it speaks to the legacy of the company, but most importantly, it speaks to how early we feel we are in this opportunity. If we thought we were in the middle innings of a baseball game, you would be a little bit more cautious or think about it in a different way. We are at the earliest stages. Only a couple of hundred companies in the world have adopted these primitives for the future. There are hundreds of thousands who can. So I think we're looking at it and saying, this opportunity is the start of something huge. We're in an investment phase. We're going to invest, but we're also going to show people how large we think this can be. So I guess it depends on what kind of coach you are.

Yes. [ I guess the questions is ] -- so please unpack the path to get there, if you will. Obviously, a key part of that is Compute@Edge. And you mentioned it's early days. But if you can just sort of walk through some of the top use cases that you see today, sort of part one of the question. And then the second one is sort of the timing of inflection, right? You've talked about reaching 100,000 developers, I think, by the end of this year. What are some of the intra-milestones that you guys want to hit to get -- to give you confidence of that path to that more durable 30% growth profile?

Yes. So let me start, and then Ron can talk a little bit about how that path carries on. I think we started very clearly with the developer milestone. And why developer milestone is important is because what we find in organizations is that when we have the hearts and minds of the developer, everything else flows for us. So imagine the way creativity happens is not, although some of us would like to think, based on what an executive says and then pounds down in the organization. Creativity comes the other way. So when one of the developers at any of your organizations or mine is trying to be creative, they go online in their own time at midnight, at their computer and they start playing around with things. So we need to make that as easy as possible. We want to be within the fabric of every prototype that exists. And that is an -- the best proxy for that as enterprise developers. So I think it starts with enterprise developers. We want them to use us, we want them to try us. There are many examples in our history where we were a prototype in someone's mind. That prototype got traction. No one ever asked whether Fastly was embedded or ran that. It just was the future. Like this is the future of my company. It was written by this group. And only at the point that we get put into production, the people ask questions of like what's behind this? Why does it work so well? Why is it so fast? Why is it so secure? Why is it so scalable? So we need to be there at the moment of inspiration. So I would say if you look at that trajectory, it starts with developers. The next element of that trajectory has to start seeing our acceleration of new customer adds. So we've been exceptionally good at nurturing our customers and growing them. We need to be equally good at getting new ones. And that's been something that, based on the history of the business, we grew through word of mouth. It takes some time for us to get that engine going, but we're confident about the investments we've made and where we're going up. Ron can talk a little bit about the inflection around some of the other points.

Yes. I think the important part about that customer acquisition has to do with a lot of the investments we've made in the sales organization. A year ago, we brought on a new head of sales. And he's been building up that sales organization. We focused on reorganizing the sales base. We've relooked at our comp plans. And those reps are now coming up to speed. And so if you look at the 9-month ramp for a lot of those reps that are coming on board in the typical sales cycle, as we get to kind of where we are right now, we should start to see acceleration in terms of new customer acquisition from these things we started doing a little over a year ago. And these new customer acquisitions, as they ramp traffic, are going to drive real revenue growth as we move into the second half of 2022, as we move into 2023, and really leverage the engagement from the developers around the edge as well as the differentiation we have in our overall network that competes very well against the competitive networks.

Yes. If I could just dovetail on the point that you made, Ron. You mentioned that Brett joined as Chief Revenue Officer a little over a year ago. What has he been working on operationally to drive that better new acquisition performance? I'm sure there's an existing customer base expansion motion there as well. And in terms of what he wanted to accomplish, how far away through is his plan, if you will? Is there more to go in 2022? Or is there -- are we through the -- most of the disruption, if you will?

I think Brett brings a number of things. I mean he's built large sales organizations at a number of enterprise companies and brings that experience here. And so I think coming onboard with building his leadership team, also looking at how we're organized, moving from -- we had a vertically organized sales team. As you look at efficiency in the organization, moving that to a regional organization increased the productivity of the sales team. Looking at our comp plans to make sure that they're focused on how our business works in terms of gaining new customers. And then once we gain new customers, we see a very strong pattern of increasing usage, not only around delivery but other products. And building an incentive plan that's really focused on that dynamic across the sales organization. And then lastly, just a great deal of discipline around the sales organization and building that discipline into the sales team. I think in terms of where we are, as I talked about, Brett's been onboard for about a year, he's built his leadership team and he's done a lot of ramping, particularly around the sales organization, particularly in the enterprise space as we continue to ramp that. And we're starting to see those sales reps drive productivity and customer acquisition. The other thing that Brett has brought on, and this is also leveraging some of the stuff from the Signal Sciences acquisition is they largely sold through a partner network, which is something we hadn't really done effectively in the past. And so leveraging off of their partner network is looking at how we deliver across our whole product portfolio and really focusing on sort of that non-enterprise segment of the market and growing that piece of the business. And so a lot of those things are in place. So as I look going forward, I think we have a ramp sales organization in place. We're going to continue to invest and grow that organization as we see opportunities and the size of the market to continue to grow and accelerate customer adoption.

Makes a lot of sense. The other big piece of the opportunity outside of delivery and performance and Compute@Edge is security, right? And again, to some of those milestones you laid out, you talked about increasing the security revenues quite substantially, about 10x by approximation. What is it going to take to get there, right? So we have Signal Sciences as a big piece of that equation. We'll talk more about Signal Sciences. But are you going to get there organically? Is there more organic R&D that you need? What's sort of the road map for that security business in that longer-term framework?

Yes. Let me start with the road map question, and then Ron can tackle the organic/inorganic side. I think that we are looking at a web and API security market. So that we're looking at people who own large websites, large APIs, mobile applications, they want to protect themselves. There's another side of this market, which is the IT market. That's not a market we're in today. So we're looking at the API and the web side of this business. If you look at that market, we have competitive products in all of the arenas that we need to have competitive products on. So the web application firewall, which we acquired through Signal Sciences, and we recently relaunched with an option on the Fastly platform. That completes that vision, but there's still incremental work that we need to continue to do. We have denial service products, protection products. Those continue to evolve with the actors, but we've got a very large network and provide protection to some of the largest organizations in the world. And the last piece of that is the bot protection side. And we have a bot protection product. You're going to continue to see R&D. So I think if you were to sort of look at incremental R&D on the DDoS and web application side, we are putting in more resources into the bot side. Because one of the things that's amazing and actually is escalating with the supply chain challenges is this bot problem is starting to affect not just shoes and tickets, right? So if you think of the classic example, it was wherever you can get arbitrage. So I can go buying Adele tickets for $200 and in the 7 minutes that I've got Ticketmaster holding it for me, I can sell it for $1,000. That is a beautiful arbitrage model. We can get bots to do it, and we can pay humans to do that job. So we and others have been battling these incredible arbitrage opportunities. And traditionally, we thought of them as like sneakers, PlayStations. The reality is that when we talk to retailers, they're suffering this across normal commodities that all of us would think should not be scarce, right? So they've got a toilet paper challenge or that toilet paper challenge turns into a mask challenge. The reality is they don't have price leverage. They can't go and increase the price of a PlayStation. The only thing they can do is make sure it gets in the hands of a loyal customer. Because the worst thing they can do is have their entire PlayStation inventory now on eBay or now on Craigslist. And so what's interesting is this bot problem, be it that or scraping of gift cards or trying to get into account takeovers. Like these are the problems of the modern company, even though they thought they weren't in this problem space. So you're going to see a continued investment on the bot. I'll hand it over to Ron for the organic/inorganic side.

Yes. So I think broadly speaking, as security becomes more important, as Joshua talked about, we're seeing it across sort of every deployment. Every delivery has security as part of it. The other thing is we're even seeing, given the capabilities and the differentiation on the Signal Sciences product, that's also something that we're seeing where we're selling it into sort of non-Fastly delivery. And so across all of that, that's what's driven the growth rate. So coming off of Q4 '21, the Signal Sciences growth rate adjusted for accounting was about 40%. So a very strong growth rate that really plays into that 10x thing. As we look at it going forward, we're largely looking at achieving this on an organic basis. There's additional organic investment in R&D, as Joshua talked about, as we continue to evolve our DDoS and our bot-detection capabilities. And we'll continue to evolve those in the future. But the view is that most of that comes from an organic basis.

Can you talk about the go-to-market behind the security, which you mentioned the focus on web and API security. And so does that imply that you have the relationships on that side of the house already? Or is this a set of purchase decision makers that you're going to have to go out and recruit? And in terms of like Signal Sciences, how is that -- what is the go-to-market strategy been over the last 12 months? And how is that going to change going into next year?

Yes, it's a great question. Signal Sciences came in, 85% of their deals were done through a channel. That was very different than our business. We've learned a lot. So I would say we want to really continue that. We want there to be a channel-driven market. Imagine the security channel is very different than -- the delivery is effectively been a monopolistic industry. So if you've got one provider, so we can all go look at who the top 10 banks, the top 10 insurance companies, up until a few years ago, the top 10 e-commerce customers. Every top 10 was 1 company. And when you have one company that is still stuck in the 2000s, i.e., a product that is undifferentiated, not -- you can't code on it, you can't extend it. There are no real APIs. You actually can't get value-added resellers to add any value. Because the only way anything is done on that platform is throughout a dent chart, a project management team and a whole bunch of consultants. So that makes a partnership model in delivery almost obsolete. When you combine that with a monopolistic position, there are no delivery partners out there. And so we started in delivery. So we encountered a world where we would have loved to go through channels, but they didn't exist. Now security is completely different. If you think of how security has been run, pre-cloud, it was appliances that sat all in line in front of each other. Those appliances were put into place by value-added resellers for the most part, and they were deeply trusted by the client. And so we are trying to look at that market and say, okay, our largest competitor really can't go through partners. We think that's a unique leverage point, and we're going to go through partners. So we are seeing really strong partnerships, not only with the hyperscalers, which is a slightly different vein, but equally important. But also through this value-added community, where they have been delivering boxes and those boxes don't have a place to live anymore and they need a central unified cloud security story arc. So I would say it's very, very important to us, and we're seeing really significant traction.

So it's not so much of a question of identifying the end buyer. It's really about, hey, this is channel out there. We can get a lot of leverage from. We need to build that muscle, if you will.

Correct.

Great. So one of the things I get asked most about coming out of last quarter's results is really on the gross margin front, right? And so the gross margins have ticked down for a couple of quarters. It looks like it's going to sort of stay in that sort of mid-50s range. Correct me if I'm wrong, Ron. But can you give us a sense of sort of -- essentially, you guys are telling us there's more traffic that's going to come on. You're investing in front of future consumption. What gives you that confidence that traffic is, one, going to come on board? And two, how long of a -- how long is this period of investment on the gross margin side going to sustain before that sort of ticking up of traffic?

That's a great question. And I think we talked about in 2021, a lot of the investment was really around expanding our footprint geographically and spinning up new locations. And a lot of that was to support traffic that we knew was coming because a lot of our multinational customers that currently use us today, as they expand to new countries are saying we want to use you there. And so that process of spinning it up is an incremental investment. You build it before the traffic is there so it's ready. And then the customers typically ramp over time. So that initially, there is a little bit of a gross margin drag until they've ramped their traffic up to their full level. Secondly, as we look at some of our largest sites, we're looking at the architecture in 2022 in terms of trying to take where we've got some of the largest sites. We've got multiple pop locations. And today, they operate as individual ones. We're looking at having those -- all operators kind of one site that drives a lot of efficiency and effectiveness in our market. But that transition requires us to sort of run both in parallel for part of the year. And so that is going to be a headwind to gross margin this year as we continue to build out our state-of-the-art network. A lot of that redundancy goes away in the second half and so gross margins in the second half of 2022 should recover as that redundancy goes away. And as that fully goes away, we should see accretion in gross margin and improving gross margins going into 2023. So it's really been about making this investment in our global market. And then as we ramp into '23, we see improving margins.

Yes. I have a follow-up on the gross margins, but before I get I want to see if anyone in the audience had any questions for the team before we hit some more topics. All right. Looks like we're doing a pretty good job. So on the gross margin. So we're talking about the shape of that recovery, right? And then ultimately, the question is like, well, where does it go long term? And I think during the pandemic, you guys have proven out sort of low 60s, I think, 63% gross margins. What -- how long do you think it will take to get to those levels that we saw during the pandemic? And what's the path sort of to the longer-term gross margins?

Yes. I mean as long as we're seeing significant growth in terms of traffic and an opportunity to build into that investment and sell into that investment, there will be some level of investment that's going to have some impact on gross margins. However, given a lot of the geographic expansion we've done, as we move into 2023, we should see continued accretion in gross margins in terms of those improving. And as we get to those utilization levels that we saw in 2020, the low 60s are very possible gross margins in terms of where we see in the medium to long term those 60%, 62% gross margins are very achievable.

On pricing, like the equation I've always had for kind of the delivery market was sort of traffic growth minus some element of annual price erosion times market share gain or loss, right, like your basic construct. And so on the pricing side, how would you characterize -- it seemed like 2021 is actually a pretty decent year for pricing. How do you feel about pricing going forward in 2022, particularly as you bring on more workloads and that's particularly as edge compute starts to come online? What does that do to pricing?

I mean I think we're in a few markets here, and they do coalesce, but we do see different dynamics. We've called out the video market. Ron said, and we've seen this over the last 2 years, and I think we'll see it again this year. We've seen a deceleration in the price compression. I think that's a combination of discipline on the side of the other vendors, which I think is great. I think we are also seeing the fact that we have an inflationary environment. So that plays into it as well. So I think we will continue to see that trend. So we're going to see a deceleration in pricing environment on the video side. It's really important. On every other part of the business, there is no price compression at all. I mean you have a monopolistic player who owns every single account, right, minus the few that we have. And again, this is so early that I'm saying a few. We've got 5 or 6 of the top 50 e-commerce customers, they've got 45. And there is literally no price compression that's happening in that market. In fact, as we bring new services to the table, particularly on the compute side, we actually see those going up compared to what our competitors do. So I think you have to split this into 2 very different worlds. And that's sometimes hard to think through. Now the reality is this video traffic is central to our thesis. Like we need that traffic in order to best leverage our infrastructure, in order to get a long-term cost structure that makes sense, in order to be in a unique position with vendors. We're seeing that right now. So like imagine that there are people -- as we all know, we're putting satellites up to disrupt the Internet market. If you look at that market, they don't have the space to go have 10 of us on each satellite. They're going to pick the 1 or 2 that actually have the traffic profiles that will work so that they can do their job up in the sky. We set ourselves up to be, I believe, a very clear option when you look at those kinds of decisions. So strategically, video is very important to our business. It also long-term changes our dynamics in terms of our infrastructure costs. So I think that there is a misunderstanding about the strategic value of being in this market. Now there are huge chunks of this market that are not interesting, right? And we've been very, very disciplined and continue to show discipline in that regard. But this is highly strategic for us. Video is highly strategic, and I think long term, that will be proven out.

What are the areas you're trying to avoid and sort of...

Sure. So I mean I think the easiest way to think about it is video that you consume that you don't pay for, we don't want to be in the market for. Or gifts that get sent around the Internet that none of us are paying for. Like there is an entire industry here, and I just would think about it from your perspective. So if you are associating yourself with a brand, that brand is charging you $12, $15, $20 a month. And that brand, even though most of it is in the video-on-demand side, the reality of video-on-demand is totally changed. So when something drops like a new series of Billions or a new series of Succession, that actually feels like live television. We're all talking about it on Slack or messaging our friends. We're all watching it at the same time. So if you have a degradation in performance, it has that same effect as when the Super Bowl is degraded or the Masters has degraded. So what we forget is that we actually have moved to what is much closer to a live experience in video-on-demand. And it's in those scenarios where it really matters. I can tell you, when a new series of Clifford the Big Red Dog comes out, like my kids don't complain, it doesn't matter. But when Succession comes out or Billions comes out or Games of Thrones comes out, it really matters. We all talk about it, we all notice it. So the markets that matter is all of us are shelling out money from our -- like it hits our credit card, we actually notice it, and we choose to keep paying it, that is probably the best indicator that you should see Fastly in that transaction, be it video-on-demand or live because live is a totally different beast. We screw up Tiger Woods walking down the 18th green of the Masters, you all know about it. We get in trouble. And so there is like -- live has always been that way, particularly in the sporting world, but we're now seeing video-on-demand move in that direction as well.

I'm the worst streaming customer. I signed up for streams to all these subscriptions. I never cancel them. It's like the worst.

Many -- some of the streamers would wish they were more like -- more customers like you.

I want to hit 2 more topics. I want to see if there's any questions out in the audience. We've got one in the back here.

Okay. Yes, on Compute@Edge, you have this wonderful program, right, like for your customers. Can you give a little bit color around that? How is that going? Would be the first one. And then the second one would be on your growth, right? It decelerated quite materially, like we are more or less on market growth. When do you expect you can come back to this 30s you are expecting midterm?

Sure. So I'll take the first one, and Ron can take the second one. I think you're right. We have announced a number of things on the compute platform recently. All have been very successful. You're probably referring to the credit program. But there were 2 aspects of what we talked about at our last earnings. The first was, we want developers to try it, so we created a free tier for them. That's been incredibly successful. We talked this quarter about one of our partners, Glitch. They've got 2,000 enterprise developers in a week. Those are the kinds of really encouraging signs that enterprise developers are attaching to the platform and using it. We also announced a credit program because one of the things when we get to new technology is we need to make sure that CFOs don't get angry. And when we go and say, hey, look at this amazing ROI when you buy our product, you need a month of your Amazon bill to prove it's true. So we do need to like sort of walk with the customer down that path. We introduced that program. We had 40 customers who attached in the quarter. We announced it halfway through the quarter. We extended it through Q1. And I think what that's telling you is we're at that next inflection phase. The first phase is I want to try it in the lab. The second phase is I want to try it in production, but I want to try to -- I want to make sure that I don't get myself in trouble with my financial team. And that's really what that credit program is going to do. If you look at almost every evolution, you see that we have to partner historically in all of DevRel tech, you've got to partner with them to get through that financial barrier, and that's exactly what we're doing. And Ron, speak to growth quickly.

Yes. I think real quickly on the growth, as I talked about, since Brett came onboard with a lot of rebuilding of the sales organization. And as that team ramped, you didn't have sort of the acceleration in new customers that we're going to see going forward. Given that he's been here about a year, that sales team is ramping, the sales cycles are closing. And so we expect to see an acceleration in the number of new customers this year, and those customers will really contribute to revenue in 2023. And so we would expect to see an acceleration in growth as we move out of this year and into 2023.

And with that, we're all out of time. Thank you so much, Joshua and Ron.

Thank you.

Thank you.