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

All right. Good afternoon. Welcome back, everybody. Thanks for joining us. My name is Rishi Jaluria. I cover software here at RBC. I'm delighted to have with me from Fastly, Joshua Bixby, who is the CEO; and Ron Kisling, who is the new CFO here. Joshua, Ron, thank you so much for joining us today.

Thank you. It's a pleasure. Thank you for having us. It's nice to be with RBC. As a Canadian, it's nice to represent up here.

I love it. Yes, the Canadians have been very nice to me since I've joined. So great to be here.

I'm glad to hear that.

And for any investors on the line who have questions for Joshua or Ron, please submit them to the box or e-mail me directly, [email protected].

So I want to start maybe high level before we get into a lot of the specifics. And Joshua, maybe just talk about the evolution of Fastly, right, from being kind of this more developer-focused CDN company into this edge player and growing security presence. Maybe talk a little bit about that.

Sure. And it's an amazing 10 years of evolution because 10 years ago, we made the bet that developers would be the new decision-makers when it came to how applications were delivered and secured. I mean that was the bet. The bet was that our founder, Artur, who is in a position, extremely frustrating trying to find ways to deliver off website in the world, how do you do that if at the end of the day, you can't control the delivery? And what had happened was this market had evolved very much as many other markets had evolved where basically, all the decisions are being made by professional services people in 1 company, and you could code on the platform, you could program it and it never felt like your own. You had no control and visibility. So foundational to Fastly is the developers' edge, this idea that developers or the decision-makers within the organization, they're the ones driving digital transformation, the ones putting the tools together, and they need their jobs to be made easier. And the way to make a developer's job easier is to give them full visibility and control, to give them security that doesn't slow them down, but speed them up, and to give them the ability to have a platform that feels and acts like one they would build for themselves. And ultimately, the foundation of Fastly, the thread that carries us from our birth all the way to today and I believe very much in the future rests around allowing developers the power to deliver scale and secure the applications of the future. That's the thread that connects us.

Wonderful. All right. I'm going to jump straight into what's a big question on investor minds, and that's the 2025 targets that you introduced on the Q3 earnings call. Obviously, really great aspirational goals. Can you talk a little bit about, a, the decision to introduce them on an earnings call? And then what is it that's giving you confidence in being able to set these targets, which, we've all done the math, is a meaningful acceleration from where you're going to end up this year?

Sure. I mean from our perspective, we have 10-plus years of history here that we can look back on. There is 1 year where we haven't grown at sort of the target rate that would get us to these targets. But ultimately, we look back and we think that there are a lot of good reasons why our business will continue to grow and to your point, accelerate. There are 3 main areas of targets. The first one, which is foundational, actually building on what I just talked about, was our targets in '22 and '23 to not only bring more requests on to our new compute platform, but to bring more enterprise developers onto that platform. What we know from history is that when a developer uses us in a moment of inspiration, he or she brings this and drives it through the enterprise. So that was the first set of targets that very important to us that we maintain that. The second sort of area of targets is around the security business. And we talked about the fact that we have a meaningful security business today. We believe that, that will only grow at scale in the future. And we said that we believe very strongly that business will be 10x what it is today in 2025. That is very much based on the evolution that we're seeing in the market. Security is elevating. It's becoming part of every conversation. And in the past, you may have had a conversation where you talk about delivering an application or an API, but not securing it. Those are long gone. Every conversation starts with those 2, both delivery and security, as front and center and inseparable. And then lastly, as you talked about, this goal of increasing our revenue to $1 billion. And I think all 3 of these targets are based in our history. We have seen this kind of growth. It's hard for people because 2021 is an anomaly, I think, to see who we are in our core. And we thought it was really important at an earnings call and at this moment where we see some of that confusion. We understand why it's out there because 2021 has been an anomaly. We understand why it's out there. We want to make sure everyone understands what we see, and we love the public accountability that will come with that. So we've never been shy of that. That's something we've always stuck to in our history. So I think it's an exciting time to be sharing that, and we see a really exciting path ahead of us.

Wonderful. And so now as we think out to 2025, right, with that $1 billion target, there's obviously kind of hyper growth modeled in security in there. But to the extent possible, how should we be thinking about how much of the business is going to come from the core, whether we want to call it CDN's performance, that kind of core area versus what are actually edge computing use cases versus security? How should we be thinking about that split at $1 billion in revenue?

Yes. So we plan to 10x the security side of the business. And as we get to that $1 billion, the remaining is going to come from our delivery, our computed edge and additional projects that will continue to roll out as we increase the capabilities of our edge delivery network, which ultimately makes it easier for companies to move toward innovating at the edge and take advantage of our edge network.

Got you. Got you. Okay. Fair. Drilling down under the security side, right? Signal Sciences, it appears to be doing well. But can you give us a sense for how integrated has it become within the rest of the security portfolio? And there were areas where Fastly had existing security offerings and those overlap with Signal Sciences. How does that integration look like where there's not an overlap? Is it certain pick and choose? Is it combining them? And how should we be thinking about how well the rest of the security portfolio has been performing?

Sure. So from day 1 when we've looked at this, for us, it was this emergent quality, which is Signal Sciences value proposition and product offerings plus our value proposition and product offering even in the security space would create more than 1 plus 1 equals 2, it create 1 plus 1 equals 3 or more. And that's the philosophy we've taken. So if you look at integration, we've taken the strengths of both dives. We've merged them together. We did that both on the sales front, where we've learned from them and they've learned from us, but particularly on the product front. So what you see is it's almost inseparable almost a year later when you think about what's Signal Science and what's Fastly, we try to bring it all together because that's what customers want. What we hear from customers all the time is they're sick and tired of all these disparate solutions, either from different companies or brought together through acquisitions without real thought to making that an integrated story. So that's been the premise. And I think I look at the centralization of security as a massive driver for some of the growth that we see in the future, obviously, when we look at the targets we put together. You can't separate these 2, but ultimately, we also see how important innovation is. And so you see, we've made significant investments in innovation, and we will continue to. There's a great product road map ahead of us that will be delivered in '22 that is really the culmination of those 2 worlds coming together. The rest of the security portfolio continues to grow. And as I say, it's almost inseparable. So I think by the time we get to '22, we're not going to be talking about what's Sig Sci and what's Fastly. This will be 1 story deeply integrated with a number of new products being added on top. It's an exciting time to be in the security business.

Got you. All right. I want to rewind the clock back a little bit. So 2020 was obviously an anomalous year with the pandemic. Just how do you think about the sustainability of some of those tailwinds we saw, especially we can think about the work from home trends, but especially related to the stay-at-home orders, right, that led to everyone consuming significantly more media, especially in the early stage of the pandemic?

Yes. So as you point out, I mean we saw digital transformation in 2020 accelerated the pace that was sort of unprecedented. And as you know, we have seen that level off in '21, but it remains exceptionally high. If you remove that elevation that we saw in 2020, 2021 saw a return to 2019 seasonality. And also, if you look at kind of the 2-year view of the business from 2019 to '21, our growth rate was north of 30%. Now that does include some of the inorganic impact of Signal Sciences, but you did see that health. In addition, coming out of that, you see health in terms of the number of new enterprise customers that we added at the end of Q3. So we saw the highest number of new enterprise customers on an organic basis since we went public. And so our guide for Q4 is consistent with previous themes. We expect traditional seasonality will return in '22. And I think as we return to that traditional seasonality, which is kind of what we saw in '21, it should make some of the comparisons when you start to try to understand our business and our year-over-year growth a lot easier to understand going forward than this year given the outsized impact of the pandemic on our last year results.

Got you. Got you. Okay. I want to then turn to the question, again, another one I get a lot from investors, which is around customer concentration, right? You still have a pretty meaningful customer concentration with top 5 customers, 22% of revenue as of last quarter. Do you think about this as a major risk, right, especially given what happened last year with ByteDance/TikTok, and then this year with your large customer that moved a lot of traffic out post outage and maybe it's bringing it back? But maybe help us think through the customer concentration and how we should be we thinking about that.

Yes, that's a good question. I think it's clear we do have a meaningful level of customer concentration. It does create variability in our results. And we talked about that both certainly last quarter and we talked about how we take that into account when we start to provide kind of our own internal projections. But there are a number of things that we're doing to sort of reduce that risk over time. I think, first, as we accelerate our product innovation, and given the work that Brett Shirk, our new Head of Sales, has done, who joined about 6 to 8 months ago, with our sales organization, we're growing our enterprise customers. As I mentioned earlier, we saw an acceleration in that growth in Q3. As we increase and have a larger number of enterprise customers, we're going to reduce that reliance on an individual customer and some of that volatility will subside. Secondly, one of the other things is as we start to focus on the SMB market, leveraging the partner network, this further reduces our reliance on some of these large customers that can move the needle individually in terms of our forecast. And then I think, thirdly, the security business, which is today the Signal Science business is predominantly SaaS-based, which is a recurring revenue streams, that business is growing at a faster rate. That's the business along with our security offerings. It's going to grow 10x. As that grows and becomes a bigger piece of our revenue, that's a more predictable stream of revenue and will further reduce some of the volatility. So taken together, we see -- we've gotten some great big customers that are very important to our revenue growth. But these 3 things are going to continue over time to reduce the impact that any one of those customers have on variability and we'll get to a more predictable state.

Got it. I just want to follow up on that, which should we worry about pricing pressure with customer concentration? And as that customer concentration, as you talked about, Ron, comes down, does that become less of a concern?

Well, I think as you -- certainly, as customer concentration comes down, I think pricing pressure, particularly from an individual customer, does tend to come down. I think the other important point is you look at pricing pressure and while pricing is always an issue in this industry, what we've seen over the last year is really a stabilization. We've seen the rate of decline in pricing across our customer base decelerated in 2021 compared to rate declines that we saw in both '19 and '20. So we think the trends are right as we expand and grow our customer base. That also becomes less of a variability in terms of our outlook.

All right. Now I want to pivot and talk a little bit about edge. Obviously, that's increasingly the messaging here. I guess I'll break it into 2 parts. Number one, how do we think about actually sizing the true opportunity just given how nascent edge is and enterprise is today? And maybe as an illustrative example, help us understand where do customers use edge cloud versus a central cloud?

They're both great questions and I think really topical because to your point, edge is both nascent, but also exploding in terms of its capabilities, technical capabilities and therefore, the use cases that can be put on them. I was talking to a CTO about a month ago, who's had this great sort of story arc of how they solve a traditional data center. And actually, in the story, we see how the cloud and how compute is seen. They said, listen, for every $100 I used to spend on compute in my data center, I spent $5 on security and I'd spent $5 delivering that content, right? So there's $100, you take $5 on security, $5 on delivery. I think in the cloud world, and this is the discussion we are having, okay, how do we see this in the cloud world? And where does edge compute play in? And their answer was, listen, it's no different. I'm spending $5 to $10 on security through cloud services. I'm spending $5 to $10 on delivery. But then I'm looking at that $100 I'm spending on compute today and storage and other costs, and I'm saying, how much of that can I bring over? And I think the answer to that lies in the use cases. So compute -- edge compute is not made for everything, right? There is a very short list from my perspective of things that will not be done at the edge, right, large persistent data stores, all these legacy applications that are built directly into these large data stores, machine learning training. There are things that are so compute intensive that the financial sort of story arc doesn't make sense. Ultimately, I think the answer is everything else does in some form or another. So if you look at some of the use cases that we're seeing today, personalization is a great example. Today, in order to personalize for you on an e-commerce site, that's heading all the way back to a central data center stuffs being crunched. In that time, I've lost your attention, you've gone somewhere else, someone else is faster. So every developer today is facing this, do I personalize? That's going to slow it down. Do I not personalize and choose speed? Edge computing takes that decision away. You can do both. And that idea of you can do both is central to this. Before, I need to make a decision about making an AB test. Am I going to give you this test or that test? Well, that's been really slow. So not a lot of people have tested as many as they should. E-commerce has been central to this. They've shown that if I do test after test after test, I can look at that data and make a really good decision. Well, e-commerce companies are now testing using edge compute. So all of these examples lead to fundamentals. Speed, it's faster when you use edge computing. Scalability, getting this out all across the world right next to you on many, many servers in a serverless environment, more scale. And ultimately, it leads to tremendous security benefits as well because I keep the bad folks farther away from me, and I have the ability to make decisions at the edge where all of my crown jewels don't exist, right? All my credit cards aren't there. All of my personal information isn't there. So I think the way to look at this is it is a chunk of that $100 that's spent. We don't know if it's 10 or 30, but I think the opportunity is for it to push up. And ultimately, we are in a decision-making process where we're sort of asking the question, what should be on there. And I think the best question is what shouldn't be on there because everything else has a role at the edge. And fundamentally, most of the applications that are being delivered by the sophisticated and most sophisticated companies in the world are leveraging the edge in amazing ways even today, let alone when that gets adopted by everyone who follows those leaders.

Got you. Got you. All right. And then I want to think, look, we talk about central cloud versus edge cloud. The big guys who control the central cloud, right, your Amazons, Microsofts, Googles, is there a worry that they become more competitive in this edge over time, right? I mean we've already seen Amazon house Cloudfront. Obviously, they use everyone else for edge right now and for core CDN. But how should we be thinking about the potential for those 3 to get into edge?

Well, I mean I think the confusing part about the term edge, we're part of this confusion and so we have to be really careful about terminology here. There are many edges, right? My phone is an edge. The cell tower is an edge. Where we sit, which is the last place developers could really cross what we call the developers' edge, that's an edge. There's an edge tier data center. So I think part of it is to really be clear about what edges we're talking about and what the purpose is. When you think about our edge, we're talking about the developers' edge. This is the last place you can put code in a place that's PCI-compliant, SOX-compliant, meets all your requirements and that you know is a safe operating environment. It's very different than the phone, right, as an example. And so if you look at the edge that we're defining, that's very different than when you see edge come out of some of the hyperscalers. Some of the hyperscalers' edges are very different places and they're trying to do different things. Fundamentally, I don't think this comes back to what a hyperscaler can or can't do. What this comes back to is what customers want. And what we hear from customers all the time is like in the perimeter of the data center where you had a Cisco or a Juniper load balancer -- or edge router. You had an F5 or NetScaler load balancer. Like the reason those 2 worlds existed separate from the server world was because what every CTO wanted was the ability to add Dell servers or HP servers or whatever. They didn't want to have complete lock-in. And so what we're hearing from customers all the time, which I think is going to set the blueprint for this, which is setting the blueprint, is you want a neutral third party that sits outside of the hyperscalers in order to centralize your security portfolio. For example, you don't want to have a different WAF in Amazon, a different WAF in Google. That's not what people want. But they do want to have the ability to go to those hyperscalers for what they do best, and what they do best is raw compute and raw storage. That's what they do best. And so I think the answer is what customers want always drives the market. What customers want here is the ability to use multiple central clouds, not necessarily for the same application, we don't see a lot of that today, but across applications. We want -- what they want is centralized control and visibility into what's happening. And that's really where Fastly plays in. Ultimately, you can't blanket all the clouds together either. You've got Amazon in a dominant position. You've got Azure, GCP and others, Oracle, for example, in a challenger position. And I think over time, they're all going to act differently because of the positions they're in. But I don't think it matters. What matters today is what customers want. What customers want is not complete vendor, lock in, in a walled garden. The worst thing you can do is go all in on 1 cloud.

Got it. That's really helpful context. And then I want to think about the competitive environment. I think you versus Akamai, really well understood. But increasingly, I get asked all the time about Cloudflare, especially because you're making a lot of investments in ads, right? They just introduced their storage offering. How do you think about where you sit versus Cloudflare and both from a market perspective as well as a just pure technology perspective?

Sure. I mean what's happening in the market is an evolution we see taking place over the last 10 years, which is you have -- for the large enterprise web workloads, you have Fastly and Akamai battling for the top e-commerce customers, the top media players, the publishers, the top industrialists. Like that is the competition at that kind of scale of that kind of level. And that really hasn't changed materially. I think what has started to change is in that mid- and lower end of the market, we certainly are seeing a much more competitive environment than it used to be. And I think that competition is being fed by, in some cases, a SaaS bundling types approach where it's like all you can eat. That's something that's attractive to certain CFOs, and that's something that we are very cognizant of. One of the benefits of bringing Signal Sciences in was our ability to really price and learn about that SaaS business. So I think that's really interesting and something we're learning a lot about. I think on a technology front, enterprise technology at the scale and the level that we're operating and Akamai is operating is just a different piece. I mean we can't experiment in the same way. I envy what some of these other players can do in terms of bringing technology to the market that isn't yet ready for prime time, but it's fine if you're a restaurant or a dry cleaner. Like it works really well. The other element to understand in the market dynamic, which I think is something that's poorly misunderstood today, is we are not in the IT security business. We're not in the SaaS business. That is there are competitors who are getting really good traction in that legacy side of the business. But if you went today and looked at who are the top 20 e-commerce customers use in the world and you look at this list a year ago, you were only going to see 2 names on that list because we're the ones working at the enterprise level. So I think ultimately, we're seeing great execution out of other companies. They're doing it in slightly different markets, and we really don't come up against each other most of the time.

Got you. Got you. All right. I got 2 investor questions, so I'm going to pivot to those. The first is, you've seen your DBNER and NRR both decline to lows this quarter. How should we think about these metrics going forward, especially to underwrite your long-term targets?

Yes. I think similar to the earlier discussion about some of the comparisons of '21 to '20, DBNER, LTM, which are annual metrics, are impacted by some of the activity we saw in 2020. More specifically, if you look at the quarter-on-quarter declines, if you look at last year, those metrics included kind of in the numerator, if you will, our Q2 activity, which was increased very significantly due to the COVID dynamics in terms of Internet traffic. And so as you move from Q2 to Q3, that activity sort of dropped out of the metric. And so on a quarter-to-quarter basis, you get this weird comparison that shows the decline. So I think if you look at kind of what those metrics have been this quarter, they were still north of 100%. And look at our historical ranges, you'd expect -- we would expect to see -- continue to see those DBNERs and LTM NRR north of 100% and at those levels. So I think, again, some of the anomalies of 2020 are showing that quarter-to-quarter decline.

Okay. Correct. And then how should we be thinking about the potential impact of supply chain issues on your business? And when does this start to become a concern?

So unlike a lot of companies that have come to the market recently and said supply chain, supply chain, supply chain, that's not what you've heard from us yet. And I think that's because of how -- 2 dynamics. One is how we source is a little bit different than other companies. We go directly to the manufacturers. And uniquely, although we don't have a massive scale compared to some of the others, we have really unique use cases, and we use people's hardware in ways that is unique and special to them. We often become a test bed for new things. So we have a unique sort of access into the supply that others don't. And I think that's part of what has an advantage to us ultimately. In the future, and I'd say ultimately in the past 2, but really in the future, what helps us the most is how efficient we are. If you look at how many servers we have in our network, it's a fraction of what our competitors have. And we do similar amounts of work with a fraction of the surge. We're more performant than our major competitor all around the world. But also we do it with orders of magnitude less hardware. And that ultimately insulates our capital efficiency, which I would put up against any capital efficiency in the open market. Like it is astounding how much more efficient we are, even to us, but we also really focus on it. So I would say 2 things are helping us. The direct relationships that we've had and the fact that we are innovating and working innovative customers that helps on the supply side. The other one is how efficient we are. But these supply constraints are affecting everyone. I think the way we will see it affect us, and Ron has talked a little bit about our sort of guidance, our story arc around gross margin, is that ultimately, we are seeing some inflation in some of the underlying pricing elements. We were able to lock in 1 or in some cases, 18-month deals, but those will expire, and we are seeing some components go up in price. There's no question. So I do think we'll see an impact but significantly less when you're buying orders of magnitude less stuff to do the same amount of work.

Yes. Got it. Got it. All right. Ron, I've got a question for you that, fortunately, is not a entire gross margin question, but something a little bit hopefully more interesting, which is like you've been in the seat for a few months now. Maybe can you talk about what sort of learnings you can bring from your past experiences, right, at Fitbit, at Google, into Fastly and where you see that manifesting itself?

Yes. So I think there's a couple of things I've seen. I think one of the things I talked a little about earlier, which is kind of working with the head of sales in terms of just increasing that connection between the 2 of us and that visibility into where we see variability and where we see risk in our pipeline to drive better and better forecasting. I think that's 1 opportunity. I think secondly, I think there's an opportunity, and Joshua and I talked about this, is really taking a fresh look at, one, what are the metrics we're using internally that we're monitoring on a weekly and monthly basis that ensures we're on track for some of these goals that we've set out and that we've shared publicly to help us really monitor the business and make those sort of real-time adjustments that we need to make in the business. And I think ultimately, some of those metrics may make their way out to what we're sort of sharing publicly as we get into 2022 for those that are really meaningful in terms of where we are. So I think those are some of the 2 things that I think is one of the things we really need to be paying attention to, to drive execution. Because if I look at the opportunity out there, we talked about the growth, the robust growth of the market and the technology that we have. It's really about execution in terms of driving awareness at the computed edge, building out a sales team that is focused on the key focus of adding enterprise customers, efficiently adding customers in the SMB space. And so how do we make sure that we improve that execution? That's something we can influence. And to me, I think that's one of the biggest learnings is applying what I quoted in the past, to just drive sort of relentless and focused execution against those goals.

Got you. All right. Last question because we're coming up on time. Joshua, as we think about the long-term targets, one of the things you said you want a little bit more -- you're putting them out there. There's that level of accountability to public investors. How should we on the outside be gauging the success of these initiatives of the trajectory before it's too late, right? I mean the obvious answer is to wait until we get guidance for 2023. But by then, if it hasn't been working, that might be a little late. So help us understand what should we be looking for? How should we be gauging your success in progressing towards these goals?

Yes. So I think if you just look at -- fundamentally, our business has 2 engine. It has an engine that goes and capture new customers. It has an engine that takes those customers and grow them. If you look at our history, we have a world-class exceptional engine that keep customers and grow them, more churn, is near industry lows, continues to be our DBNER, and particularly, like last 12-month metrics, although as Ron explained, why they might tick up over a little -- a shorter period of time are world-class. So we have a world-class engine. So I think I would continue to focus on the things we have talked about, LTM NRR and the idea that we take these customers, those cohorts grow. I think that's super important. As Ron said, there's going to be some hiccups if some of that Q2 and Q3 story arc plays out, but that is, I think, a real focus. The other element that I would really -- that I think is super important is the number of net new customers and ultimately, enterprise customers because that's a trailing metric. So I would focus on both of those things. I think enterprise customers we'll see in the second half of the year start becoming more predictable as the work that Ron has done and Brett has done starts to really come to fruition around a discipline. But I do think those are the areas. New customer acquisition and the growth, those are the 2 areas that I think we report every quarter, and we will definitely see as the indicators of progress. Now we're also going to report against the goal. So it's not like we're going to be shy about saying where we are at, at the right points in our trajectory.

All right. Wonderful. I think that's a great place to hop off. Joshua, Ron, thank you so much for joining us. And thanks, everyone, for tuning in.

Thank you.

Thank you.