FSLY - Fastly Inc. (FSLY) Presents at Annual William Blair Growth Stock Conference (Transcript)
2023-06-06 19:13:06 ET
Fastly, Inc. (FSLY)
Annual William Blair Growth Stock Conference Call
June 06, 2023, 05:00 PM ET
Company Participants
Ron Kisling - Chief Financial Officer
Vern Essi - Head of Investor Relations
Conference Call Participants
Jonathan Ho - William Blair & Company
Presentation
Jonathan Ho
Hello, everyone, and thank you for joining us for our Growth Stock Conference in today's session with Fastly. My name is Jonathan Ho, and I'm the cybersecurity analyst for William Blair & Company.
Our speakers today are CFO of the company, Ron Kisling, and Vern Essi, who's the Head of Investor Relations.
Before we begin, I'm required to inform you that a complete list of disclosures is available at our website as well as conflicts of interest at www.williamblair.com.
With that, I'll hand it over to Ron just to -- maybe give us a little bit of an overview of Fastly and a little bit of a level set session. Thank you.
Ron Kisling
Thanks. So, Fastly -- at Fastly, we're really focused on the user experience, and we do that through our edge cloud platform that delivers content, security, as well as compute capabilities on a single platform to the user. And the platform automatically caches and purges data so that the right data is there so that users get low latency, fast performance, as well as personalized experiences from their compute capability, all to enable fast, safe, and engaging experiences, which is what companies want with their consumers is an engaging experience.
Our security portfolio includes the components that you would see for web application, a firewall -- our web application firewall, which we acquired from Signal Sciences, rated top product by Gartner. It's a leader in the market space. And its accuracy is so high that over 90% of the customers run the firewall in blocking mode. Our second component of this web application is DDoS protection, denial of service protection. We have great technology, and we're in the process of productizing that into a product this year. And then, the third component, which is bot detection. We are releasing a new version of that this year as well to fill out our security portfolio.
In September of 2022, Todd Nightingale joined as our new CEO. And his focus has really been on transforming and driving execution within the company. And then, one of his first areas was really looking at our go-to-market efforts. And in that, was looking at how we've organized the sales organization, and launching a simplified packaging to sell our products with less friction and open up more of the mid-market to our product.
We've also driven financial rigor in the company with the initial focus on our cost of revenues in terms of driving improvement in our gross margins. We started with building a very robust forecasting process to better align our investments in our network with the expected traffic, and that work is largely done.
The next step really is continuing to drive engineering efficiency in terms of how we use bandwidth and hardware, increase our use of peering. And the results in Q1 of 2023, we saw 300 basis point improvement in our gross margin. And we expect these efforts that we're putting in place now to drive continued accretion and improvement in our gross margin, which is a key part of our path to profitability.
And then, looking at OpEx, we've created financial rigor. We've established a purchasing function, and we've been eliminating a lot of duplicate SaaS that's come up over the course of the year. And the results of that have really driven a meaningful improvement in our bottom-line results. Our operating loss is improving from an 18% of revenue in 2022 to 10% of our revenue in 2023 at the midpoint of our guidance.
We'll be talking a lot more about our path to profitability and our long-term model in our Investor Day on June 22 at the New York Stock Exchange, and we look forward to seeing many of you there.
Question-and-Answer Session
Q - Jonathan Ho
Fantastic. So, Ron, Fastly has been able to deliver a lot of efficiency in both the operating and CapEx lines over the past year or so. Can you talk a little bit about where you found success in driving that leverage? And maybe, where you still see further opportunities?
Ron Kisling
Yes. I think the first thing I just touched on a little bit earlier was really putting in place a forecasting process that we didn't have in place before and using that as a guide to investing in line with the expected traffic. And that's largely in place, and so that drives an efficiency in the network all of its own.
The other efforts are really around engineering efforts where we're able to deploy as a software-first network, software across the fleet of servers to drive efficiency. We recently were able to increase our servers by 20% to 30% where we can run 20% or 30% more traffic through ours -- the same number of servers just with the software update. And there's opportunities to do the same thing around managing bandwidth more efficiently. And then, continuing to negotiate on our bandwidth rates, our hardware rates, as well as increasing our use of peering as we grow our business with low cost.
And so, those opportunities will play out over the next year or two, and we expect to continue to see accretion in gross margins. And I think as we get into 2024, you can kind of look to the margins we've seen historically back in 2020 in terms of what's possible in terms of gross margin.
Jonathan Ho
Fantastic. Fastly has also driven a lot of its growth from net expansion with only maybe a small number of net new enterprise customers as part of the model. Can you talk a little bit about the dynamic there between that new customer addition, as well as how you think about large customers and how they typically ramp with Fastly over time?
Ron Kisling
Yes. I think there's two motions. I think the expansion with existing customers has been something that we've done really well. And I think it speaks to the performance that we have. What we find a lot of times is, as we enter a customer, particularly amongst the large enterprise customers, they allocate traffic based on the performance levels. And since our performance latency, our efficiency of caching data is better, we typically gain share of traffic within a company once we've deployed there. We also see that happen over time.
So, if you look at kind of the growth metrics, if we sign a customer in the first year, the second year revenues might be 200% on average of what that first year was, going to 40% increase in the second year -- third year. So, you see really good expansion within those customers as we deploy and they see our performance levels.
One of the reasons we talked about some of the go-to-market efforts around package simplification and reorganizing the salesforce is to develop programs to increase our acquisition of new customers. So, as we increase the acquisition of new customers, and then apply this expansion capability that we've been so successful at is to accelerate our revenue growth by increasing the number of new customer acquisition.
Jonathan Ho
Makes a ton of sense. Fastly has historically been focused on some less commoditized areas of content delivery market. Are these use cases growing faster than the general market? And can you maybe talk about some of the drivers of that growth? And why is it that Fastly is differentiated around some of these commoditized areas?
Ron Kisling
I think one of the things that if you look at kind of the evolution of kind of the web experience, I think it's moved from just a delivering content to where an immersive experience, latency matters, personalization matters and driving customer engagement. And by delivering this higher performance, that, I think, plays into what companies are now looking for in the web to give them a competitive differentiation. And so, I think that's why in the space of -- versus sort of commoditized versus sort of value add area, I think we've done very well. And I think there's an opportunity, I think, to continue to grow that.
And I think if you look at our success, we had early success in industries where that mattered a lot. Commerce, where time to cart directly plays into cart abandonment and conversion ratios. Streaming, where reliable streaming is important, particularly if you're selling a premium service.
What we're starting to see is more and more companies realize the value of customer engagement. We're seeing new customers come to the edge in areas like travel and health care, where they're looking to drive better engagement through more immersive experiences at the web.
Jonathan Ho
Fantastic. You talked a little bit about sort of your acquisition of Signal Sciences as a next-generation WAF. And so far, the growth has been very strong there. And what's driving customers to select Signal Sciences over the competition? And are there any sort of emerging drivers we should think about in that WAF market?
Ron Kisling
It's a good question. As I mentioned earlier, the Signal Sciences WAF is a great product. Gartner ranks it number one. It is extremely accurate in terms of identifying a threat versus safe traffic. And as I said, most companies run it in blocking mode to actually stop threats. And so, that's why it's done extremely well. And what we've seen is we have deployments of our Signal Sciences products in customers that aren't running Fastly delivery. But it's a great opportunity because some of those customers over time can become Fastly delivery customers. So, we deployed it not only in our own environment, but other environments as well.
I think if you look at the overall portfolio at the web application level, I think, as I mentioned earlier, there's sort of three products that are sort of the core that you need at the edge. To elaborate a little bit more on DDoS, we have exceptionally good technology. It hasn't really been productized. We've deployed it [more in an] (ph) engineering method. And customers are really happy with where it is, and they've deployed in some really areas where they have particularly challenges in terms of denial of service attacks. So, we're productizing that this year.
And then, our bot detection, which is particularly important, as vendors really want their product to go to their consumers, not resellers. We're launching a new version of that product later this year to build out the complete security portfolio. And, again, focused really on web application, we're not focused on the IT shop, but the web application, our platform. And all of these run on our platform. We have a modern architecture where delivery, the security, and compute all run on one platform.
Jonathan Ho
Can you give us a little bit more detail as well on sort of the bundling and packaging that you talked about earlier? And how does this work with sort of traditional DDoS and other services that you're sort of bundling together into that single offering?
Ron Kisling
Yes. So, you're talking about the simplified packaging offerings.
Jonathan Ho
Yes.
Ron Kisling
So, we come from a history of really selling into the enterprise space. Our early customers were some of the largest customers. They're the ones where the differentiation and performance really mattered, and we gained traction there. Those customers like to be very specific in terms of what components they buy. They want to price each component separately. It's a consumption-based business. And so, those are very complicated transactions and billing arrangements. And that's generally how we build for our product.
We created quite a bit of friction with a lot of customers, and particularly as you get below that top customer, they just want to be able to buy a solution, and they wanted predictability. And so, what we are launching is a delivery package, which is everything you need to run delivery on our platform, a security package and a compute package. And they come in different sizes based on your volume, and it's a fixed amount each month. And it drives predictability, it includes everything you need. And so, it dramatically reduces the friction in the sales cycle. Every deal doesn't have to go through legal or our deal desk approval. And I think importantly, it also opens up the mid-market to take advantage of it, because the enterprise sales motion we had was really not well suited for those customers.
And so, this allows us to a product that I think will resonate with the mid-market. It's a product that can be sold to the channel, and using channel to augment our own direct sales efforts, and grow us into a segment of the market where we have not deeply penetrated to date.
Jonathan Ho
Absolutely. Traditionally, DDoS has been an area that maybe Fastly has lagged a little bit relative to the competition. Like, where do you see room to improve the product? And what does that ultimately mean for the financial model, if you can improve there?
Ron Kisling
Yes. I mean, I think if you look at kind of the DDoS space, Cloudflare actually comes from that, they were the DDoS player and then kind of moved into delivery. They have a great product. I think if you look at what we've done and I think if you go back, we have really great technology, and we actually have very good DDoS technology. We hadn't really put it into a package. I think as we package it, I think we will have a very competitive product in the DDoS space.
So, I think we'll go into the security space with, I think, a leading product in WAF, a very competitive product on DDoS, and a new product on bot, which is really the triad of the security for web application. And I think what it does is it means, given that we deploy on one platform, it makes it a lot easier to cross-sell, increase that expansion motion that we talked about earlier. In terms of deployment, we have very competitive and even leading product cum security that run on the same platform. By the end of the year, we'll run on one control plane. It's a reason to adopt those, and that reduces the friction, and I think will absolutely increase the upsell motion that we also see that's driving that expansion.
Jonathan Ho
Makes a ton of sense. Yes, I'm not sure if we can get through any presentation without the mention of AI. The obligatory question is, how should we think about the emergence of things like AI, large language models? And how does this impact things like edge compute, which suddenly are becoming more relevant?
Ron Kisling
I tend to think of it a lot like a lot of the other things in terms of edge, which is a lot of the content, a lot of the big compute, heavy compute is going to be done kind of at the central cloud. But there are certain things you need to do at the edge so that you're responsive is there. And so, I think inferential calculations on AI will be done at the edge. I think they'll be done at part of compute. So, I really see AI as it comes as being another application for compute at the edge.
And similar to where I think the overall web architecture of central data that you cache at the edge for the user experience, you're going to do something similar with AI, which is the heavy calculations will take place in the central cloud. You'll deliver, you'll do inferential updates or whatever at the edge for a quick responsiveness. But it is another application of compute. And I think there's a lot of applications for compute from personalization to gaming, all of which, I think, have an opportunity to contribute to -- compute being an important part of, what I call, a platform. It's not about content delivery, it's about being able to deliver an experience to the user, which includes secure, safe, engaging, not just delivering content.
Vern Essi
Yes. And it's also worth noting we've released ML APIs on our platform, like, two years ago. So, we're not, like, newcomers to the idea of AI on our platform, and we do engage with a lot of customers as well, all under NDA unfortunately, but we're working on it.
Jonathan Ho
Makes a ton of sense. In terms of the edge compute market in general, this has been a little bit slower to develop than we anticipated. How close are we to seeing maybe the greater monetization of the edge compute platforms? And, maybe what has to happen in order to make those use cases more of a reality?
Ron Kisling
Yes. I mean, I think two things. In terms of the timing, I think a lot of people started talking about it way ahead of time, so I'm not sure if it was delayed or we just talked about it too soon, I mean, broadly across the market. What I can say is I think we're seeing a lot of interest. A big driver is going to be identifying use cases for compute.
And I think one of the things that we found as we started talking to customers is one of the barriers to actually rolling out some of these use cases as they tested is we don't know what the response is going to be. Are we going to get 10,000 requests, or we going to get 10 million? That's a big difference in terms of billing. And so, when we started talking to customers about our package, all of them said, "Well, yes, let me sign up for this. I want to roll out some compute applications."
So, I think we're in the, right now, in terms of really looking at how many customers are starting to adopt it rather than counting revenue. And I think we're seeing a significant expanded interest in it. I think we're going to see that customer -- number of customers increase pretty dramatically over the next year. And then, I think, beyond that, we're going to start to see that become meaningful monetization, maybe 12, 18 months from now.
And the way we look at the business is the core piece of our business is delivery. Security is where we're seeing kind of the big growth engine. And then, we have incubation, and incubation includes the compute piece. And then, as that matures, it probably moves into the growth and becomes a meaningful revenue contributor.
Ultimately, when you look at whether it's the content responsiveness or compute, it's all about delivering an experience to the user that's safe.
Jonathan Ho
I think that makes a ton of sense. And, at this point, I think we can take questions from the audience. If anybody has anything that they'd like to ask, go ahead.
Unidentified Analyst
[indiscernible]
Jonathan Ho
Yes, repeat the question.
Ron Kisling
Yes. So, the question was, when you find yourself in a competitive situation with Cloudflare, and they bring their -- up their Workers platform, what's our competitive response?
I think two data points. I think, one, largely, we haven't run into them a lot. As I said, our sort of legacy history is really with the enterprise customers. I think they built really good traction in that mid-market and SMB market and where their Workers platform is. I think certainly, I think as we move into the mid-market with our packaging, I think we will see them more.
And I think what we would do is -- we talked about our compute platform. And our compute platform is written in Varnish, which has, I think, some benefits, particularly in terms of maybe more larger scale, [indiscernible] comes from the enterprise space. But I think if you're looking to do a robust application, I think we have a solution that would fit very well for those customers, and now we have a package that actually we can go to that market with.
Jonathan Ho
Other questions from the audience? Go ahead.
Unidentified Analyst
[indiscernible]
Ron Kisling
Yes. So, the question was, talking about the competitive environment beyond just Akamai and Cloudflare to, like, the Limelights or Edgios, and Lumens and some of the other players and how we're truly differentiated.
I think there's two or three things I would highlight in terms of differentiation. I think, first and foremost, is our performance. We have a modern architecture that software led, and so we have a meaningful performance issue in terms of latency. Our caching and purging is automatic based on content levels, and that really delivers a meaningful performance differentiation versus our competitors.
We also are very efficient with our hardware usage relative to the competition. And so, our hardware intensity is a lot lower. What that does is really allow us over time to be competitive with pricing and the cost structures of our competitors and still drive a meaningful gross margin. Now, we have some internal issues that we've been addressing, but ultimately, we can achieve the margins we talked about and be competitive on price.
So, those are those are two pieces. I think the third piece of it really is around -- and I talked about kind of the differentiation from an efficiency perspective. This -- we also run the entire compute, security and content on one platform. And so, one, that gives us the ability to optimize our content deliveries, heavy bandwidth usage, compute tends to be heavy CPU uses. And so, we use our network fully across that, but we're not use -- having to support and run multiple platforms to deliver all this. And for the user, that means ultimately one control plane. They can control all of this, so it also creates efficiency at the user side.
Jonathan Ho
Maybe just following up on that question around differentiation. Like, how do your customers determine which CDM player to use, which security product to use? Like, what does that evaluation process maybe look like? And what sort of stands out for Fastly relative to those competitors?
Ron Kisling
Yes. I mean, I think you're -- I believe a whole lot -- different lot of other things, I think you'll do a lot of proof of concepts. I think one of the benefits that we have when we do that is, for the core piece of sort of content delivery around their network, it's very easy for us to deploy that in a demo and demonstrate our performance capabilities. So, it's part of that proof of concept. We can very quickly demonstrate our low latency and our performance and the caching efficiency that we do. But that's really kind of the evaluation. They're looking at performance. They're looking at the cost of ownership. They're looking at your cache efficiency. On security, they're looking at, it's that firewall, the performance of the firewall, or the performance of the various ones, and they'll test it.
I think one of the things is once we get into a customer, some of our largest enterprise customers run multi CDN. And some of the most sophisticated ones really look at traffic allocation based on performance, whether they're measuring throughput or error rate or cache efficiency, and they'll move traffic around just almost automatically based on performance. And so, we like the multi-CDN environment, because our performance is one of the things that drives that expansion motion that you talked about earlier in terms of being able to gain additional shares of traffic once we're deployed in a customer.
Jonathan Ho
In terms of CapEx spending, I mean, how should we think about that as, I guess, an indicator around either future growth or as an opportunity for you to derive some of the cost savings and some of the cash savings?
Ron Kisling
Yes. I think, we mentioned earlier, we're very efficient from a hardware perspective. And I think if you look at, our outlook for this year is, our CapEx would be 6% to 8% of revenue. And that's down from, I think -- in 2021, it was, I think, 12% to 14% of revenue. Some of that was investment beyond traffic needs. But as we built an efficiency into it, we're looking at a 6% or 8%. I think that is a near-term sustainable level of CapEx investment, and that includes both hardware as well as the capitalization of internally-used software that we're required to do when we're developing particular features for our product.
Jonathan Ho
Excellent. Any other questions from the audience? I can ask a couple more if you want to keep going.
Ron Kisling
Go right ahead.
Jonathan Ho
Sure. So, when we think about sort of the competitive environment, can you talk a little bit about some of the consolidation trends that have been happening? And, in particular, when companies are choosing to go from maybe a multi provider to, like, a smaller number or vice versa, how do you think about sort of the trends there? How does Fastly get either benefit or face headwinds?
Ron Kisling
Yes. I mean, I think, we've started to see some vendors who are multi CDN in terms of three, four, five, six players consolidating down to maybe two. And they do want the redundancy of two vendors, particularly the largest enterprise players. I think their rationale is it's a lot easier internally to manage two vendors rather than six. And there's also the realization that if you're taking your traffic and splitting it to two vendors versus six, you're probably going to get a better price. And so, certainly, very large companies are doing that.
What we've really seen in that is, we actually like that motion, because a lot of times that decision is going to be based on who are the best performers and who have provided really great customer service. We have great customer service as evidenced by our better than 99% annual retention rate. So, we actually like that motion when they consolidate, because our performance usually means that we're going to be one of the two that they consolidate down to, which increases our traffic and share of the market.
Jonathan Ho
That makes sense. And maybe just one final question. When it comes to the ability to, I guess, take on more traffic or take on more sort of advanced use cases, when we look at that market, is there an expansion in terms of new opportunities over time for you to take on that traffic?
Ron Kisling
Yes. I mean, I think, if [indiscernible] capacity, I think, given the efficiencies that we continue to build into the network, I think we -- it's very easy for us to fairly quickly expand capacity. I would say, there is the overall capacity against our existing footprint. It's a little bit of a different dynamic if we're expanding capacity maybe in a market where we don't have a presence. That takes a little more planning. And when I talk about the 6% or 8%, I think over a 12-month period, we will be -- expect to be in that range. It will fluctuate from quarter to quarter. If we're rolling out a new site, particular quarter might be a little bit higher, might be a little bit lower. So, it is a little bit different dynamic if we're expanding capacity to a new site. If it's just overall capacity in the existing network, we have a lot of ability to sort of flex that network capacity from a bandwidth perspective, we've got a lot of capacity in our hardware capabilities.
Jonathan Ho
Excellent. Well, thank you so much for doing the presentation today and fireside chat, and we'll move to Jenny for the Q&A session. Thank you.
Ron Kisling
Okay. Thank you.
For further details see:
Fastly, Inc. (FSLY) Presents at Annual William Blair Growth Stock Conference (Transcript)