Good evening. My name is Rob, and I will be your conference operator today. At this time, I would like to welcome everyone to the Fastly third quarter 2019 earnings conference call. All lines have been placed on mute to prevent any background noise. After the speaker's remarks, there will be a question and answer session. If you would like to ask a question during this time, simply press star followed by the number 1 on your telephone keypad. If you would like to withdraw your question, press the pound key. Thank you. I would now like to turn the conference over to Maria Lukens, Vice President of Investor Relations. Please go ahead.
Hi, everyone. Thank you for joining our third quarter 2019 earnings call. We have Fastly CEO, Artur Bergman, President Joshua Bixby, and CFO Adriel Lares with us today. Before they start, I want to remind everyone about the format of our call. We published a shareholder letter on our investor relations website and with the SEC about an hour ago. We hope everyone had a chance to read it. Since the letter provides a lot of details, we'll begin the call with brief remarks before opening the call for your questions. During this call, we will make forward-looking statements, including statements related to the expected performance of our business, future financial results, strategy, long-term growth, and overall future prospects. These statements are subject to known and unknown risks, uncertainties, and assumptions that could cause actual results to differ materially from those projected or implied during the call.
Please take a look at our filings with the SEC and our Q3 2019 shareholder letter for a discussion of the factors that could cause poor results to differ. Note that the forward-looking statements on this call are based on information available to us as of today's date. We disclaim any obligation to update any forward-looking statements except as required by law. During this call, we will discuss certain non-GAAP financial measures. Reconciliations to the most directly comparable GAAP financial measures are provided in the shareholder letter on our investor relations website. These non-GAAP measures are not intended to be a substitute for our GAAP results. This call is being webcast and will be archived on our website shortly after the call. With that, I'll turn the call over to Artur.
Thanks, Maria. Welcome, everyone. We're deeply appreciative for all of you joining us today to discuss the 3rd quarter results. We're making great progress. We're continuing to execute the strategy that we laid out when we went public earlier this year. We provide a scalable edge cloud platform that allows developers to build agile, secure applications all around the world. We provide control, real-time visibility so they can grow their businesses. Today, we're excited about our future as we were 6 months ago at the IPO, and we're really pleased with our results for the 3rd quarter. As you can see in our shareholder letter, we had a strong quarter, generated $50 million in revenue, up 35% year-over-year. We increased our enterprise customer count to 274 from 262 in the previous quarter. We expanded DBNER to 135% and increased the average enterprise customer spend to $575,000 from $556,000.
These results reflect an increased adoption of our edge cloud platform by both new and existing enterprise customers across the globe. We continue and will continue to identify opportunities to drive operating leverage and expect to maintain our path towards profitability in the years to come as our network scales. Developers continue to use our platform, and a big reason for that is that we deliver new tools and functionality that enable developers to build and create beyond what was previously possible. We're starting to see results from last quarter's launch of our developer library, which includes ready-to-deploy code and solution patterns that makes it faster and safer for developers to discover, test, customize, and deploy edge cloud solutions.
In the third quarter, we've launched the newest version of our HashiCorp Terraform provider, the highly requested provider from our customers, so they can integrate more tightly with their operations and deployment systems. We also announced the expansion of our real-time log streaming to support Apache Kafka and Elasticsearch endpoints, giving developers and customers even more options to experiment and discover big data. Finally, as you may have seen yesterday, we launched our powerful new language agnostic compute environment, Compute@Edge. This new product in beta is a major milestone. It's the next evolution of our edge computing capabilities. Purpose is to empower developers to build far more advanced edge applications, more secure, more robust logic, and a new level of performance. It offers developers our best edge computing technology with instant global scale.
We notably added tooling to support Rust as a second programming language, in addition to the Varnish configuration language. We chose Rust due to its growing popularity with forward-thinking organizations. We also plan to add more languages in the future. We're still only a couple of quarters into our life as a public company, but we're excited with the progress we've made so far and we look forward to the opportunities ahead. We raised our full year guidance for 2019 to reflect the ongoing momentum in the business. Total revenue is now expected to be between $194 million and $198 million, compared to our prior outlook of $191 million to $195 million. Non-GAAP operating loss is expected to be between -$38 million and -$34 million compared to our previous outlook of $40 million or $34 million.
Before we move to Q&A, I want to thank our employees, partners, customers, investors for their continued support of Fastly. They're the backbone of our efforts. We're grateful for dedication to Fastly and our mission. With that, I'll turn back to the operator for the Q&A portion of the call.
In order to ask a question, you will need to press star one on your telephone. To withdraw your question, please press the pound key. Your first question comes from the line of Brad Zelnick from Credit Suisse. Your line is open.
Excellent. Thank you so much, and congrats on all the success. I've got two questions. The first, in this past quarter, two of your competitors highlighted record seasonal traffic growth. Can you talk about the trends you saw in the market in general and as well competitively? As we look to Q4 in 2020, can you talk about the impact of a potential change in the mix of traffic on your network from upcoming OTT launches and how that might impact the leverage we see on the gross margin line?
Sure. Thanks, Brad. It's Joshua here. I think on the seasonal traffic growth, I think it's important to remember that a lot of the legacy competitors are really mostly media plays. If you look at Fastly, we are an edge compute platform. We are an edge cloud. We have seen sustained growth, but that sustained growth is really in our core customer base, which is applications delivered differently. We obviously, and we've talked about this in the past, have video traffic, but our media business is really highlighting on the high margin segments in that market. Things that really drive very different business than sort of traditional VOD. I think we are seeing some of that, but I think it's really augmented in some of the legacy customers who rely only on the VOD segment to progress.
I think if you look at Q4 2020 and ongoing, one of the secrets to our success has always been really relying on this developer-friendly application driven mandate. I think Artur spoke about this, but our Compute@Edge launch, I think, only continues to further that as we look out in the future. I wouldn't anticipate any significant differences in that traffic. You're going to continue to see us drive. Most of our traffic is going to be security-based and application development-oriented traffic.
That makes perfect sense. Thank you for that. Maybe onto something even more exciting. I did want to ask a question just about Compute@Edge, and I appreciate it's still in beta, but it seems your differentiation in part will be the speed and low latency in executing a request. There's also a lot beneath the covers in terms of the technology. Perhaps, Artur, can you talk about the technological differentiation and perhaps defensibility of how you've solved the challenges of concurrency and isolation and multi-language support? How should we think about the monetization model there and the margins associated with deploying these type of workloads at the edge? Thank you.
Yes. Thank you. I think on the technology side, we set up a couple of requirements when we started this project a few years ago, which was really we wanted to be able to support many different languages, right? We know that tying customers or developers into one language is not optimal. We needed it to scale to do many thousands of requests per second on a given piece of hardware. Notably, we also wanted kind of a secure environment, not per customer, but per request, so that even if a customer's code would have bugs in it wouldn't potentially leak data between two different requests in that customer. We started working together with Mozilla and some other companies on the WebAssembly project, which was originally developed to allow multiple languages run closest native code speed as possible inside browsers.
This is attractive to us because browsers they have a lot of resources, and they are an area where they have to spend a lot of time and money on making sure it's secure. We took that, and we have developed some open source technologies to Lucet as a compiler for WebAssembly. We have then embedded that into our One Network, and we will allow customers to upload WebAssembly and deploy it using all the tools, logs, streaming, all the things that make Fastly great. Now they can write much more complex applications in their language of choice. Initially, we're supporting Rust, but since it's WebAssembly underneath, once we are comfortable with that and have that developed, we will expand the tool chain to support other languages. Reception from customers that we've been talking to, both before and after the announcement, have been great.
On the revenue side, the forecasts that we have given in the past have not included the additional edge compute revenue from this new version of our edge compute, the Compute@Edge. We aren't ready yet to give guidance on what that will do, but it's not been baked into the previous numbers.
Thank you so much for the color. Really appreciate it.
Thank you.
Your next question comes from the line of Will Power from Baird. Your line is open.
Okay, great. Thanks. Yeah. Congratulations on the results. I wonder if you could, I guess, peel back the results a little bit as you looked at the upside in Q3, maybe just help us understand where any positive surprises might have been and just maybe a little more color on the key drivers there. I guess, really the same thing for the 2019 guidance. What's giving you that confidence those trends continue into Q4?
Sure, Will, this is Adriel. I think in general, what you can see in the increase in DBNER from Q2 to Q3, in general, what you're seeing is, even though we believe over time DBNER could compress over time, just given sort of law of large numbers, what we saw in the quarter was continued usage, expanded usage of the edge capabilities that we have amongst all of our customers. The fact that the trends related to our enterprise customers remain positive. We added a few more enterprise customers, more importantly, you'll also have reflected in terms of the average spend of enterprise customer that went up to what it is in Q3. That trend seems to be continuing as we think about the end of this year.
That's historically borne out, and I think what you're seeing in our guidance, is that we see that in terms of what we have in our customers today. We look at our enterprise customers as sort of the main driver of that, and that seems to be continuing.
Okay. As you look at that Dollar-Based Net Expansion Rate, you're at a higher level than where you were in Q2. Can that tick up further in Q4? You're obviously at a pretty high level as it is right now.
It's tough to speculate how that will be in Q4. We do have new customers that are continuing to ramp up with us. We shall see how those continue to ramp up into Q4. I think at this time, it's difficult for me to speculate in terms of whether it'll be DBNER that's going to be growing continued in Q4, or whether new customers are going to be ramping as part of that. I'm trying to give you in the guidance is just that at least some combination of the two that pleaded me to be a little bit more bullish than I was for the full year than I was previously.
Okay, great. Thank you.
Thanks, Will.
Your next question comes from the line of Jonathan Ho from William Blair. Your line is open.
Hi, good afternoon. Congratulations on the strong results. I guess maybe starting out with some of the success that you've seen with the developer library. Can you talk a little bit about how that is maybe impacting your ability to reach customers as well as ease of adoption of your solution?
Sure, Jonathan, this is Joshua. That's a great question, I think a really exciting development. When we talked about the Edge Compute Platform, we spent a lot of time educating this audience about our reusable modules, about how we could have this real platform effect where you have this flywheel where customers come onto the platform, developers develop code, that code solves a problem, and then there's a way to bring on that next company that may be in the same vertical or may share some commonalities such that they can get to market faster, develop less code themselves, and we're really seeing that flywheel turn on the developer library. Artur talked about this. Giving people the ability to take these advanced solutions and simple configurations, I think what you're going to see is an acceleration of that revenue and the stickiness of the platform.
I think we're seeing that in the numbers. We're starting to see that in the numbers, and I think we'll see that play out in the future. I think what we're seeing is a real stickiness, a faster time to ramp for these customers. They're getting to value quicker. There are a lot of examples of that where we're just seeing significant traction. I think the really exciting part of the Compute@Edge launch, and obviously it's still early, is the acceleration of that. I think we're going to see more of these things come to light as we allow people a bit more flexibility. Very exciting flywheel, and we're seeing it in real effect right now.
Got it. Can you talk a little bit about the investments that you guys have made on the go-to-market side and Salesforce, and maybe what's been effective? What are you seeing opportunities to invest more in? I just want to get sort of an update in terms of where we stand there. Thank you.
Sure. It's Joshua again. We talked about this when we did the roadshow, which this is a muscle that we are building, particularly on the marketing side. We have traditionally not invested a lot in this. It's been much more word-of-mouth business that is wonderful business, but not a place that we've invested. We've started to invest on the marketing side. That's going from near zero, it does take time. You have to build a team. You've got to find smart ways to invest that money. The signs are very, very positive. We talked about that in the last call, and I think the signs are even more positive. We're seeing significant attribution to our marketing efforts.
I just got a picture of one of our billboards in Times Square, which everyone was very excited about, and a picture from a friend of a billboard on the highway. We're starting to sort of come out of our shell and tell that story, and those results are showing up. Remember, this is an enterprise sale, this is not something where we're going to see that effect instantly. The money's being spent, the metrics are moving in the right direction, and we're starting to see investment. I think I'll pass it to Adriel to sort of speak to the sales ramp side, right now, lots of momentum.
Thanks, Joshua, I guess, Jonathan, the only other thing to point out there is from a Q3 year-over-year, we did clearly increase spend on sales and marketing and even as a percentage. We're making that investment, but to Joshua's point, it is still early days, and we're still trying to figure out the actual to fine-tune the go-to-market motion. We still feel good about the ramping of our sort of end sales reps that we want to achieve near the end of the year, and we'll clearly give you a sense of where we are at the end of the year when we report in Q4. Overall, I'm pleased with where we are in that ramp and we're hoping to sort of get some bear fruit of that labor hopefully sometime next year.
Great. Thank you.
Your next question comes from the line of Walter Pritchard from Citi. Your line is open.
Hi. Thanks. Adriel Lares, for you, on the gross margin side, I think this quarter, things surprised a little bit, at least versus our model on the positive side and reversal from what you saw in Q2. Could you help us understand what the swing factors were? I think just as I look at, I know you're not guiding to it specifically, but as I look at consensus for Q4, it looks like the street's looking for a 200 basis points or so sequential increase in gross margin. Just curious if that's realistic and what the drivers are to get there.
I'll answer in reverse order. Last year we saw a similar increase in gross margin from Q3 to Q4, to the tune of what you just mentioned. I see no reason why we would be no different than that this year. Everything seems to be going that direction. Specifically, with respect to Q3, in terms of the gross margin progress we're making, I think in general there's still areas within our investments into support, labor allocations that were still investments we need to make. We will reach critical mass there, hopefully here in the near future. From a sort of an incremental bandwidth side, some of the other areas that we feel on a variable side, we are making good progress there, and I think you're beginning to see the beginning of that in Q3.
We hope to, as I've mentioned before in past calls, contribute leverage in gross margin on a year-over-year basis, and we expect to do that in 2019 over 2018. From a high level, that's kind of how I see it.
Great. Just relative to the industry, maybe for Joshua, on pricing, I know from an industry perspective you're embedding some compression in pricing as time goes on. I'm wondering what you saw in Q3 and how that compares to what it's been historically.
Yeah, we certainly haven't seen any difference in pricing. Again, I think it really depends on the market you're in, right? Because we're not in the low end of the commodity space, which we hear is comprising quite quickly, that's not the space we're in. When you think about price compression, obviously it's a direct function of how much value you add to customers. I think when you look at the DBNER and you look at the growth in our large customers, what that is speaking to is the significant value we provide. I feel like that price compression question is really, in many ways, a commodification question. Being the only real leader in the edge compute space, that gives us a huge leg up.
We made significant sort of bounds leaps, I would say, even farther in that with the launch of Compute@Edge. We're not only in the lead, we're increasing the lead, and that dramatically affects the pricing continuum.
Okay, great. Thank you.
Thank you.
Your next question comes from the line of Michael Turits from Raymond James. Your line is open.
Hey, guys. Congrats on the results. A couple questions. I would imagine, and granted, you're not in the commodified end of it, but my understanding or my expectation would be still that you may get some benefit in 4Q from the two prominent, new big OTT launches. How much is being built into for that? Even more important, how much variability is there? In other words, I think there's a general sense that we don't know how fast those will roll out. We don't know how much will go to one provider or another. Does that introduce more variability in this particular quarter?
Let me talk, Michael, this is Joshua here, on the OTT launches. As we said all along, we're not an events-based business. We don't build out for events, in the typical way that you'd see how people sort of throw a lot of CapEx and then wait a long time for it to get used. From our perspective, it doesn't really introduce variability. We've got a plan. It's lasted and worked for us for the last eight years, and we continue to work towards that plan. I think you'll see over the year, and Adriel can speak more to this, that we have a pretty good certainty of what our CapEx numbers are going to be. They're quarter to quarter, as we talked about last quarter and this quarter, there is some variability.
We work at the high margin side of that, and we don't see that introducing tremendous variability into the business. Adriel, I don't know if you want to add to that.
Yeah. The only other thing I'd add, Michael, is just in general, second half of the year and specifically in Q4, you do sort of see an acceleration in revenue across all of our business, not just what you, I think you're referring to as OTT, but you also have our e-commerce customers with the busy shopping season. A number of those do impact things. I think, what we're trying to endeavor to get better at every year is clearly from a forecasting standpoint. The range you see there is probably a bit wider than you would see when we're at, call it a $500 million run rate or several hundred million dollars north of where we are today. I think there's nothing, I think, more unusual today, than what we sort of saw last year.
Okay, thanks. You hired a CISO and security is part of your offering. My sense has always been that there's certainly a roadmap there and a lot more opportunity for development and things that you can do. Anything you can share with us about that roadmap at this point?
Yeah. I think we've been pretty clear that security and delivery is core to our business, and they're really equal players. If you look at our customers, they are buying security from us. It's sort of a given now. You don't go to one of these e-commerce or platform or high-tech companies and just offer part of the solution you're offering compute, you're offering delivery, you're offering security. I think that the web security market is changing rapidly, and it's a huge area of investment. You saw the announcement, which we're quite excited about, around significant new talent joining the team. I think what you will see in the roadmap is continued investment and significant investment in that area in 2020. This is a large growth area for us. Our customers are demanding more and more.
We are investing more R&D into this every year because the demands are pretty significant. If you think back to the story arc of the disintegration of the data center and all of these sort of antique software appliances that sit in data centers that are all going away, the reality is a large portion of those were security appliances. When you think about the real value that we provide, it's actually building out the ability for those functions to, one, modernize, and two, go out to the network. One of the areas that's extremely exciting about the Compute@Edge launch, and Wasm in particular, is the ability to allow developers, our customers, to innovate on the security side. I think you're going to see a couple areas of innovation and sped-up innovation because of the ability that we're bringing with the Compute@Edge.
I would anticipate more announcements on that over the coming year and a significant R&D investment in terms of moving dollars, not net new dollars, into that area.
Great, Joshua and Adriel. Thank you very much.
Thank you.
Your next question comes from the line of Tim Horan from Oppenheimer. Your line is open.
Thanks. Artur, can you maybe just help us understand why Compute@Edge is so important and why this is so revolutionary? Maybe just talk about some of the new use cases that you're discussing with your customers and, I guess, longer term, how large of a market can this really be? Thank you.
Sure. Thank you. I think the revolutionary aspect, if you look back at how the kind of serverless container market as a whole, and some of the function in the cloud, there's a lot of overhead with running large JavaScript interpreters that you have to either start up or keep running. They use a lot of memory. They use a lot more CPU. That might work when you are operating in large data centers in the middle of nowhere. When you want to deploy to the edge, you're in a more constrained environment. The fact that the overhead of WebAssembly, not only in the execution time but also in the spin-up time, the number we mentioned in our press release is around 35 microseconds. Even I do that mistake sometimes. Microseconds to spin up a container for that request is really critical. Same with memory efficiency.
It really takes this idea that write code, write it in the language that you are using in-house, that you have tooling for, but be able to deploy it in a very safe, secure, efficient manner as close to the user as possible and kind of handle that load globally. I think that's where the really amazing engineering work we have done shines through, is that kind of the efficiency of doing this. Not just doing it, because that's been possible in the past with large servers and virtualization, but just really doing it at scale in a very tight environment. On use cases, we've been hearing from customers for a long time around things like GraphQL. For example, like the ability to assemble API requests at the edge.
You might render a page and think of a page that's made out of a bunch of different modules. Some of those modules are shared by all users, some are not. With this, you can fetch the different modules from origin. Some of them might be cached, and you can assemble the page and stream it out to the customer as soon as things arrive. Some of it personalized, some of it not. You really can get a much better feel, like a response time on pages, on API calls, offload more on the cache because you can personalize. We're hearing a lot of that. We're also hearing more advanced use cases in IoT and mapping that customers really want to explore new things to do in this environment that they haven't been able to do before.
Thank you.
Your next question comes from the line of James Fish from Piper Jaffray. Your line is open.
Hey, guys. Congrats on the quarter. Sorry, jumping around here, so apologize if it's somewhat been asked. Artur, maybe for you, if we could start. How did the opening up of the WebAssembly this quarter go for kind of net new ads? Or did you see any demand as a result of that openness?
I can't share the exact numbers, but we're very happy with the demand of the open or the public beta that we're opening up. That demand is coming both from existing customers as well as divisions inside our customers that weren't customers, as well as brand new customers that are really interested in the technology. The initial, we announced it yesterday, but the initial response is we're really happy with it. It's really exciting. Some of these customers really are very advanced technically, and so we feel it's always exciting when you have customers that really can use the technology you provide and also help guide us in what's most important in the short term to deliver to them and make them successful.
Got it. Then, Adriel Lares maybe, or Josh, could you guys give us a sense as to what has evolved in terms of the talks with the major OTT providers in terms of coming share of the new services and if you guys think you need to continue to add more capacity to the network? Just trying to get a sense of how much capacity of the network you're even using today. That's it for me. Thanks.
Thanks, Jim. I'll take that one. It's Adriel Lares. I think it's something that we continue to try to keep at just from a safety standpoint. That drives how we manage our CapEx purchases and in general bandwidth purchases for capacity in the network overall. We're constantly having these conversations with our current and prospective customers in terms of what we think we may win in terms of new business and new usage. That stuff is an ongoing, this is a week to week, month to month look forward. What I try to build into our guidance when we give out is effectively that.
Which is, hey, here's what we think that we see out there from our current business as well as our prospective business is beginning to ramp, or excuse me, that we believe will ramp. As it relates to on an annual basis, when we think of CapEx, that you see on the cash flow statement as a percentage of revenue. Eventually long term, we believe that'll get to around 10% of revenue. This year we think we'll be similar to where we were last year as we begin to ramp our CapEx and bandwidth purchases. That feels about right and that's all incorporated into that guidance. I know it's a bit of a long-winded answer, but I was trying to give you just a bit of a sense of how we think about these things.
Yeah, makes sense. Thank you.
Thanks.
Your next question comes from the line of Jeff Van Rhee from Craig-Hallum Capital Group. Your line is open.
Hey, guys. This is Rudy on for Jeff. Couple for me. One, the DBNER up nicely this quarter, second straight quarter. I was curious, you guys gave this cohort analysis in the S-1. I was curious, if you look at the 2018 customers now in their second year, is that sort of the cohort that growth that you guys were seeing, almost $3 in year two for every dollar in year one, four in year three for every dollar in year one. Is that still trending sort of in line or what's changed there?
I think the short answer is yes. It's still trending sort of as you'd seen historically. I think what you saw in the S1 with the 2018 quarter in general, they were growing actually a little bit faster. I think that's what you see sort of impacting our current DBNER are those 2018 customers continuing to follow the trends of previous cohorts.
Got it. Secondly, I think you guys touched on this, that you're on track for year-end sales hiring goal. I think you commented last quarter that you were looking to the end of the year at 60 quota-bearing reps. Is that still the goal? Can you share where you guys are at now?
That's still our goal. We're sort of running fast and furious trying to keep up with that goal, and we'll share where we are in our next quarter in terms of where we end up.
Okay. Got it. Great. Thanks.
Thank you.
Thanks, Rudy.
Your next question comes from the line of Jeff Kaplan from Stifel. Your line is open.
Hey, guys. Thanks for taking the question. Just a quick one on the enterprise customer growth. Just the absolute number of net adds was down a little this quarter relative to the last two. Just curious if you have any additional color on that. If it's just a reflection of typical seasonality and just trying to figure out how to think about the growth of enterprise customers going forward on a quarterly basis.
Yeah, I think a lot of these metrics, just given the size, I think you have to look at them over a LTM basis rather than sort of a quarter-to-quarter basis. We're really happy with the continued trend of customers growing. In fact, we're seeing actually an acceleration of customers ramping into that enterprise class faster than historically. I wouldn't read too much into that. I think you have to look at this really over a longer period of time. Seasonality plays into it to some extent, but I think overall that's a healthy growth and you'll continue to see a healthy growth, but there will be some variability.
Got it. Thanks.
No problem. Thanks for the question.
There are no further questions at this time. Mr. Adriel Lares, I turn the call back over to you for some closing remarks.
Thanks everybody for joining us on our Q3 call. We shall see hopefully many of you on the investor relations circuit. I believe next week we will be at the Needham Conference in New York, so we hope to see some of you there. Thanks very much.
Ladies and gentlemen, this concludes.
Goodbye
today's conference call. Thank you for participating. You may now disconnect.