Cloudflare, Inc. (NET) Earnings Call Transcript & Summary

Okay. Welcome back, everybody. I'm Brad Zelnick, software analyst with Crédit Suisse, here on day 4 of our 34th Annual Tech Conference, typically in Scottsdale, Arizona, hopefully, where we'll be next year. But today, streaming live right into your office, right into your living room, right wherever you are. For this particular session, truly delighted to be joined by Thomas Seifert, who I call friend; and as well is CFO of Cloudflare. Thomas, thanks so much for joining us.

Always a pleasure. Thank you for having us.

For sure. And I think on video as well, we have Jayson Noland, who is head of all things Investor Relations. Jayson, thank you as well. Format of this presentation is meant to be a fireside chat. So we're going to talk about all sorts of fun things, and I've got a bunch of topics to go over with Thomas. But as well, I will keep my eyes on my inbox. If somebody has a topic they'd want me to interject into the conversation, I will try to weave that in. But maybe just to jump right into things. Thomas, Cloudflare is truly a unique business. Before we get any further, can you just maybe level-set folks with a brief background on your mission, products and the market opportunity that you see?

Yes, I'll gladly do that. So Cloudflare was formed on this paradigm shift that the world was moving from on-premise hardware and software that you buy to services in the cloud that you rent. And as you watch that packaged software turned into SaaS applications and physical service migrated into the public cloud, it was clear to the founders that it was only a matter of time before the same thing would happen to network appliances and network firewalls and network optimizers and load balancers and routers. And the myriad of other hardware appliances that are on-premise would turn -- inevitably turn into a service that is provided from the cloud. And we built that platform that allowed us to operate a large network and provide these services of these networks to our customer. And with that, you'll make anything that wants to connect to the Internet, more secure, more reliable and fast.

Helpful context. And maybe just to dive a little bit deeper. As it relates to that mission, I'd be curious to understand the significant spike in traffic once COVID hit. And I think on your first quarter earnings call around that time, Matthew had said that Cloudflare had seen as much growth in traffic on the platform in 12 weeks as you saw in the prior 12 months. So you -- I mean, I guess, maybe the question in that is, obviously, a lot of opportunity to help customers. But how were you able to handle traffic growth of that magnitude? And how did the platform perform?

Yes. I mean this truly goes back to the core of the Cloudflare business model and what is in the DNA. In order to provide those services that we just talked run about, load balancing, firewalling, routing from the cloud, we needed to come up with a rather unique architecture. So the first way to cope with that is how the business model is built and it's off-the-shelf hardware. And on this hardware that runs now in more than 200 cities and more than 100 countries, runs a fully integrated software stack of all the products we just talked about on every server in every country, in every location. And with that, the complete network surface becomes your decrees of freedom to handle and manipulate demand and supply. There's always some part of the world asleep where you have underutilization in your network, and you use that in order to mitigate that. The architecture and the network is also built in a way that, remember, we need -- if you want to disrupt the hardware-based competitors, you need to be able to price accordingly. You find -- you need to find a pricing model that allows you to price as a subscription, your Palo Alto firewall is not going to send you a higher bill at the end of the quarter just because traffic was up. So we needed to be able to offer those subscription-based services fixed and not with a lot of utility or utilization variables in there. And the network had to be designed to digest increases in traffic spikes because of that. So we peer -- we have peered now with almost 9,000 networks globally. And how we are peered and how we are collocated really allowed us to turn a lot of what other competitors have as variable cost into fixed costs, and that's why you have this unique flexibility to digest really significant spikes in demand, literally without any impact in our gross margin. I mean we stayed in our target range of 75% to 77% despite the spike in traffic. And our network CapEx stayed right where it was supposed to be around 11% in the third quarter despite the increase in the traffic.

Very impressive to see how the business model and architectural design is able to match the needs of customers even in that type of anomalous spike scenario. And I want to ask about the predictable pricing. But before I even get there, if I just look at the health of the business over the last few quarters. You were able to maintain greater than 40% growth, and growth actually accelerated this past quarter to over 50%. And as we think about those dynamics, can you walk us through where the strength came from, what you saw from a pipeline perspective?

Yes. The third quarter was a unique quarter and everything that could go well did go well and fell in our direction, so that was a truly remarkable quarter. But the strength was driven across a couple of categories. The one category, for sure, is the growth we see in our -- not only in our enterprise business but within the enterprise business and with our large customers. So the business was originally founded as a freemium and pay-as-you-go business, where people gave us a credit card. We -- the enterprise business only just started to develop in 2016, but it is today already more than 80% of our revenue. And within that segment, our revenue with our large customers has strong growth. Large customers are customers for us that do more than $100,000 of business with us per year. And that category grew now consistently above 60% every quarter since the first quarter of '18 and is now almost 50% of revenue as of end of third quarter. So within -- one very strong reason for the process is the growth in the large customer cohort. We've also done well from a global footprint perspective, very strong business in Europe, almost 58% in the second quarter and north of 50% in the third quarter. So we are increasing our global footprint, opening offices in Paris 2 weeks ago and Sydney and Tokyo and really trying to follow the large customers with more localized support. And the third topic is, I think, that is driving growth is we have been consistently been able to enlarge the TAM we address. In the beginning, it was really the core of network control as a service, load balancing, firewalling, routing. We started with Cloudflare for Teams, that is VPN replacement, gateway, browsing as a service, DLP and added another concentric circle and the TAM enlargement, for sure, that is the third vector that is helping us to grow.

Thank you for that. And maybe just to go back to the subscription-based pricing because somebody who doesn't understand the model well enough, seeing that type of acceleration and understanding what it is that you do for customers, it would seem intuitive that it was a function of some surge in increased traffic. I guess maybe from a go-to-market perspective, how important is it to price that way? And I mean, is there ever a time or a reason. Are you fully committed to that for all products? Aren't some of your products still tied to consumption? Or do I have that wrong?

So today, less than 5% of our revenue is usage-based. And that was important for a couple of reasons. I mean most of our money comes from disrupting on-premise hardware solutions. And as I said before, your firewall or your load balancer that is on-premise and hardware-based is not sending you a higher bill at the end of the month just because traffic went up. And that is what we wanted to compete against successfully, so we had to price accordingly. We also found it very helpful to disrupt other business models. A big product, a very successful product of ours is DDoS protection. And there are a lot of DDoS protection products out there that charge you by the amount of attack traffic you see. And there are some competitors out there that throw you off their network if the traffic becomes so high. And we thought with the advantages we have, we can literally un-meter that product. So if you're an enterprise customer at Cloudflare, you get DDoS protection with that, and it doesn't really matter how big the attack is. So it was first a means to disrupt and compete with on-premise hardware solutions. And then it also turned into advantage, into usage-based pricing competitors and then be successful against those. We are now entering a world of metered products. They're not really usage-based in the sense of bandwidth consumption, but licenses you take. So when we talk about Cloudflare for Teams, the VPN replacement product, browsing as a service, that will be a seat-based product, where customers pay dollar per employee per month. So it will have a metered component, but not in terms of using bandwidth as a usage-based concept, we will shy away from that.

And Thomas, as you think about the portfolio and you think about future products, I mean, even as we think about -- and we think about Workers, and I'm thinking also along the lines of other offerings, server-less type offerings that are out there in the market. Is there ever a time that you can envision that you won't be able to smooth out the consumption with the supply that you have to where you're forced to then move more into metered-based pricing? Because I would think that Workers is an example, maybe you could speak to this, that it would get to a point where if you had a customer overly consume, there's a hard cost associated with that, no?

Well, Workers is going to be a product that is going to be charged on request. So the amount of compute requests you have at the edge of the network is driving that pricing. With that said, it's less about bandwidth and traffic flow. It's literally how much compute capacity do you have at the edge of your network in order to accommodate a product like that. And here is another very unique advantage of our business model because we were never designed to be a CDN. We were always designed to deliver a product and security and performance products. The amount of CPU capacity that we have at the edge of our network is rather unique and higher than everybody else. So even there, we don't see an impact. And then Workers in terms of its architecture, so now we go really down into the weeds, this isolate architecture we have makes for a very low overhead in terms of processing power and cost to try. So we are -- we are able to offer Workers a product that allows you to be deployed at COVID, the edge of our network at very competitive prices, and they are still highly margin accretive to their Cloudflare business model.

Got it. Maybe just pivoting back to the core opportunity. I think you talked about 3 sets of competition, legacy or the hardware players, point solutions, and then you also have other cloud providers or central cloud providers as well. Is the majority of growth attacking a displacement opportunity? How should we think about that mix?

So today, the majority is still coming from the hardware-based replacement opportunity. And this is another great example where the business model really works well. We are not a rip and replace business model, right? We are -- we win what we call this war of attrition. So you take our load balancer, you put it in front of your on-premise in load balancing infrastructure. And then over time, instead of upgrading and putting additional on-premise capacity in place, budget dollars migrate to Cloudflare and your hardware-based environment disappears or it becomes a significantly smaller footprint. That is where the majority of our revenue and business opportunity is still coming from. And then I would say it's point solution providers, the Akamais, the Fastlys, the Limelights, so more and more Zscaler is falling in that category, too. Companies that have a similar business model and network, and they offer services from a network. We see each other more often now, but I would still say the prime opportunity is on-premise disruption. And then there are the cloud providers, as you rightfully pointed out. And if you were to ask Matthew, what keeps you up at night, he would say there's one name. It's Amazon, but we don't really worry much about the other 2 categories.

I think, Matthew, in that regard would be in good company with every CEO and every industry at this point, right?

I am [indiscernible].

Maybe just to expand a little bit more on the competitive point. Investors, I think, often approach the world through pattern matching. They like to put you into a bucket, and whether you're in the CDN bucket, the cloud security bucket. But in reality, I guess you don't necessarily fit neatly in any bucket. But to that point, on the disruptor side, who would you say, if any 1 or 2 players you most run into?

I mean if you accept the premise that loan for money comes from disrupting on-premise, then these are the guys that you run into, the Ciscos and the F5s, and the Riverbeds and the Palo Alto Networks and the Check Points of this world. On the point solution side, there's Akamai, which we admire. There is now sometimes more Zscaler more and more often. And on the cloud provider side, it's for sure, Amazon.

Got it. And maybe if I can turn back, Thomas, to what you talked about as helping drive some of the recent success, specifically in enterprise. Correct me if I'm wrong, but I think as of Q3, you had over 730 customers spending over $100,000 in annual run rate with you. Can you break apart some of that success? Is it coming from larger lands with multiple products? Or is it existing customer expansion? Or how should we think about the drivers of that? And how do you continue that momentum from a go-to-market perspective, from how you productize in serving that market or even partner in some ways if that makes sense?

Yes. I would say if you're -- rather consistently across the last quarters, it is about 50-50 in terms of expand and land, getting the growth in large customer count. And the land is interesting because we come out of this freemium and pay-as-you-go world. So the products are really efficient to get installed. So if you, for example, like come to us and say, I want to put my family page or my blog post behind Cloudflare, that takes you 5 minutes. But if Credit Suisse comes to us and says we're under heavy DDoS attack, that onboarding takes a couple of hours. And we don't have to send any Cloudflare employees to your site, it happens very fast and efficient. So we land really fast. And then once you're on our network, we have tremendous visibility into what hits your website. We can tell you where attacks come from. That -- I don't know, 40% of the traffic that hits your website might be malicious parts, and if you add a bot mitigation product, we not only help you save cost, but you also have a better security posture. And literally, every additional feature and product is [ a most take way ], whether you are Brad Zelnick or whether you're Credit Suisse. So this allows and drives very efficient expansion. And as we have increased our product suite now, and we talk about Cloudflare for Teams, we launched a product that's Cloudflare One that really bundles all the offerings we have in a rather unique way. We also enable ourselves to go after significantly larger customers from a network infrastructure footprint perspective. So the lands are becoming bigger, too. So -- and we've been able to drive both vectors, expansion and landing larger customers in a rather continuous way.

Maybe if we could turn to Cloudflare for Teams that you mentioned earlier. I believe you introduced the product earlier in the year, and you made it available for free for trial up through September 1. Is that right?

Yes.

And the conversion, from what I last heard on the Q3 call, has been tremendous. Can you maybe walk us through, Thomas, what is exactly the offering and the value prop? And what's driven the conversion success and how you'll continue to grow and monetize this opportunity going forward?

Yes. So Cloudflare for Teams is literally 4 products. It is what we call Cloudflare Access, that is their VPN replacement product. So Zscaler would call it the zero trust product, it's a VPN replacement product, and that was launched first in January, and that is what we gave away for free. I'll come back to that in a second. And then there are 3 more products to follow it later this year. The first one was gateway. The second -- the third one -- second one that followed is our browser isolation product where you offer browsing as a service. And it will be followed up by a DLP product in the first quarter of next year. So Cloudflare for Access VPN replacement product, we launched it in January, then COVID hit, and we thought we wanted to offer it, especially to small and medium-sized companies who struggle to deal with their remote workforce. And then we increased that offering to any customer who wanted to come, in some areas by government asked. If you need to help out the infrastructure, you should offer that to a broader base, and we did. So that product -- that offering run -- ran until end of September. And then we started conversion to getting paid for the services. And we were, however, very cautious. So the last thing we wanted to do was showing up on customers' footsteps on first of October, saying, here's the contract. So we were very, hopefully, very diligent and prudent. There are still companies out there that suffer tremendously under the pandemic, and we -- so we didn't push hard. We also knew that there are 3 more products to come, and there's a bigger opportunity for us from a bundling perspective if we don't push too hard on the first product, but rather wait for an opportunity to talk about a bigger offering to those customers. So we try to be very mindful of how we manage this transition. And despite of that, we had a higher than 78% conversion rate. But to be honest, there is -- there are lots of cats and dogs in that conversion. We have very large customers, large airlines, large pharmaceutical companies. The largest deal was, I think, 40,000 seats for one company, but you had also a lot of small companies in there, and we are proud of those 2.

Sounds like it's being very well received. But we've come across it just in our work on some other companies that have similar things in the market, and it sounds like people are very pleased with it. Maybe if we can turn to Cloudflare One, which seems to be a different way of packaging some already existing solutions. Why do that?

Well, first of all, I think the corporate networks have evolved. And COVID and the pandemic have accelerated that conversion in terms of how do you protect data centers, how do you protect offices, how do you protect now users that are, and employees, that are sitting out there in remote places. And managing this hybrid world has become a huge issue, and you have lots of idle capacity, all of a sudden, of traditional network infrastructure that is obsolete because of all the changes that happened. So Cloudflare One is an offering that is the result of literally hundreds of customer discussions, of what they need in order to accelerate their transition to this cloud-based world that we all move towards. And we found it as a unique opportunity to evolve our business model. Their -- a lot of the companies that we admire. There's Microsoft, there's Salesforce. A big part of their success is, how they evolve their bundled offerings over time. And Cloudflare One is that approach for us to find a new way how we bundle, how we price because Cloudflare One will be a seat-based offering like Cloudflare for Teams. And it goes well with what we talked about before. It allows us now to go after customers and land at customers with a larger ACV out of the gate. It also adds, over time we think significant opportunity to increase our TAM because now you can obsolete for our customers all the very expensive MPLS links they need in order to run this infrastructure. And one customer said that now, the Internet is almost good enough for running your network. And if you combine that with our services and products, you might just get there.

Yes. Thomas, forgive me, I've got so many topics that I want to touch on, and we're going to run out of time in just a few minutes. And the one, maybe I'll go to, and we should have started with was edge computing, Workers, which I remember, you and I chatting, it's probably a good over 3 years ago when you were newer to Cloudflare, and you were telling me all about how exciting the opportunity was. And I think back then, you described Workers as, this is the optionality that people don't realize, but this is where the world is really going. And since then, earlier this year, you introduced Workers Unbound which is a server-less platform. I've got many questions. I guess everybody is trying to understand the size and scope of this overall opportunity. I have my own view. I think a lot of people struggle to put numbers to it. I don't know that it's necessary. If we can all agree that when we think about application architecture going forward, we know that there's a lot that happens that's not going to all happen in the central cloud, and this needs to migrate to the edge to meet the demands in terms of data volumes and concurrency and security and performance and everything that people need out at the edge. So we can just say, I think, it's big. We can probably agree that it's big. But in terms of differentiation and how many winners there's going to be for this opportunity, I'd be curious to hear your thoughts on how this all comes together, and why I shouldn't be really excited because I am.

Yes. So we are shying away from even sizing it. It gets big rather fast, and then you tend to become not credible when you throw numbers around that are this large. But we think it's a big opportunity. Workers is the platform that all of our products are built on, right? So it's a way for our customers now to deploy code at the edge of our network. It's also a way for our customers to take our services and products and make them the most customized, load balancers, firewalls, whatever is out there, that's why the attach rate is so high. We attach it now to north of 20% of all enterprise contracts we write. But we also say this is, for us, what we call a wave 3 product, and it's priced for adoption. We wanted to be in as many hands as possible, as many developers, not only looking at it, but really writing code and deploying code at the edge of our network. And on the Q3 earnings calls, we gave some numbers how the number of developers really writing code is exploding. We think we are unique from a variety of reasons. And especially when we look at competitive offerings. And to be honest, the only serious offering that is out there, in our opinion, is Lambda from Amazon. And we developed -- when we talk about what developers need when it comes to edge computing, we talk about this hierarchy of needs. And what we learned over the last 1.5 years is that we always thought it's all about speed. And we have really interesting use cases about speed where people take artificial intelligence and push it out to the edge of the network or offload bill of materials from hardware devices, from drones or self-driving cars into our network. But the bigger opportunity now seems to be compliance. And in a world and an area of data sovereignty and privacy where countries and regions of the world make it really hard to move data, enabling that seems to be one of the biggest opportunities. And this is where we make the biggest strides in terms of featuring out the product. So our customers can use it as a tool to be compliant and efficient in how they run data. And there's one feature we launched called durable objects, where you can flag and describe data where it should be stored, how it can move, give complete transparency when it moves. And this has huge uptake for people redesigning shopping carts when you start to input a private credit card data, chat functions, color variation tools, people sharing the same document across multiple data sovereignty regions. So compliance is really important, and we are really unique there in differentiating ourselves from competitors. Then it's about ease of use and what it takes to work with the product. And I think that's another area where we are super unique because we go to the developer. You don't have to learn a language to work with us, whether it's Python or COBOL or anything in between, it's enabling you to use Workers. Because the architecture is so simple, the costs that we have, we talked about it before, are very low. So we can offer it to significantly discounted to whatever competitive offerings are out there and still make it a margin-accretive product for us. And then consistency and speed is really where we excel. Because the architecture is so simple, it's an isolate-based architecture, our spin-up times are, I think, 5,000x faster than Lambda. So you excel in all what we call the 5 categories or 5 needs of developers in this edge computing space. So we are thrilled by the product we have. And as I said, it's not about revenue. It's really about adoption and making sure we get it into as many hands as possible.

Very exciting. With that, unfortunately, we're out of time. I congratulate you and Matthew and the entire team on all the success to date. I hope that you keep it up. Jayson, as always, thank you. Thomas, it's always great to see you. And hopefully, soon enough, it will be in person and safe when the world is an even better place. So thank you for Cloudflare for keeping the Internet up and running and keeping us all connected.

All right. Thank you. Proud of it. Thank you.

Have a good day, guys.

Thank you.