Excellent. Thank you everyone for joining us. My name is Keith Weiss. I run the U.S. Software Equity Research franchise here at Morgan Stanley. Have our European software analysts in the front row, so I have to qualify that to just the U.S. or I'll get them mad at me. Very pleased to have with us from Cloudflare, CFO Thomas Seifert. Thomas, thank you for joining us.
Thanks for having us.
Excellent. Before we get started, a brief disclosure. For important disclosures, please see the Morgan Stanley Resource Disclosure website at www.morganstanley.com/resourcedisclosures. Excellent. With that out of the way, Thomas, again, thank you for joining us. I think maybe just to sort of set the bar, macro has been a kind of top concern for investors for the better part of the last, let's say, two years. Could you talk to us a little bit about what you're seeing in the environment today and how Cloudflare has been navigating it, sort of at best a volatile macro backdrop?
Yeah. Boy, it's gonna keep us busy, I guess, for a couple of more quarters. You know, we saw what we think was a downturn early. We talked about it early, in the Q4 of not last year, the previous year, and we adjusted early in terms of spend and hiring and focus on cash flow. Over the course of last year and into the Q3 , we saw the fundamentals actually starting to improve again. You know, we talked about it on our earnings call. Pipeline in Q3 improved, was even better in Q4. It's pretty stable in Q1. There are inflection points in the business.
However, we see that transitioning from pipeline into ACV into business is still driven by lots of uncertainty. You know, the behavior we put into Cloudflare, you know, we lowered spending limits. I see more purchase orders at lower levels going out and it's reflected in customers too. Sales cycles are moving out. What we also see is that, you know, we have probably more larger deals in the pipeline than ever before, but they take significantly longer. The impact on how does an improvement in fundamentals in better pipeline translate into ACV and then ultimately in revenue is still attached with lots of uncertainty, and that has not improved at all, in my opinion.
You know, it's a complex picture, I would say.
Right.
If we take. Sort of multiple moving pieces. That, like, pipeline seems to still be robust. There's a lot of. Larger deals and sort of the expanse of what Cloudflare is going after is expanding, but the conversion rates and then the close rates are still. Up in the air. The number one question I get from investors, like at the end of every single conference call, is, "Is the outlook de-risked? When you guys gave your forward outlook, it doesn't look for a lot of deceleration, and I think it sparked concerns that it's not a de-risked outlook. How do you answer that question on terms of the level of conservatism in that outlook?
Well, we try to, as always, when we give guidance, we try to be prudent and take into account what we know and what we control. There are, on the one side, the improved fundamentals, but there is the uncertainty of that, you know, how is pipeline ultimately translating into revenue. I think we saw in the earnings call, we actually feel better about the year than, you know, the risk in the short term. And I think that hasn't changed. You know, we see good traction with the products, especially with the new products. Pipeline is healthy and sound, but we deal with the uncertainty of seeing the inflection in the fundamentals really turning into revenue. You know, I think it's unchanged from what we.
How we talked about it in the earnings call. For the year, we feel we've done our job. Contrary to what people expected, we think there's probably more risk in the short term than there is in over the course of the year. You know, we'll work our way through it.
Okay. shifting gears.
Yeah
Sort of focusing on the Cloudflare story from a much more high level perspective, I remember meeting Matthew Prince years and years ago, and you met him and asked him about his company, and he said, "Well, I'm building a better internet.
Yeah.
You're like: Wow, this guy, he's got some major aspirations. One of the most amazing parts of the story is the innovation cycle and how you guys have come out with act one, act two, act three products and really starting to build out that vision that Matt had a decade ago in terms of building a better and safer internet. Can you talk to us about sort of that massive expansion of that solution portfolio?
Yeah.
How has that sort of expanded the TAM that Cloudflare is going after?
Well, you know, I always go back to when I started, which is now more than five years ago, I thought I had some idea on how disruptive Cloudflare could be. Now looking back, like, you know, I'm surprised of what we are doing, and I way underestimated the potential. It comes back, you know, with a couple of meetings today where people are new to the story and they say, "Well, what is the best entrance into understanding us?" It's really understanding the architecture of the network and how the offerings, the services really sit on top of that hardware stack, and literally every product running on every server in every location. The bigger we get, the more flywheels are flying.
We see so much data now that. Even just optimizing the utilization of the network is an innovation source. We, you know, there's unutilized capacity in parts of the network at certain times, and how do we fill that? How we can use excess compute or storage capacity or bandwidth capacity to innovate is huge. The whole organization is built around it. Our product development organization is unique. We almost have two parallel organizations. One is working on products that are GA'd, and one team is working on products that we call wave T plus one wave, so to speak. We started as, you know, with what we call wave one products.
It was, we sized that as a $35 billion TAM at the time of the IPO, and it includes the, you know, routing and DDoS mitigation application, this protection services, if you want to say so. The second wave of products came, very timely, I have to say. We launched our first Zero Trust product, our access product, VPN replacement, if you want to say so, in the Q1 of 2020, right when COVID started. That is now a fully-fledged product line of Gateway, Browser Isolation, email is part of that, and now also threat intelligence. We'd say that's another $35 billion of TAM. There's what we call our wave three products, where R2 and D1, the storage, products are in there.
Workers is not even sized into that market, we think we are well north now of $100 billion of TAM market that we disrupt. you know, the flywheel is going strong. you know, we'll have Security Week coming up in two weeks. I understand what is in the pipeline, there's a relentless pace of innovation that will keep us exciting. the markets and TAM we disrupt expand, you know, as we move forward. Super exciting. Yeah.
I think the other really interesting part of the Cloudflare story is on one side of the equation is this innovation engine, which has been fantastic.
Yeah.
The other side of the equation is the network, right? To me, one of the most difficult things in software is to find durable competitive moats.
Yeah.
The technology evolves so quickly. It's so hard to, like, just use innovation alone to protect your sort of differentiation, to protect your competitive moats. The network is interesting in that way. Could you talk to us about sort of the scale of the network, how that creates differentiation? You've talked in the past about how you could use some of that gross profit dollars from that network to actually further protect and create, a competitive interest.
Yeah. Yeah. As I said before, you know, if you want to understand us in the competitive moats, you really have to look at the architecture of the network. You know, our CTO would say, JGC would say, "You know, we can change a lot of things. We cannot change the speed of light." We have to move as close to the eyeballs, as close as what wants to connect to the internet as possible. You know, that determines the footprint of the network. It used to be websites and mobile devices and IoT devices, you know, as we evolved. The hardware stack is off the shelf. There's no uniqueness on it, and that helps in terms of procurement. It helps us in terms of delivering these gigantic network at superior capital efficiency.
Our CapEx ratio now is 12% of revenue. Looking at the infrastructure that is there, that's a very competitive number, and it's part of why the margins are where they are. All the offerings we have run in one homogeneous, fully integrated software stack on that hardware. The software stack is completely independent from the hardware that is under it. Doesn't matter whether we buy Intel CPUs or AMD CPUs or Arm CPUs or NVIDIA GPUs. The software stack is agnostic to what is under it. With that, it means that every server we add, regardless of where in the world, you increase the surface of this network, and you can use the surface. It increases your degrees of freedom in how you manage costs, supply, demand, bandwidth across the surface of the network.
With every server, as I said, it gets bigger and bigger. With the network getting bigger, we see more and more traffic. It's almost like an Apple Watch on the internet in terms of performance and congestion, security, threats, that become an input also in the innovation engine itself. There's an endless amount of opportunities of what else we can put on this network. As I said, it drives innovation faster. It's also unique in the cost structure because if you look at the first wave of products, they determined literally the size of the pipe, how data moves through our network. You don't pay for bandwidth by the amount of data that moves through this pipe. You literally pay for it by the size of the diameter.
All the first wave products are, you know, determine the size. Now, if these traffic, the coming back that, that can fill this, the pipes, and that's why products like Zero Trust that bring traffic back, so to speak, from the eyeballs is literally at hardly zero marginal cost. You have a second wave of products that is coming at close to 100% gross margin. You can use then, you know, this margin profile, not so much to discount, but to disrupt. You know, when we talk about pricing philosophy, it's never about, you know, okay, let's be 5% cheaper than this competitor or that competitor. The question is always, what is the most disruptive move we can make in the market? Unmitering DDoS was such a move.
You know, if you are a large customer, like, the bank that is hosting this event here. If you are an enterprise customer, unmetered DDoS is just part of your package. We can afford that because it doesn't really matter how big a DDoS attack is. The size of the attack doesn't increase the cost we have mitigating the attack for our customers. All of a sudden, this, the architecture of the network and how products funded it determines a really superior margin structure that we can use to disrupt further and to innovate further into new markets. It's quite impressive flywheel on its own, if you wanna say so.
Yeah. Definitely fascinating. I wanna dig into sort of the various product categories and maybe starting with those wave one products, the application services that Cloudflare got started with. I think the number one question I get from investors is how much more runway for growth is there left in these application services?
We think a lot, you know. The first, you know, answer when you get a question like that is it's not a static offering, right? We continue to innovate in these base services on threat levels, you know, DDoS mitigation from Layer four attacks to Layer seven attacks, where you defend, how you defend. Offering a product for financial service institutions in Europe that offers Layer seven DDoS mitigation that is GDPR compliant. It's not a little feast, and we think we are superior just because the architect of the network. You have the first wave of products are not static. They get innovated. They disrupt a significant market in terms of installed on-premise infrastructure that needs to get replaced. We think it's gonna drive growth for years to come.
In terms of growth rates, however, what we call our second wave of products, the Zero Trust products, will overtake application services from a growth rate perspective already this year. In terms of size, that will come, but in terms of growth rate, Zero Trust related products will overtake application services already this year.
Got it. Can you , have you given investors any kind of, perspective on the relative scale of sort of Act One versus Act Two versus Act Three products?
We have not yet. We are thinking about how we do that. And it's a bit harder for us to do because we're not really selling products. We sell bundle of products. When we got started, you know, we described that, you know, the revenue structure is literally 50% of our revenue, is security-based product. Even with the start of the Zero Trust wave, that has not changed much. Revenue has grown significantly. It might be more slightly biased towards security products, but we have not broken out yet, of how investors and potential investors can measure us better and you folks can measure us better. Well, our analyst, our investor day is coming up in May.
I think we'll have some ideas how we describe that. We are referring now, if you go back to our earnings calls, we put a lot of focus on describing where the wins are and what verticals are we successful, how big the wins are. You have some idea where market traction is and how we are doing from a competitiveness perspective. We understand that we have to become a bit more transparent in showing success in this field because it's going to be a big part of us moving forward, and it is going to determine the growth profile for years to come.
Got it. If we think about the growth dynamic in application services in particular, is it still geared mostly towards new customer additions, or is it more so kind of uplifting those existing customers to become kind of a higher ARPU type of-
It's both. Even in customers that have grown with us from the first minute, almost from $10,000s to multi-million dollars of ACV per year. We still have customers where we have potential to migrate to the second wave product. There is room in that customer base, both from an expansion opportunity as well as from a new logo opportunity. There are quite a lot of more traditional verticals out there that are sitting on significant on-premise infrastructure that needs to be disrupted, where business models are a little bit slower to digitalize and transform. Some of that has accelerated during COVID, but there's still a lot of runway left.
Got it. I wanna shift gears to sort of some the act two. Before we go to Zero Trust, I wanna start on network services. I would describe this as probably the hidden jewel within Cloudflare. Probably a part of the equation that a lot of investors don't really fully understand. Maybe you could help us talk a little bit about stuff like, the Magic WAN and some of those network services and the value proposition that gives to the end customer?
Are you talking Zero Trust now or?
No, the network services.
Well, you know, it's all about sitting behind one control plane in a way. Regardless if you have infrastructure in branch offices or headquarters, in subsidiaries, how do you bring all that infrastructure behind one control plane and get the performance. Efficiency and cost efficiency and effectiveness and security of on-premise equipment migrated to the cloud. There, you know, at one point you will just take a cable and plug it into a wall, and this infrastructure, that branch office will sit completely behind the Cloudflare network with all the services we offer. That's a compelling offer. Customers are excited about that. It's probably of all the products we have the most complex onboarding because you.
Especially with large global customers, you deal with a very heterogeneous corporate networking infrastructure in, you know, in some cases distributed about 15, 18 global ISPs, and getting this all migrated is sometimes not a small task. It's an overwhelming good value proposition. You know, it's if you. I think we launched a product, talked about it in 2019, end of 2019, the first time. The buildup of the architecture and the product was almost gearing to that event that literally all you need is a piece of fiber that you plug in a wall and everything else becomes Cloudflare and is delivered by Cloudflare. It's a huge lift. It's a huge opportunity.
We onboarded one of the largest oil and gas companies on this platform in the Q4 of last year. There's traction, but it's also from a go-to-market motion, a heavier lift than the other products we have.
Got it. On the reverse side of the equation, I think it's also one of the most sort of, well-defined market opportunities because and correct me if I'm wrong, like this is a MPLS replacement, and last I checked, there's something in the order of like $45 billion spent annually on MPLS.
In some of the logos we work with, it's still the largest spend category in the IT budget. Is MPLS replacement. It makes for extremely good ROIs, and, especially in times like this where budgets are flat or shrink, the ability to take significant spend out that can be redeployed is a big lever. It comes at a, you know, a, a steeper implementation and a high implementation cost. You have a trade-off that you have to go through.
Got it. I think one of the areas that's kind of difficult for software investors is there's a lot of different vendors who talk to us about their wide area network products. What's the real competitive environment out there when we're thinking about Magic WAN and how do you differentiate? Like what makes the Cloudflare solution stand out?
Well, there are a couple. you know, once you are deployed, you just talked about MPLS spend, the ROIs are really superior. Because the products on our side are highly margin efficient, you know, that's normally a win-win on both sides. It's the consolidation of offerings that is becoming a bigger and bigger part, especially in an environment like this. It's one control plane that becomes more and more important, especially if you look at the very heterogeneous and global fragmented infrastructure of offices, manufacturing facilities, branch offices, and you see performance security and, you know, and efficiency all in one control plane across all the offerings you have.
Yes
Whether it's routing, firewalling, DDoS mitigation, other secure bot mitigations, and now of course, also Zero Trust, on-ramps.
Got it. Let's switch gears to Zero Trust because it's been a really hot topic for a while in security, but also a very fragmented market. A lot of different vendors coming out of that space. One, who would you sort of term as your core competitors in that space, and how do you look to differentiate the Cloudflare Zero Trust solution?
Well, you know, similar to what we've seen on the application services. You know, the first set of competitors are on-premise solutions. You know, they're, I don't want to talk to too many names, but the Ciscos and the Citrixes of this world, on-premise VPN infrastructure, that needs to be replaced, that it has a hard time adjusting to a remote working environment. That is certainly a big part of where our business come from is replacing on-premise infrastructure. Then, you know, there are names that are probably have been here on the conference too, and Zscaler falls into that category. Netskope falls into that category that you see, that we see more and more often.
I would still say today, you know, while that is hyped up a little bit, most of our opportunities are still replacing and displacing on-premise infrastructure.
Got it. Then, you guys rolled out Cloudflare One, which has been sort of a bundle of the solutions. What's the reception been thus far? How aggressively do you have to sort of push on pricing in that bundle to sort of garner traction?
I think it's big. I mean, we've been working on. You know, one of the problems of, you know, or opportunities, I think is a better word to describe it, is once you have such a rate of innovation, once you expand your product portfolio so aggressively in short periods of time, is how do you deal with this in front of the customers? How do we bundle offerings in a way that they help us get efficiency in our go-to-market, enable our customers to understand better what we do? We've been working on this for a while. The reception is I think is really good. We looked at companies that we admire that have gone this path before, Salesforce, Microsoft.
That's how you buy Microsoft 365 Office today.
Right.
You're not trying to figure out how many people are using PowerPoint and Excel and, you know, what's the best distribution license. You just buy the bundle. This is pretty much what you do with Cloudflare One across the offerings we have. I think it allowed us to get efficiency in our go-to market, from an enablement perspective, how we structure the account teams. It has been rather successful. From a pricing perspective, we don't have to discount our products a lot because we are never the price leader when we enter a new market in the first place. You know, our prices are competitive because our cost structure allows us to be competitive out the gate and still have superior margins.
Pricing pressure is not really, is not really one of the biggest topics we have. We don't go out with high prices that we discount deeply, just because the business model allows for a much more differentiated approach, I would say.
I want to make sure we touch on Workers as well. Probably one of the parts of the solution portfolio that investors get most excited about, and it seems to be almost an open-ended market opportunity in terms of what can be built.
Yeah.
On top of Workers. on the other side of the equation, it's the part of the TAM that you can't quantify yet.
Yes.
How do we rectify that? Like, what is the unknown that makes it so difficult for you to quantify kind of what that opportunity is?
So for, you know, It's a way of how you deploy code at the edge of our network. That is really what it is. You know, before a product like Workers code could run in a data center where it was all about speed of execution, or it could run on your endpoint, where it was all about latency. Workers now finds this third spot close to the device, so you can execute code at the latency almost as if it were on the device, but at the speed as if it were in a data center. That was not developed in terms of a market opportunity. It was first a means to solve our innovation problem.
If you innovate this fast, put out so many products, we felt that our own structure and tools were hindering our rate of success. This is how Workers was born. Today, Workers as a platform is the basis of all the products we have. If, you know, the first use case then is you buy a product from us, like, I don't know, a Load Balancer. You attach Workers to it, you can use then Workers to make this the most customized load balance on the planet. This is where it all started. The use cases have changed over time, and what we thought would be interesting has also changed over time. There's a really interesting blog post on our website that Matthew wrote, I would say three years ago.
It really talks about Workers and how it works and what the architecture is. He says when he launched it, we thought it's all about speed applications. We had really interesting use cases of automotive manufacturers trying to offload compute capacity in the car into our network in order to take bills of materials down. We had security companies pushing artificial intelligence and face recognition in front of dumb security cameras. These are all about. It turned out that compliance becomes the more interesting use case in the midterm. Today, we attach it to almost every enterprise contract we sign in Europe or in parts of the world where GDPR or GDPR-like rules are prominent. Workers allows you to really control data on a very granular level.
This is Keith's credit card number, and he's a citizen of Switzerland, and it cannot leave the country, it cannot leave the canton, it can only reside in the data center of that security level. It allows you to run really granular workloads. That has become the big opportunity. Now, you know, we cannot finish this call without me mentioning AI three times. It turns out that combining Workers and R2, our storage product, has become a really interesting use case for a place where you train AI models, where you know. Here at the conference yesterday, people talked about how GPU capacity is being outpaced by the demand for training models.
You can reside data in our R2 and then really find the cheapest available GPU capacity at whatever hyperscale it is, and then optimize your and triage for training your AI model. The use cases have evolved, and that is the transition then to dollars. We think we would enable that explosion of use cases and adoption if we push for revenue too fast. We've been talking about this for a while. Not only is it difficult to size, but we think adoption and getting it in as many hands, especially in hands of developers, is the goal today and probably for the next two years before you push that towards revenue. Sizing is difficult, but I think it would be counterproductive.
We don't want any use case being discouraged to run through Cloudflare, just because we price it too early, we price it too high. It's like how Cloudflare was built, right? We, we want to eliminate any disincentive to move a byte of data not through the Cloudflare network. I think it's the same for Workers.
Got it. I'm gonna try to sneak one last one. I'd be remiss if I had a interview with the CFO and didn't ask a margin question. You guys did see really nice margin expansion, almost 500 basis points of margin expansion in 2022. Matt has talked about wanting to continually reinvest in the business.
Yeah.
How should we think about that balance between growth and profitability? Are we at a point where both could be going in the same direction?
Well, you know, we are growing up. We are getting to scale, and we understand that there's a trade-off, and I think that is what you saw, last year. We saw growth slowing down, and, you know, we also were able to show the elasticity in the business model that we were able to translate a reduction in growth rate immediately into an increase in operating margin and, you know, delivering, finally delivering free cash flow then in the second half of the year as we promised. We understand that, and that is a trade-off we keep in mind. There are minimum hurdle rates we have in terms of, you know, how fast we need to grow and what rules we want to follow.
As long as, you know, the macro environment is muted and the outlook has uncertainty, you know, we understand that we will deliver, you know, more profits accordingly. It worked out really well last year. We saw the slowdown early. We slowed down hiring and expenses really dramatically. That prevented us from doing more need of drastic actions like some of our peers. We managed quite nicely through last year's environment, you know, hopefully we'll be able to do the same this year.
Outstanding. Well, thank you so much for the time, Thomas.
Thanks for having us.
My pleasure.
Yep.