F5, Inc. (FFIV) Earnings Call Transcript & Summary

Welcome, everybody. This is Meta Marshall. I head up the networking research here at Morgan Stanley. We're very pleased to have F5 here today. We have Francois and Frank, CEO and CFO. I have a short disclosure to read, Francois is going to jump into a disclosure as well. For important disclosures, please see Morgan Stanley research disclosure website at morganstanley.com/researchdisclosures. If you have any questions, please reach out to Morgan Stanley sales representative. Francois, did you want to touch on your disclosure?

Thank you, Meta, and hello, everyone. Before I respond to your questions, Meta, I need to get our safe harbor on record. Please note that our discussion today may contain forward-looking statements, which involve uncertainties and risks. Our actual results may differ materially from those expressed or implied by these statements. Please see our SEC filings for more information on those risk factors. Thank you, Meta.

Perfect. Great. So F5 has undergone a great transformation over the past couple of years from having dominant share into what people consider kind of a slowing traditional market to a company that's leveraged its position, its knowledge of application delivery to expand into security, modern applications, revised edge. How much of this was kind of the vision that you had, the Board had kind of a couple of years ago and how much of it has evolved as different opportunities would come along?

Well, Meta, I think it will be 4 years in 4 weeks since I joined F5. And if I go back to 4 years ago, I think there were a lot of questions, actually, outside of F5, but also inside the company of -- on whether we were a data center company, a networking company, an appliance company and really what was the going forward identity of the company. Through a lot of discussions, what was became very clear to us was that if you looked at the core mission of F5 and where we were at our best and what really differentiated us from other companies was that we had -- we were really good at solving our customers' biggest application challenges, and that we have to put -- we had built technology and a team of people and a support organization and a number of things that made us very good at solving our customers' biggest application challenges. And the fact that those things could be done in hardware, in software, in the public cloud or on-prem, like, all of these other things were kind of secondary to the core. And so the strategy, really, if you have to summarize it in one word, it was applications, and it still is applications. It was -- they say, look, the future of F5, the cornerstone of the future identity of F5 is going to be about applications. Now what we saw about applications already then, we saw 3 big trends that we knew would be big factors that we have to become leaders in if we're going to fulfill our mission of solving our customers' big applications, challenges in the future. Trend #1 was that more and more, our customers would want to deploy their applications in software-first environment and in cloud-first environment, evolving to multi-cloud. Trend #2 was that customers would either modernize their applications or increasingly build modern acquisitions from scratch that were in container environments. And trend #3 was that application security was going to become the problem -- the security problem in the next decade because more and more of the most sophisticated attacks, the most damaging attacks and, in fact, the most frequent attacks would be on application security. And so all of that we kind of -- was part of our original vision. And then from there, we went through a process -- quite a disciplined process of do we -- to adapt to this changing world and to continue to be able to secure any application anywhere, do we build, do we buy or do we partner. And so if you look at trend #1, which was, hey, applications are going to go into multi-cloud environment, we felt that we had all the technologies to support that, but we have to refactor our technologies, and a lot of that work was organic. So the choice was built, and we did that with BIG-IP. And a lot of the work we've done around automation and orchestration and integration with public cloud providers and new commercial models has really unleashed consumption of F5 in multi-client environment and has generated this growth in software. As it relates to trend #2, we looked at it and felt these container-native environments, we did not have a form factor that we could get to quickly and organically on our own, and that led to the acquisition of NGINX. Now this was really a build and buy because we acquired NGINX, but we had already started building a controller, which we merged with NGINX and is now getting a lot of traction and accelerated velocity with NGINX, but that put F5 in front of modern applications. And as it relates to trend #3, application security, we continue to build organically, and we doubled down with the acquisition of Shape Security, which now puts us in a privileged and unique position to not just protect application traffic, but also protect how applications are used. And so I think those 3 elements of the transformation, if you will, Meta, we kind of saw them, I would say, 3, 4 years ago. We knew -- we didn't know exactly how we would execute on them, but we knew that we would adapt as a company to these things. What has emerged more recently, if you will, is the issue of application distribution. It's only, I would say, in the last year that we realized that the kind of the issue of the next decade would be not content distribution, but application distribution because, now that applications are built in container and microservices, developers have an opportunity to put a lot -- select a lot of these microservices and put them very close to an IoT sensor or very close to a user and create a much more dynamic experience. And we went through the same process of that to build our supply, and we decided to go and build an edge platform Volterra, which gives us that app distribution, but also gives us the opportunity to deliver all of the F5 portfolio as a SaaS offering. So that's how things have evolved. But if you net all of that, Meta, if you come from playing just in a traditional ADC market, that was $2 billion and kind of grow in low single digits, to now we estimate that by 2023, we will be playing in a market that's $28 billion, and so much, much larger market, still with the same mission of securing and delivering every app anywhere, but having much broader role and much broader reach.

Perfect. Maybe diving into NGINX a little bit. Probably, the biggest question that we get is understanding that there is a need for kind of this need within the modern applications market, but just being kind of skeptics around the different open source solutions or that you guys will be able to monetize that market. You've seen traction here with NGINX having implementations that were up 109% year-over-year. So what do you think kind of investors miss about -- what they might miss out on with kind of an open source solution?

Well, I think the way to think about NGINX is a little bit to understand the success. So if you go back to the success that F5 had with BIG-IP, the reason the platform became so successful is because we built a best-in-class proxy architecture that was programmable, and that allowed us to consolidate a lot of functionality in a single package, a single package, software or hardware. But we started with low balancing, and then we added application firewalls and identity and authentication and encryption and decryption. And for customers, it basically allowed them to use F5 as kind of their gateway into the data center and not have to use 10 different vendors and basically from 1 vendor and 1 box or 1 type of software get all of that. Well, you think of NGINX is similarly a programmable platform that can do all the same consolidation, but for modern applications in container-native environments. And so I think what people missed is, at the time we bought NGINX, they had just started monetizing a very thin slice of that platform, and that's what we saw. What we saw in NGINX is a platform that we could create the same consolidation of services on this platform, but it was very early in the monetization, which essentially, at the time, were just load balancing plus, plus. And -- but it could do -- it could be an ingress to a Kubernetes cluster. It could be a sidecar to a Kubernetes cluster. It could be -- now it's a WAF as well, a security solution. It's an API gateway, and we're building API management on top of it. It's a map server. It's a web server. So all of these consolidation of modern services for modern applications in one platform is the same playbook, and NGINX has the incredible attribute of being able to scale very elegantly. The reason it's used by 500 million websites today is because it's allowed millions and millions of websites to scale by 10x in capacity with a very marginal increase in cost. And so you apply kind of the same properties to application services, and that's the power of the platform. The 109% growth in revenue that you mentioned, Meta, that was essentially achieved largely by NGINX coming into F5 in a larger scale organization with larger distribution. And so the growth to come from the monetization of a broader range of services, that growth is in front of us. It's yet to come, and it's not a catalyst that has played out yet.

Got it. Maybe touching on Shape for a second. This acquisition ended up being kind of incredibly timing -- timely, which is fraud detection becoming critical as a lot of businesses or just activity moved online over the last year. Clearly, there's strong demand for the product. Do you think you've kind of leveraged that for cross-sell yet or that's still ahead of you?

No, I think that's largely ahead of us, Meta. We've said we have increased -- at our Analyst Day, we said we had increased customer count by 15%, and we had grown ARR roughly 44% since the time of the acquisition. But the penetration of Shape into our customer base is still in the early days, so there's a lot there to come, and that cross-sell will happen not just on selling the Shape platform, but also taking the Shape capabilities and moving them to other platforms. So Shape is going to go on to -- it's already integrated now into our Silverline as a service solution, and we're actually seeing stronger demand for Silverline as a result. It's going to be integrated in BIG-IP, and it's going to be integrated in NGINX. And so we're going to monetize it across these platforms, and it is a priority to integrate Shape onto Volterra to be delivered in a self-serve SaaS edge solution. So there is still a lot of penetration to be had with Shape in its current form in our customer base, but also new variants to be delivered on other platforms, and all that's ahead of us.

Got it. And with Volterra, maybe people kind of think of F5 going -- don't think of you traditionally as an edge company and think of you more as data center centric, something that we kind of discussed in the first question. And so how do you think about evolving towards the edge? And how did Volterra make sense as part of that?

I think, Meta, first of all, and I think that was kind of the shift 4 years ago in terms of the way we think about ourselves, we don't think about ourselves as a data-centric company. We think of ourselves as an application-centric company. And kind of the moves we've made organically or inorganically are really about solving application challenges for our customers. And so we're on a mission to deliver and secure every app anywhere, and so the way that Volterra fits into that is pretty simple. And I'd say there's 2 big aspects to that. Number one is there are a number of customers that -- or a number of applications within an existing customer entity that don't want to consume technology as packaged software or packaged hardware. They don't want to own it. They want to consume it as SaaS, right? They want to be able to go on a portal, click on a few things and have the service come up because it's lower friction, there's lower demand on resources for them. For customers that have that preference, they have not had that consumption model from F5, and so they want us to think this method passed efficacy that we have and make it available as SaaS. And then the second aspect is, as more and more customers distribute their applications to the edge, at the edge, they also want to be able to deliver right there the kinds of security and delivery services that we -- F5 is known for. And Volterra is an edge platform that allows us to insert our security and delivery services at the edge as opposed to in a central data center, in an on-prem or in a public cloud. So that's a big shift, but all of that fits into being able to secure and deliver any application anywhere.

Got it. And a big question we get from investors is just on kind of Volterra versus some of the next-gen solutions that are out there, whether that be a Fastly or kind of other names. How much of what Volterra does is differentiated today versus it can become more differentiated as you layer in kind of all the different F5 capabilities?

So I think if you look at what's differentiated today, Meta, there's -- I would start with -- there will be 2 things that are really important for customers. The first one is that Volterra is built for application distribution, not content distribution. All these other players start from the perspective of solving the problem of content distribution, and then they're adding more app delivery and security services there because they have the footprint, and it's an opportunity. Volterra started from the standpoint that the problem that developers have is they want to distribute their applications, and what they've built is a platform that is intelligent about app distribution. They have basically done all the work around API networking -- intelligent API networking and orchestration and built a globally distributed Kubernetes platform, such that if you are a developer, you build your container once and you deploy it to hundreds of thousands of locations, and you don't have to worry about anything else. You don't have to worry about networking. You don't have to worry about firewalls. You don't have to worry about all of that. And you certainly don't have to go and rewrite in the language of a closed platform, as you would have to do on a cloud player or Fastly to deploy your application there. As long as you wrote it in a container that's orchestrated by Kubernetes, you're good to go. So I think that's a big difference from Volterra to the others. The second difference is that Volterra is also a universal edge platform, meaning, we don't intend to go and compete with public cloud providers on who's going to build the best infrastructure out there and the best point of presence. We believe public cloud providers are going to build infrastructure out to the edge, and we want to leverage that infrastructure and our partnership with them. And so Volterra is a software-defined edge platform that allows us to access all locations from whether it's Equinix or public cloud providers or, frankly, anywhere as close to a user as there is any form of compute available. So those are differentiations today. The differentiation that's going to accelerate with F5 is security. So oftentimes, what you'll see is CDN providers will offer security, which can be good enough, but for large enterprises really don't want to compromise or define their security strategy by which CDN they happen to be using because best-in-class efficacy is absolutely key for large enterprises. And so the -- with the sort of best-in-class security capabilities from F5, the Volterra platform is going to be security first. So it's going to be best-in-class security delivered at the edge. So I think those are the 3 elements that we see as differentiated.

Got it. That's helpful. Maybe I guess, for both of you. But with over 2/3 of F5's revenue recurring, you're an 80-plus percent gross margin company, 30% plus operating margin, free cash flow margin, F5 often kind of still mistakenly gets grouped into the category of a hardware transition versus an accelerating growth story from profitable software company. We just spent the whole first 15 minutes of this and didn't talk about an appliance once. How do you find -- what do you find is like the best way to kind of effectively get past concerns from investors, particularly as hardware or kind of that systems revenue has undergone a resurgence in revenue over the past year, just that this really is a software story that's past the hardware transition piece of the story?

Well, Meta, it's an interesting one. I think the first thing about F5 is there is no mix penalty with F5 because what -- you see our product revenue, you get 85% gross margins on hardware. So there's not -- whether it's hardware or software in the end, you end up with a gross margin profile that's very similar. But we definitely see ourselves, Meta, in the category of accelerating core company. If you look at the numbers, we did $360 million in software last year, and our software over the last -- from 2017 to 2020 grew on average at 50%. And we guided Horizon 2 to 35% to 40% growth, which means, this year, we should be in the vicinity of $500 million software business, growing at 35% to 40% per annum. And so I think we definitely see ourselves in that category. I think the big thing within that, Meta, is not so much just the hardware-software transition, but that the software business is -- the vast majority of it is now transitioned to subscription, and that is also going to continue to grow. I think we're about 75% subscription-based software business, but I think that percentage will only grow, and what that will do is kind of smooth out our growth rate and the predictability and the visibility that we have. We're already starting to see that in a kind of a flywheel of a broader base of subscriptions that have way more customers now other that their own subscription, starting to see as renewals come up and the true forwards and expansion of the subscription happen. So it's giving us strong confidence in the growth of the business, but also kind of the predictability of the business in the future.

Got it. And then for Frank, Francois, just kind of address some of this irons itself out just as you kind of get towards more recurring revenue versus some license base, but maybe with -- touching on the ELAs as well, there's just been questions that come up from investors about how to think about kind of some of the volatility that you see on the software line and just how not to kind of get caught up in the ELA component of that.

Yes. Thanks, Meta. I appreciate the question. So just to level set everyone, in our software growth, we've got several different consumption models that come through that line in order to meet the customer demand. So we've got perpetual, which is a shrinking portion of the software revenue. We've got flexible term agreements. We've got utility revenue that acts a lot like SaaS revenue in the sense that it's recurring and sort of recognized ratably in that period. And there is part SaaS subscription, and the SaaS is ratable over the period of that time. And so when we take a look at the software growth rates that we've seen quarterly and why they've been choppy, it's really for a few reasons. One is the perpetual license still is recognized upfront. And even the flexible term agreements, which we had previously and probably mistakenly called ELAs, have got us an upfront -- a smaller upfront, but an upfront portion and that goes in the product, and the balance is ratably recognized in the services line. So with the -- especially when you've got larger deals in earlier stages of those transitions, you're going to get some choppiness in that software growth. The second factor is that we've had some acquisitions over the last couple of years that have put new software revenue and contributions from NGINX and Shape, and this is the first quarter where we have lapped that and really not have anything of significance to talk about in terms of contribution from acquisitions. So in the last year, though, we've seen a much broader base of software adoption with fewer bulky deals, and we talked about each quarter having more and more of these multiyear subscription agreements that we do a record number almost every quarter. And as software revenue grows, no one particular deal is going to skew this dramatically one way or another. And so, particularly as we start to lap some of those multiyear subscriptions, which we see in the back half of FY '22, we think that the growth rate starts to smooth, and you don't see the same choppiness that we've seen before.

Got it. Probably the biggest pushback I get from investors on the F5 story is just the Horizon 2 target seemed aggressive, and that's something that they almost got a little bit more aggressive with the addition of Volterra. Where do you get confidence in just like where the wiggle room is there? Is it how fast applications move to the cloud, how fast new applications ramp, how much cross-sell there is? Just kind of what type of wiggle room do you guys think that you have in that number that just gives you confidence about the Horizon 2 targets?

Well, Meta, we -- there's a lot of work that went into these targets, and we -- I think you have to go back to the drivers of -- that give us the confidence in the growth, and there are 3 drivers, right? The first one is continued growth in software and subscription demand driven by our customers deploying more in multi-cloud environment, which we serve with BIG-IP, driven by the growth in modern applications and the growth of NGINX to support modern applications. Those are strong drivers that clear and present, and we're seeing them. The second driver of our growth target was application security. And you said earlier that the timing of the Shape acquisition was good because we see more demand for fraud as people move to more digital channels. And generally, application security is a bigger issue for customers, so that demand is pretty healthy as well. And then the third driver was that we felt that the decline in systems would moderate. Now if you remember, in 2019, it was in the double digits, and we said it would moderate to high single to mid-single digits, and that we are seeing happening. In fact, it's been happening better. If you look at the Q1 result, it happened better than what we thought, in part because we think customers are -- in part because acquisitions going forward, also these customers are getting more comfortable with hybrid architectures going forward. So those 3 drivers, I think, are pretty healthy, and we felt at the time that we put these targets together that there was some sensitivity analysis around, well, what if this one doesn't go well or that one doesn't go well. And what was kind of clear to us was that we did not all 3 of them to go well for us to hit the targets. So I think the wiggle room we have is that we don't need all 3 of them to go perfectly well, and that's what's giving us the confidence in these growth rates.

Great. Perfect. You've been acquisitive. That's been a good use of cash over the past couple of years. At the Analyst Day, you laid out kind of a commitment to capital return. Just how do you think about kind of, I guess, capital utilization between M&A and share repurchases?

Yes. I mean, so look, it's, obviously, Francois told us that we have spent about $2.2 billion on acquisitions over the last 3 years after not really having a history of doing many acquisitions. And this was all very centered around the strategic vision that Francois laid out at our Analyst Investor Day in 2018, and it has been obviously allowed us to go in the direction to serve every app everywhere in a very secure manner. By going forward, we see a balanced opportunity now that we've got the major building blocks in place where we can sort of share the capital that we're generating for -- reserve for strategic purposes with our commitment to return capital to our shareholders. At our November Analyst Day in 2020, we talked about our commitment of repurchasing $1 billion worth of shares over the next 2 years, this year, in an accelerated share repurchase format and next year more ratably over the quarters. And we initiated that, as we talked about in our Q in the beginning of February. We also said that we expect M&A to support our adaptive application strategy. And at this time, we noted that we expect to pursue targets that really accelerate that vision and top line momentum. But we also said that we're not going to sacrifice the commitments that we made to Horizon to non-GAAP operating margins as well as EPS guidance that we gave. And when we announced Volterra in the beginning of January, we stayed within those guardrails, and that's what we're committed to doing. And so we expect that we will continue to sort of balance out the capital towards M&A, but over the next 12 to 18 months is we are really committed to focusing on the integration effort, all of the wonderful things that are going to come out for sales of both Volterra's core products as well as what we're integrating out of F5's product portfolio, such that this will be a great acquisition for us. But that's going to take time to integrate. That's going to be our focus. We're really -- we're not focused on anything bigger than sort of $100 million tuck-in acquisitions over that period of time, but we're going to continue to balance that going forward on FY '22, with half of our free cash flow going towards share repurchase and the other half building our strategic war chest.

Got it. Well, I could ask you questions all day because I think it's a fascinating story. So congrats to you guys on the transformation that you guys have built over the past kind of 4 years and look forward to talking with you guys more as things go on. For any investors, if you have questions to me or if you have questions to Suzanne DuLong at F5, and we could get you any follow-up questions. So appreciate you guys being here today.

Thanks so much for having us.

Thank you.

Thanks.