Joining us, everybody. We have the pleasure of having F5 with us. Kara and Cooper, thanks for making the trip out from Seattle.
Yep. Thanks for having us.
Appreciate you guys making it again this year. I'll, I've got a bunch of questions to run through, like I did on the last one. I'll open up towards the end if there's any questions from the audience. You know, maybe just to start, obviously, you know, we talked about systems and software with you guys. Maybe just on the systems side, as we think about, you know, with the backlog flushing through kind of faster, better availability of parts that you're getting, how far out do you guys have for visibility at this point on the systems side? You know, is there a way to, you know, shape this, that we can see beyond a quarter or two, at this point?
Yeah. And, before I answer, I do have to get our safe harbor on record. So 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 these risk factors. So with that being said, on the subject of visibility with hardware, yeah, so we talked about our backlog, going into the year was about $230 million. We expect to have substantially worked down that backlog over the course of the year. So what that's done is it's improved our visibility on, the supply side of the equation.
So that's where we've obviously had a lot of, kind of ups and downs on what we've been able to ship and record as revenue. And so from that perspective, we feel like we've got a lot better visibility as to the availability of our, of our product, supply. And so, so that's a good thing. The demand side of the equation, I think our visibility is actually pretty good on the needs of our customers. So we do a lot of work, kind of looking into our installed base with our customers, the aging of appliances.
We know kind of what's a healthy level for customers in terms of the utilization of their capacity, the aging of appliances, and clearly, over the last year, demand has been a little bit suppressed, and that's really kind of a macro-driven event, where customers are sweating their infrastructure. We've had experience with this with prior economic cycles. That's a common pattern. What we've typically seen is that customers, when they go through these periods of sweating assets, it's lasted kind of four to six quarters or so. But ultimately, their capacity needs don't go away, and it's a heightened state of risk that they're running their applications in when they continue to kind of elongate those life cycles.
Our experience is that customers tend to snap back to the demand levels that they've seen historically. We would expect that to happen again. We've seen really good traction with our new appliance series, the rSeries, and so that tells us that customers are continuing to move forward with hardware-based architectures for those legacy applications. So, you know, we don't have perfect visibility as to when that demand would bounce back, but kind of based on prior cycles, we would expect that to be, you know, sometime over the course of FY 2024.
With that being said, you know, how should we start to think— You, you're not going to guide here for fiscal 2024, we all know that. But how should we think about that fiscal 2024 systems headwind, given the, the backlog normalization or, you know, normalized systems growth, even beyond that, as, you know, horizon three, horizon four, whatever you want to call it?
Right. Right. So really, we're talking beyond FY 2024. Based on what we're hearing from customers and how their applications are growing and the needs of those applications, we continue to see the systems business as kind of call it a low single-digit declining business with stronger growth on the software side.
Got it. You know, Kara, something you and I have chatted about frequently, and I think it's important to talk about.
Look, we've kind of gone through this system software transition to a degree, but we've also had this iSeries to rS eries.
Mm-hmm.
The rS eries, the frequent advantage we hear about is, you know, standardized components, and just brings up the question of, like, why do you guys need to be in the box business anymore, as opposed to, you know, making that a pass-through? You know, what are your thoughts on that?
So it's true that the rS eries does rely on more standardized components in terms of there's not the same dependency on specialty networking semiconductors, and then also a lot of the components in the rS eries are from more modern process nodes, and so the components are more broadly available. But that said, there's still a lot of special sauce that F5 engineers directly into our physical systems, and that special sauce comes in the way of accelerating many of the functions that we have in our BIG-IP software. And this is the approach that we've always taken with hardware.
F5 is, and has always been, fundamentally a software company, and our core value proposition has been these robust and highly performant application delivery and security services, that are part of a software capability called, you know, first called BIG-IP, and now we've expanded that to NGINX and also F5 Distributed Cloud. So the purpose-built hardware has always served two functions. One is it allows acceleration of certain functions in that software, and you can think of use cases like TLS offloading, or also, you know, think of super high throughput, service provider use cases like CGNAT.
But then the other purpose that our hardware serves is that many of our customers really prefer these end-to-end integrated systems because it reduces their organizational dependencies on other parts of their company, to help provision hardware, and manage things for them. A second point I want to make is that the rS eries appliances and VELOS are very different, differentiated from prior generations. So in investor conversations, we have a lot of conversation about the difference in the bits, the parts of it, but the reality is, what the customers really care about is the fundamental innovation that we've packaged into these things. And, you know, just a few things I would highlight, our r Series and VELOS are far more automatable than prior generations.
And secondly, we've built them so that they have a multi-tenant platform, meaning it can host multiple instances of BIG-IP side by side. And that—when you, when you ask customers how, how they're looking for operational innovations in their ADC infrastructure, that is a huge deal for them because it means that they can achieve the, the panacea of hitless upgrades. They can basically put one instance of BIG-IP next to another one and then bleed the application traffic very slowly, and it means that the application doesn't suffer downtime, and that's a super, super cool innovation.
And then the final point I would make on this one is what we hear from customers is that the choice that we give them in terms of deployment model flexibility between hardware, packaged software, SaaS, and managed services is a very valuable benefit to them. And they like being able to say, "Okay, maybe my data center exit plans have been postponed. I'm going to have to stick with hardware for a bit longer," or, "I want to accelerate this part of my business, let me take the software form factor." And using F5 products and capabilities across all of our product families gives them that optionality, which they value tremendously.
Moving over to the software side, you know, we started the year. I think it's important to set the table here.
Mm-hmm.
We started the year, you know, expecting greater than 15% growth. Obviously, the market environment changed a little bit overall, but, you know, is there anything else to think about beyond kind of the macro demand impacting that software business in terms of that initial expectation?
Yeah, I mean, that really is the main factor in the lower software growth rates. And as we went into the year, we talked about, you know, over half of our business was still software business is new software projects, and we didn't expect to see growth given the macro heading into the year. And it's actually what we've seen is there's been a decline on those new software projects. It's been pretty substantial. The renewal parts of our business, the ratable parts of our business have been very healthy. We've had good net retention rates, but really, it's these larger kind of digital transformation projects, a lot of them have been put on hold.
In some cases, it's been approvals at the last second that have, you know, gone further up the stack, where these projects didn't get funded or they got pushed out. And that really is where the weakness has been on the software side. The good news is, last quarter, we started to see some of that. The more irrational ends of that behavior start to stabilize. We saw more projects getting funded and starting to move forward. It's still a challenged environment, but we do think that those projects ultimately will get funded and move forward. And to the extent that the rest of our business remains healthy, and we're seeing good expansion on the subscription side of the business, we think that that is an opportunity to see a return to growth for software.
Makes sense. And one of the debates that we frequently hear is load balancing in the cloud.
Mm-hmm.
So, obviously, F5's the dominant player on-prem and in private cloud. You know, as we transition towards cloud-based environments, how does F5 maintain that dominance? And do you actually need as much load balancing in the cloud?
Absolutely. And one point I would think, one thing I would point to is when you look at the ADC market, for which load balancing is a big part of the ADC market, ADC as a Service is the fastest growing segment of that market and is now actually outstripping software-based and hardware-based ADC. Now, the majority of that market today for ADC as a Service is led by public cloud-native solutions like ELB, ALB. We see real opportunity for a third-party ADC as a Service offering that offers customers a capability to deploy and manage their traffic across a hybrid and multi-cloud environment. I'd say the other thing I would say is your question kind of ignores the fact that BIG-IP was, you know, ADC, that's such a decade ago.
In reality, we have a lot more capabilities in the portfolio beyond load balancing and even more core ADC stuff, including a security portfolio now that makes up more than a third, more than a third of, more than a third of our overall business.
Yep, fair enough. Is there a way to think about the demand for within software for NGINX versus virtual BIG-IP, when you use one instead of the other, if there's a difference in deal sizes?
So no, there's not a material difference in deal sizes. We see, you know, a significant range in the deal size for both. And secondly, I think an important misconception that we hear from a lot of investors is they ask us about substitution or cannibalization amongst our product families. We do not see a significant substitutive or cannibalistic effect between BIG-IP and NGINX, nor between BIG-IP and F5 Distributed Cloud, nor between NGINX and F5 Distributed Cloud. So what we're seeing much more as a dynamic playing out is customers, and especially the customers that we serve, which are the largest organizations in the world, they have a very diverse need in terms of their application portfolios.
They have a mix of what we would call traditional, which is more monolithic client- server or three-tier type apps, and they also have a bunch of modern and container-native apps. Generally, most new apps, built after, call it 2014, are modern and container native, but that doesn't mean that they're decommissioning all of the apps they've built before that, nor are they spending a lot of time really modernizing those apps. And so they're managing this mixed portfolio. And increasingly, what we're seeing is that they're managing these portfolios hosted across, you know, several different deployment locations. So they have their on-prems, they have multiple public clouds, they have edge environments.
In our latest survey, so we do an annual survey of, you know, 1,000 global organizations, and what they said was 25% of the respondents there were deploying their apps and APIs and hosting them in as many as six different infrastructure environments. And so that heterogeneity in the applications that they have calls for heterogeneity in the solutions that we provide. You don't use a semi-truck for the same thing you would use a sports car for, and that's fundamentally the difference between a BIG-IP and an NGINX.
I mean, I'd prefer the sports car, but,
But sometimes when you need to move a house.
I suppose. I suppose. So, speaking of F5 as a service-
Mm-hmm.
Still early days on that, but, you know, what are the main services being adopted on F5 as a service or multicloud distributed cloud? And does that change the competitive landscape for you at all?
It does. So we call it F5 Distributed Cloud Services, and that is a capability that we introduced to the market about 18 months ago. We have two use cases that we are targeting with that. The first is a web app and API protection capability, or the shorthand from analysts is WAF. That includes a signature-based WAF, API security, DDoS protection, as well as anti-bot. And we actually have an award-winning offer at this point. We were just awarded by SC Magazine to be the top web application security solution in the market. So very excited about that one, seeing great customer traction.
And what customers are using it for is for those applications where they don't mind using a SaaS or managed service-based form factor, and that's an increasingly large portion of applications, especially as our customers are dealing with massive talent shortages in terms of cybersecurity talent. The second use case that we have on F5 Distributed Cloud is what we call secure multi-cloud networking. And what that is about is for those customers that are deploying and hosting their apps in these multiple infrastructure environments, it can become very operationally onerous and prone to a lot of error if you're trying to manually configure all of the connections between your on-prem, your AWS, your Azure, your GCP, your edge environments, et cetera.
And so there's a burgeoning market there, for solutions that abstract all of that and provide, effectively, an abstraction, or a cloud overlay layer, by which you can manage your applications and deploy them to any of these locations in a fairly seamless way. F5's solution in that space through Distributed Cloud is something that spans all the way from Layer 3 in the ISO stack, all the way up to Layer 7, which is differentiated from other available options in the market, which only cover Layer 3 and 4.
So, Cooper, turning this over to you, because as we start to think about catalysts for F5, one thing that, you know, our team looks at is new versus renewal on software. And it does start to imply that we might actually get a software re-acceleration next year. So what would prevent, or give us the puts and takes behind software potentially re-accelerating, as we look towards next year and beyond?
Right.
Is there a way to think about that incremental renewals that we get just based on durations and flows?
Right. Yeah. So I mean, there really are kind of two key factors. One is just the macro, you know, that customer behavior around new projects. Do projects start to move forward? That's an opportunity. And then on the renewal side, yes, on the packaged software side, we do have a larger base of packaged software subscriptions that are coming up for that renewal cycle in the next fiscal year. And so what we've seen to date is really good renewal rates and good expansion rates, and so that is an opportunity for an inflection on revenue on the software side, and that's really on the 606, ASC 606 side of the software business, where you get that upfront recognition of software.
Obviously, we're not guiding the software next year at this point, but those are kind of the things that I'd look towards to assess the opportunity.
Makes sense. Kara, going back to something you said and brought up, you know, greater than a third of revenue being security-
Mm-hmm.
You know, this kind of convergence of application delivery, application security. On the penetration side, or however you guys want to measure it, is there a way to understand, you know, what percentage of your customers are using at least one or two security products at this point, or a better way to understand that white space opportunity within your 20-ish thousand customers?
So, I'll try to offer some intuition around this, and I'll preface it with, you know, over the last six years or so, we have purchased quite a lot of capability that expands F5's reach and role beyond just BIG-IP into these modern applications and also into SaaS and managed services, and also expands us from, you know, a very heavy focus on ADC into more security capabilities. So then, when you look at that portfolio today, which we say is the only portfolio out there that can secure, deliver, and optimize any app, any API, anywhere, there's nobody else that can do that. The areas that we are looking for growth are four dimensions. So first dimension we're looking at is, you can look at net new customer acquisition.
We are actually seeing our F5 Distributed Cloud Services be a tremendous engine for net new customer acquisition. I talked already about how we expect customers to be using more than one of our product families, but we're also seeing F5 Distributed Cloud Services bring in new customers to us. The second dimension that we think about is expanding our capabilities and our service to support more of the applications that customers have in their portfolio. So for large enterprises, the average large enterprise has around 600-700 different applications that they are supporting. And in the past, with just BIG-IP, when I came into the company and had this conversation with customers, we were supporting maybe 20% of their most important, most mission-critical applications.
And now with NGINX and with F5 Distributed Cloud Services, we have a portfolio that has the right form factor and also the right operational model that we can reach 100% of the applications. That's the second dimension. Third dimension gets to your question, which is, okay, but it's not just load balancing that we're using and we're helping these applications with. We have a litany of other services, right? So we have web app firewall, we have DNS, we have the distributed denial of service capabilities, we have anti-bot, we have API gateway. Like, there's a whole bunch of those. And so it's not just a simple question of saying, how many of your customers have one thing versus many, because many of our customers, and the majority of our customers, are using many of those services.
And then the final fourth dimensional point, too, is application growth. So let's say F5 gets one of our pieces of technology, be it Distributed Cloud, or NGINX, or BIG-IP, embedded into an application of an organization, and that application just rips. It goes awesome. They need more F5 technology in order to support that capacity.
Got it. Cooper, how should we think about puts and takes for next year? You're not gonna guide, I know that, but between systems, software, and services?
Yeah, so on the system side, we've got kind of two dimensions you gotta look at. We've referenced a bit of a headwind on the backlog that we were able to ship this year, as we got caught up with supply chain. So with our backlog entering the year a little over $200 million, and we expect to exit the year at a more normalized state, from a reported revenue perspective, that's a headwind. On the flip side, we expect to see a resumption in demand as customers need to get caught up with their capacity needs. And so that should offset some of that headwind on the hardware side.
On the software side, we do see that there's opportunity, both in terms of the renewals that I referenced earlier and the expansion opportunity, and then also as customers move forward with their large digital transformation projects. So but, you know, kind of assessing what the mix is between hardware and software, that's, that's more challenging, but we do see that that in, in concert, that's a very healthy market with our BIG-IP and NGINX offerings. On the services side, we've seen really strong growth this year. Part of that is our, our price realization with the, the price increases that we introduced last year. There is still a little bit of a tail of that opportunity heading into FY 2024.
And then, of course, the longer or the extended renewal cycles we saw in FY 2023, that revenue gets recognized over time, so there's additional opportunity for growth on the services side in FY 2024.
Got it. Are there any questions from the audience at this time? We've got a couple minutes left. If not, I can keep asking. Go ahead.
One thing that I know we've been trying to ask a lot of people is just more on, obviously, there's been a big focus on AI, and not everyone is specifically pitching AI, certainly, but a lot of people are looking internally. So how are you, like, leveraging AI internally to maybe be more efficient, more productive, or maybe just any sort of projects you're using AI for?
You wanna take that?
Yeah. So there's a number of areas that we're looking at. And, you know, first, on the engineering side, there's an opportunity to do some more of the kind of low-level coding, leveraging AI to get more efficient. So you're doing, you know, more of that work, through, you know, automated tools rather than necessarily human coding. And so we think that will help drive more efficiency within just our engineering organization. On the services side, there's case deflection opportunities, leveraging AI and automation, you know, just prompts that customers can go to self-resolve, which also helps the customer stat. I mean, I think that most of us would prefer to quickly resolve our own issues with a little bit of help rather than having to navigate, you know, a human interface.
And then on the sales and marketing side, we think there's opportunity to drive better productivity within our sales force, just with things like content generation. So we've got a number of use cases. We've got a broad portfolio that can be very tailored to certain customer use cases. And our ability to recognize where we've been successful and then package it up into a repeatable sales play in the form of good content and enablement for our sales force, AI can be very effective at quickly turning those use cases into actionable content for our sales force. We think that's gonna help enable us to get to those opportunities more quickly as an organization.
And then there's a kind of longer range of automation improvements that we can leverage on the G&A side, just to get more efficient with things like finance and accounting, some of our recruiting processes. But the biggest opportunities are really on the engineering, support, and sales and marketing areas of the business.
If I might add, you know, we believe that AI is fundamentally about access to data, and F5 sits in the line of traffic of about 40% of the internet, and we also support almost 300 million websites, and so we have access to a lot of data. And we see a lot of opportunities for us to use that data and help our customers do better detection for issues, both delivery and security issues, but then also eventually, prediction and prevention.
Maybe just last one, easy one. How should we think about, you know, the delta between op margins versus free cash flow with this transition and anything on capital return?
Yeah. So, on the free cash flow margins, you've seen a little bit of a bigger delta between free cash flow margins and operating margins in the last year or two, and really, that's been driven by a couple of factors. You know, one of them is our transition to subscription means that whereas when you buy perpetual as a customer, you typically pay for that entire solution upfront. When you purchase via a multi-year subscription, while we recognize the revenue upfront, we chart, we invoice on an annual basis. And so as you're making that transition to a subscription model, the payment tends to lag the recognized revenue, and then if you start to stabilize on that mix, you would expect that to catch up.
And so we think that you'll see over time that delta between those two metrics start to shrink. And then, with the supply chain kind of prolonging our ability to ship, what we saw is we were shipping a greater percentage of our revenue late in the quarter, and so it's a little bit longer to collect on those sales. Again, we're getting caught up on supply chain. We think that the DSOs and our receivables will get kind of more back in line with historical. So I would expect over time to see that delta start to return to kind of more normalized levels. But, you know, that may be over a period of a year or two.
Well, I think we're out of time, so appreciate everybody for joining us, but special thanks to you both-
Thank you
... for your time.
Great.