Five9, Inc. (FIVN) Earnings Call Transcript & Summary

Awesome. Hello, everyone, and thanks for joining the 2021 UBS Virtual Conference. My name is Taylor McGinnis, and I'm an analyst on the software equity research team here. And in this session, we are joined by the CFO, Barry Zwarenstein; and President, Dan Burkland of Five9. Barry and Dan, thanks so much for joining today.

Thank you, Taylor.

Awesome. And before we get started, I would just like to remind everyone listening in that you can submit a question via the ask-a-question tab in the upper right-hand corner of the screen, and we'll make sure to leave time at the end to address some of those. So before we dive in, Barry, I'll turn it over to you to share some prepared remarks.

Thank you, Taylor. So before we start, I just want to remind everybody that we'll make forward-looking statements about events and trends that affect the industry and the company and its operations. Actual results may be materially different. Please refer to the 10-K, 10-Q under Risk Factors and elsewhere in such reports for detailed information about what could cause the results to differ materially from those described in the forward-looking statements we make. Thank you.

Perfect. So maybe starting with some questions on the macro environment. So -- and the impact Five9 might be seeing there from some of the newer developments. So there's talks of talent shortages. Now there's a new COVID variant. So would you flag any of these or any other events having any material impact on demand for Five9? Have there been any recent changes in the appetite to migrate contact centers to the cloud? Maybe we can start there.

Yes. I'll take that one. Thanks, Taylor. Greatly appreciate it, and thank you all for joining today. When you look at that, and definitely, there's press out there. There's absolutely a challenge for some of our customers to find the talent for their contact centers. However, when moving to the cloud, what we have created is the work from anywhere ability for our customers. And so what they find is they don't -- they no longer have to recruit within a 30- or 60-mile radius of their contact centers for talent, they can recruit from anywhere. So some of them have really opened up and widened the aperture with which they can search for talent. And many of them have been able to do so very effectively. Others have recognized that, hey, in this difficult labor market, technology and automation can certainly help augment that. And so we've seen more customers opting in for some of the technologies we provide and say, "Hey, we can offer self-service. We can offer a reduction in average handle time by delivering things like Agent Assist and with AI." So there's lots of ways in which you can maximize the productivity of that labor force without having to necessarily expand it at the rate that you might have otherwise. And so that's been very helpful for us. And so I think it's just one of these dynamic ebb and flows of the labor market, and then also the technology helping augment that.

Perfect. Yes. And then maybe like a second part to this broader demand question. But as we get further away from the initial COVID-related tailwinds, I guess, how conversations with current and potential customers evolved? We've spoken to several CIOs and IT executives that have invested heavily in customer service and support initiatives during the pandemic. So is there a risk that some of the large deal activity and pipeline activity that you guys have experienced over the last 1.5 years was pulled forward in any way, such that growth in lead generations or deal signings might normalize or start to moderate over time?

Yes. I don't see it. I mean, we definitely saw some immediate short-term required decisions when the pandemic hit for bringing in the flexibility of work-from-home and bringing in technologies that allowed you to monitor and manage your resources when they were working from home. But as far as -- if anything, that was just a wake-up call to bring more and more companies to the forefront to say, we've got to look at our customer experience and the options that we're delivering to our consumers because they're becoming more and more demanding, if not most demanding right now about what they want from the brands that they work with. So it's opened up the process to start adoption of these technologies. Bear in mind, your question was more about, gee, will it start to moderate or slow down? It's just getting started. When you look at large enterprise, we talk about our industry being in the high teens, maybe 20% penetrated as far as move to the cloud. If you go up into the large enterprise, it's single digits, for sure. And they're just getting started with their processes and decision making to come into the cloud and to take advantage of these innovative and automation solutions. And so we see nothing of a slowdown, if anything, an acceleration and an increase in our bookings and pipeline reflects that.

Got it. And then maybe staying on the topic of some of these larger enterprise deals, I know one area of big focus for you guys has been the 7-figure deal activity and obviously, something key that investors are looking to. So it's not hard to see how some of the deals even you guys talked about can have a material impact to numbers. I know like your last Analyst Day you talked about how the $1 million-plus ARR customers is your guy's fastest-growing category. But I guess, how durable is that pace of customer adds? And could that even continue to accelerate? We commonly -- and maybe it's like a second part to this question is we commonly hear some enterprise talk about maybe a slower move to CCaaS, potentially wanting to take a hybrid approach. So I guess how does that compare to what you guys are seeing and hearing in the market?

Yes. Yes, great question. We absolutely can see an acceleration to the $1 million ARR customers. And we're witnessing that right now. As most of you know, the adoption when we close business in that size category, it takes time for them to go through their implementation cycle, go live with their first agents and then ramp up the rest onto the platform. And so that we see accelerating. Again, to my earlier point, the market -- upmarket is single digits, and those floodgates are opening, and we see nothing but strong momentum coming from that segment of customers. And so it's early innings when it comes to that. We don't see any slowdown at all. In fact, we see the acceleration from the points I made earlier. And we see it with the adoption of the automation and self-service type options with IVAs. And so that -- as those become more and more prevalent and they sell for a higher ARPU, we see an increase on a per seat basis and total spend. Just harking back to the days of our IPO, we had 3 $1 million ARR customers, and I think we have 123 and growing today. So that should continue, and that's our highest area of acceleration that we're seeing both in pipeline as well as in bookings and revenue.

Got it. And then maybe staying on this, in order to help the audience think through the current and future financial impact from these large 7-figure deals, can you provide some color on the revenue contribution we have seen from them so far, given that they typically are to what you just talked about, recognized over time. So based on the deals that you have visibility to, how do you expect those to evolve in the future as a percentage of the mix and contribute to your 2026 revenue target?

Yes. I'll get that one.

Yes. Go ahead.

So these larger deals, they take a longer time to go live, like 12 months or so versus a more classic 3 to 4 months with the smaller deals. During the course of 2021, we had some PS revenues from these mega deals and the beginnings of recurring revenue in this current quarter, but we expect a progressive ramp during the course of next year from those deals that already been booked. In terms of the future, as Dan has talked about and you alluded to, we've had tremendous progress in moving into the big enterprise. These orders would have seemed incredible. As we've mentioned them even to certainly 3 or 4 years ago, the size of these orders. And it's the fastest growing, as you said. And we expect that to continue. And by the way, parenthetically will also help us with our dollar-based retention rate because these do have a better rate than the smaller customers.

And Dan, was there something that you wanted to add to that, too?

No, I think he hit it spot on. Thanks, Barry.

Okay. Perfect. And then maybe another question going off of that. So I think what's very interesting about you guys is the size of even some of the deals that you have today for a company of your guys' size. So when you think about the 120-plus $1 million plus ARR deals that you have today, when you look at your existing installed base in terms of maybe some of the enterprise customers that are just dipping our toe in the water, is there a way to kind of like think about that opportunity and that expansion that you have today? The portion of the base that could ultimately be like those 120-plus customers that you already have?

Yes, the mix -- so we see it from both segments, right? And net new, we're getting those customers for the first time, opening up to cloud and recognizing the value that these innovations can bring them, and they can only get them by moving to the cloud. And that's why not just for Five9, but our whole industry, you're seeing that uptick in opportunity at the high end of the market. But if you look at -- so that's one. And then if you look at our installed base, we've got lots of customers that continue not only to expand the seats with us in the new departments and so on, but picking up and adopting those technologies, which gives us greater ARPU even without increasing the seats. So it's a combination of factors and both are helping fuel our growth significantly.

Got it. And then, Barry, maybe one for you. So to me, I mean, I think the highlight of the Analyst Day was really the expectation for the LTM dollar-based net retention rate in the high 20s or 120s excuse me by 2026, which is up from 123% last quarter. I think ARPU expansion and enterprise opportunities underpin a lot of this. But Barry, you seem to have expressed high degree of comfort with that guide. So I guess what are you seeing that gives you conviction this is well above what we've seen historically? And with the pandemic being a catalyst that puts this metric higher, what gives you comfort that this upward trajectory can continue?

Yes. So there's few certainties in the business, very few, but we are about as confident as one can be overall that this trend will continue upward. So 3 factors. First of all, within the enterprise part of the business, the bigger customers, that fastest-growing segment, 87% CAGR of only 1 $1 million-plus ARR customers, they have meaningfully higher dollar ratio retention rates. And that makes sense because, as Dan has also alluded to, when you go into an account, a bigger account, the probability is you're adding all of that in all regions in the flash cut are lower and therefore, a bigger opportunity to expand. The second one is that we also have, over time, ARPU increases. And thirdly, we've got a little bit of assist as well from the fact that the 123 that we did report is a blend between our enterprise business and our commercial business. The commercial business is only 16% of our business, but it is 16%, and that has retention rates of somewhere in the high 80s, low 90s depending on the quarter. So -- and that business is growing slower than the enterprise. Now I did say there are few certainties in business, again, we're confident about. We can be even more certain about the fact that there will be fluctuations on the way to the upper 20s due in large part to these mega customers coming on at different times and ramping at different rates. And that's happened in the past and it will definitely happen again in the future.

Got it. And then is there any way actually to like going off of that to kind of like how should we think about those ramp as like to what I think you were talking about a little bit earlier, that even like this past year, you haven't really seen some of the deals that you guys have talked about really flow through to the revenue line. So as that compounds over time, any way to think about that trajectory and what we can see from a revenue growth standpoint as that builds?

Yes. So with those bigger accounts, that will be -- it's part of our thinking for 2022. It's a definite tailwind. We don't have the benefit of the -- onetime benefit of the COVID that we believe we were fully retaining. But we will have that benefit. And that's part of the reason, Taylor, that we gave them in the guidance that we gave for 2022. And just to remind some in the audience, for 5 years, through 2020, Five9 gave initial annual guidance somewhere around 16% for annual growth and then proceeded to -- as the year unfolded to improve, just prudent guidance that we gave. We took that 17% in 2021 and increased it to all the way to 20%. And now we're talking about 24%. And then part of that is due to the fact that we believe we're retaining the COVID benefit, and we can go into that separately, if necessary. And secondly, these mega deals.

Got it. And maybe just because you mentioned it, and I know that's a bit of interest to people. What gives you comfort in terms of retaining some of the COVID benefit that you've seen? Maybe you can talk a little bit about your conversations with customers? What you're seeing from those that gives you that conviction?

Yes. It comes mainly from our examination of the bookings trends. We saw a strengthening in our installed base bookings quite meaningfully during the period -- at the height of COVID probably still there, but when it was really strong. And the way to think about that is that if you rewind the movie to 2 or 3 years ago, the ratio between new logos and installed base was like 2/3 [indiscernible] and now it's closer to 50-50. And based upon the trended data that we're seeing in bookings and revenue, we saw an increase in Q2 -- excuse me, yes, Q3 and Q4 of 2020 and Q1 of 2021. And we're now emphasizing the fact that those quarters, Q2 of 2021 and Q3 of 2021, we saw sequential growth rates that were similar to what they were pre-pandemic, 4% and 7%, respectively. And that gives us -- since we're retaining the same sequential growth rate as pre-pandemic, it leads us to believe that we are retaining the full benefit. And it makes sense as well because companies are not going to undigitize post-COVID. There's new buying habits that have been installed. And we're seeing it in some of our customers. Some of our retail customers, for example, compared with Q3 of 2019, up by 50%, 70%. The telehealth companies are up by 140%, 150%. So those behaviors seem to have permeated the way people transact.

Got it. And then maybe taking the question I asked earlier on dollar-based net retention and taking that to like the next step, I guess, the other side of this equation is new customer logos. And if you're talking about 120s dollar-based net expansion rate, that only requires modest new logo growth in order to get to that 30-plus percent. So could you maybe talk about how you feel about the new customer activity side? And how you're thinking about that as we look ahead?

Yes, absolutely. And the new logos are obviously crucial. As I mentioned, half new logos, half installed base currently. And so hugely important to that commitment that we've made to The Street to have many years of durable sustained enterprise LTM subscription growth. And by the way, if you take that LTM enterprise subscription and the associated professional services, you're talking about 2/3 of our total revenue. So it's meaningful. It's not just a little 5%, or 10%, or 20% of the total. When we think of the future going forward to that new logo growth, Taylor, we think of it in 3 buckets: market, market landscape and execution. As far as the market is concerned, we don't talk about it internally because we know that it's extremely strong. It's barely penetrated, mission-critical. We talk about it to investors, of course. And the reason that there's these trends that are driving a premise to cloud and digital transformation and AI and automation that are not going to reverse. And we've got all these opportunity expanding channels, expanding internationally and going up market. In terms of addressing that market, though, it comes down to execution. And we submit that with this leadership team and with over 2,000 people who repeatedly demonstrated that we deliver on our commitments collectively. Execution can go wrong, but we feel pretty comfortable about that as well. And that brings us finally to competition. And given how attractive the market is, it's only natural that we'll see more competition going forward. We've heard the news from Genesis this morning as well in terms of their increased involvement in the market. And -- but as far as we're concerned, that may -- we believe that we are very well-positioned with our innovative platform and especially with our overall service offering, that is so critical in this industry, the implementation and the ongoing support. And it's a big market. It's barely penetrated and there's room for everybody.

Perfect. And then Dan, maybe on to that last part that you talked about, but is on the competition side. So with that being one of the risks here. You just talked about the Genesis funding round announcement that was made today. And when thinking about some of the like investors on that from, whether it'd be Salesforce, ServiceNow, I guess, what are your thoughts there? Maybe you can give, for those listening in, an update on the competitive landscape with Genesis and NICE inContact and others?

Sure. Yes. This market -- thanks, Taylor. The market, as we look at it, there's a big moat and it's very difficult to get in and break -- the barriers to entry are significant. We've seen that even from companies like Genesis. They acquired Interactive Intelligence, gosh, I think it's been over 4 years now, maybe 5, time flies. But they acquired a cloud solution that was already 3 years in the making, and then it took another 4 years. So I think about that time frame for them investing in that to now get a solution that's ready for market. Now they have a big installed base upmarket that's been patient enough to wait. Some of them have. Others have made the migration over to us and other cloud solutions over the years. But that -- when you look at it, it gives credence to the fact that the cloud solutions are going to really own this market in the large enterprise. And there's really only 3 of us. There's us, NICE inContact and Genesis. And so others have messaged to come play. Amazon is doing a bit, but it's a different model. It's -- you've got to really take a foundational approach and then apply resources and really customize and build your own. If you've got a team of developers that has that appetite to want to do that. We've had many customers that have embarked on that process and gotten 6 or 9 months into it and thrown in the towel and said, we're not going to get to where we want to go for 3 more years. We don't have that ability to wait and be that patient. So there's really 3 of us at play in the market, and there's a huge, huge market for us. I mean you take the $24 billion market TAM that we've talked about over the years, and you really can take that and expand it much greater when you start applying the self-service IVAs, the Agent Assist and other technologies and automations that we're bringing into the market. So we're seeing an expanded TAM. All the while, we're seeing even more barriers to entry because customer -- companies that would want to enter this space would have to build what the 3 of us have put into our products over the last 2 decades and add to it the new innovations that are coming very rapidly as we speak. So we feel very strongly about our place in the market, and we have great win rates when we go head-to-head against those 2 other providers, but there's often times opportunities that we don't see or -- and vice versa that they don't see that we're in. So it's a big, big market and lots of upside, and we're still in the very early innings, if you look at it. The mass market in the mid-market is by most assessments is in the high teens as far as the penetration and move to the cloud. And if you get into the higher end of the enterprises, it's certainly single digits.

Got it. And when you talk about the potential for competition to increase over time, can you maybe talk about like the source of that? So is that the potential for new entrants like Zoom? Or Microsoft recently announced a CCaaS offering. I know you talked a little bit about AWS, the potential maybe for more of that CPaaS-like architecture to play a bigger role over time. Maybe you could just elaborate, I guess, on what you mean by some of those comments earlier? And where are the risks? And who do you see as some of the threats?

Yes. I'll break it down a couple of ways. So the pure play, the NICE inContact and Genesis are kind of now at this point, 100% cloud. We each have a 100% cloud offer for the market. When you look at the entrants like AWS or Amazon Connect that has come in, it's a very different model geared towards developers positioned there. And again, it's effective at the very, very high end of the market where you do want to take a foundation and expand and build your own, and it's effective at the very low end of the market if you need something extremely simple. It's lacking a lot of functionality off the shelf. So it's very simple and easy to just download and operate, but it's missing a lot. So then that -- they're kind of both ends of the market. And then if you look at Twilio has lots of solutions in and around the contact center. Very point solutions that are custom in nature. We coexist. A lot of our customers have Twilio and us. It works very well for what they do and for what we do. But if you look at the end-to-end replacement, the ones that are going to take the legacy Cisco and Avayas and Mitels and others and move them to the cloud, you've got a very select group that can do that. Others that are starting to creep a little towards the space and start to -- the CRM guys all have their digital channels with chat and e-mail and they're starting to do more of that, but they need a routing engine like ours that has real-time visibility and real-time status of all the agents to know their skills, know what they're capable of, so when those interactions come, whether they're chat, e-mail or voice, we can then assess who's the best resource for this interaction and intelligently route it to the right destination. The CRMs have no visibility to even knowing agents at all. So they get increased. They put them in a bucket, and the agents have to go cherry-pick them out of that bucket and respond to them unless they have a routing engine like ours to help enhance what they deliver on their side of things. So CRMs are starting to dabble around the edges, and looking at the value of having deep integrations with all of us. But there's also -- but to try to come into this space and build a solution and enter the market, it's a many multiyear process. And I think sometimes people have underestimated that when Twilio Flex first came in 2017, I said we're going to enter into the contact center market and then they realized to try to build the end-to-end is pretty prohibitive from an investment standpoint. And a couple of others have tried, but they've had trouble getting up market. And then we see a lot of what we call ankle biters that will enter and exit and enter and exit at the very small end of the market. So when you see companies announce contact center, be careful because a lot of it is the voice. The Zendesk Talk, where we still do a ton of deals every quarter with Zendesk, and we're a big sponsor in their user conferences and so forth because when they say talk, they're talking about just voice. It's not really a contact center. It's great. Our voice-enabled so that when you're in the CRM, you see a customer record, there's Taylor McGinnis, I'm going to click on and do a click to call or I'm going to set up a very basic conference call. That's not really a contact center as we know it. And so when you look at Microsoft saying the same thing, we don't feel threatened at all by that in the market because it's many, many years if they were to continue to develop on that before they could really be a full-fledged enterprise contact center solution.

Yes, that's really interesting. And can you maybe elaborate a little bit on that? Because I think that you bring up a good point here because I think one thing that you're seeing is okay, lots more demand for communicating or engaging with end users or customers across different channels outside of just traditional voice and e-mail. And you're seeing the CPaaS players offer -- create like offerings in this category for the "digital contact center." You're seeing the customer service and support players also enter the space. So are you saying that even though you're seeing those offerings brought to market and even if some of the SMS channels, things like that are coming necessarily through you guys, you would still need like a Five9 or a contact center offering underneath in order to route a lot about that. Can you explain that role? Because I think that you mentioned in comment.

Yes. Thank you, Taylor. -- because that's a very key point is whether you're an SMS solution or an e-mail only solution or any of these point solutions, we integrate. We have an ISV program where we'll integrate or allow them to come integrate to us and become part of our ISV program and certify their solution as an extension of our platform. But they need the underlying platform, the routing engine, if you will, like us because we're the ones that have full knowledge of every interaction that attempts to enter the enterprise and needs assistance, and we have 100% visibility into all the resources to serve that customer. And when I say all the resources, it's not only the human resources and agents that are around the globe, but I also want to know what the status is of my, call them, silicon agents or digital agents. The ones that are the IVAs of the world. So I can understand how to treat every interaction in the best possible fashion because -- and that's what the beauty of our routing engine is, is because we're able to say, [ ha ]. I've got Taylor coming to us via this channel for what we anticipate is this problem because we've looked into the databases and seen that you've got an open case or whatnot. And who do I have to best serve your interaction at this point in time? It may be the person you spoke to yesterday. Now I can see that, and see if they're available. And if so, I can route you to the same person. So there's lots of decision-making in a dynamic fashion that happens in real time that is oftentimes taken for granted by new entrants because we're a niche play that says, I'm this one thing. I'm delivering SMS messaging. Well, they don't have the broader visibility that's required as a foundation. So we think most of those, and that's why we partner with many of them are able to bring their value-add on top of the Five9 solution.

Got it. That's interesting. Maybe switching to ARPU. So it sounds like Five9 is having a lot of success in its AI, automation and WFO products with strategic accounts. And we were even surprised to hear that AI and automation is already 12% of revenue that you guys provided at the Analyst Day. So monthly ARPU has been, I think, roughly flat at around 2012. So can you help us understand the magnitude of impact like adoption -- the adoption of some of these products is having and could have to ARPU over time? Maybe you could talk a little bit about the pricing of the IVA products, the pricing of the full product suite, the impact that could have? Would love to get a little bit more color there.

Yes, I'll start and then Barry feel free. But when you look at the automation solutions of that 12%, that included AI Agent Assist and as well as our WFO suite from Virtual Observer and Verint. So those are solutions on the WFO side. They've been around longer. We're typically replacing when we go in to our enterprise and strategic accounts, but they're also coming down market more effectively, and we're able to serve customers that previously didn't have WFO. And so we're seeing a higher and higher attach rate, if you will, of those. But as far as the impact that it's having, it's -- the IVAs, I'll take your pricing question. IVAs -- again, think of it as a port or a digital agent instead of the human agent. And the compelling ROI here is we can get $400, $450 per virtual agent. And you might think, wow, that's a lot per port. But if you compare it to the human agent, which may be 10x that number, it's tremendously valuable. It has a compelling ROI. So we're always working with customers to swing and deflect as much traffic over to the virtual agents as we can because the ROI is very compelling. When you look at the Agent Assist, there's a little more customization into that pricing because it all depends what we're doing. I think of Agent Assist as we're listening to the conversation, we're understanding through our NLU, natural language understanding what the conversation is asking for. And then going and fetching and retrieving the data and delivering it back to the agent so they can be more effective and more efficient with their time. So the question is, gee, are we shaving a few seconds off of the call? Are we making the call a better experience because they have the right answer the first time? Those are more soft dollars are to justifying a hard ROI, but absolutely compelling in the right environment. So hard to answer your question about how much it's going to contribute, but we know that curve is up and to the right. It's a matter of how fast that adoption is going to occur and for how many calls, right? Am I going to deflect 2% or 3% of my calls over to the IVAs? Or am I going to eventually get to where I've deflected 20% that I've, therefore, been able to reduce my labor force by 20%? It's going to happen. It's going to happen over a decade or more as society becomes more accepting of voice interfaces as they recognize the experience is still delightful and they're willing to have that conversation with a machine rather than a human. So we'll see. We know it's a good trend. We benefit regardless of how fast that trend occurs. And we kind of win either way. If they want -- if they want to stick with human agents, we've got lots of great software to help them there. If they want to move towards digital agents, we've got software to help there, too. So we win in either fashion.

Perfect. And Barry, maybe you could talk about, from the ARPU side, any sense of when you look at the current base or thinking about how ARPU evolves over time, what are realistic portion of the base that could have things like IVA or some of these multiple products that they don't necessarily have today? Any way to quantify or frame that?

Yes. It would be difficult to quantify because we are on the AI side of it, IVA side, let me be more precise, in some of the associated technologies, it's still somewhat to be proven exactly what the uptake will be. We know that there's a huge arbitrage between saving labor at $2,000 a month and paying $400 instead for technology. We know that the technology is mature and works. It's just a matter of how many of these use cases, high-volume monotonous where's my stuff? How do I change my password type situation? On the WFO side, there has been a progression downstream. In the past, you would -- maybe wouldn't have even considered workforce management, one part of the WFO solution for less than, say, 100 agents. Now maybe it's 50 or even somewhat less with a lot of benefits. And we've got a really crack team that is working the base. But we have not quantified that, and we know it's an upside. It's a tailwind, but yet to be determined precisely how much.

Got it. And another growth opportunity aside from ARPU is international. So I know that you talked about, at your last Analyst Day all the investment that you guys are doing in data centers outside the U.S. and an expectation for revenue to be in the mid- to high teens on the international side by 2026. So I think historically, the international has been a weaker area for Five9 relative to some of its competitors. So maybe you can just talk about what you are seeing in terms of the pipeline and like the exposure that you have today? And what else you're doing aside from just data center expansions and getting that presence there in order to facilitate growth in some of these regions?

Yes. Thanks, Taylor. And if you look at our international expansion, historically, as you mentioned, we haven't put a lot there because each -- when you look at the investment of each additional dollar, we've been having such success here in the U.S. with pretty immediate, predictable returns for every head we add into the U.S. We know what we're going to get back. And in the international markets, there's more investment. It takes longer. You got to build a brand. You got to build a base. You got to put it in infrastructure. There's just costs associated with that. The great news is we've made those investments. We've done so over the last 3 to 5 years putting in the infrastructure and establishing ourselves setting up channels and now it's execution. And so right now, we're seeing tremendous growth. The CAGR, just over the last 2 years that we mentioned for EMEA, is 167%. And then LatAm, 128% and the next frontier being Asia Pac for us. So the great news is we're now in execution mode. Those investments are behind us for the most part, and we'll reap the benefits. Part of it is setting up channels. Part of it is putting established personnel in those markets, understanding the data sovereignty and compliance requirements in each of the countries so that we know where we can be most effective in providing our solutions and going to market there. And so that's why we've made a concerted effort to increase the headcount and focus there, and that's why you're seeing the uptick. It will be slow like you said, to mid- to high teens by 2026 because we're -- it's competing with a great growth that's happening here domestically. But we see that as a great growth engine, especially when you recognize that more than half of the agents in the world are outside the U.S. It's important for us to be there. But also bear in mind that even though we show 9% of our revenues being international, over 30% of our seats -- of our actual users are international or outside the U.S. So when we sell a multinational global company, we measure it on the bill to address and we typical to the headquarters. And even if they have agents spread throughout the world, that shows up as a 100% domestic revenue, even though it's really not. So again, over 30% of our business is actually outside the U.S.

Got it. That's helpful. And then maybe in the last 5 minutes here, and to anyone listening in, if you have a question, you can always click the Ask-a-Question button in the upper corner in order to ask one. But maybe in the last 5 minutes here, maybe we'll turn to some financial-related questions. So Barry, if you look at the guide that you guys provided for 4Q of this year, I think, historically, pre-COVID, your 4Q guide has always sequentially implied anywhere from 2% to 4% quarter-over-quarter growth. And I think the 4Q guide this year implies 7% sequential growth and even as we look into the guide for next year, the 24% that you gave for 2022 is higher than it's typically been. So I guess, how should we interpret this change? Is this largely a reflection of what you're seeing in the pipeline at a more positive view on the demand environment, or a change in guidance philosophy? And I'm curious how we should interpret some of that?

That's a fabulous question. And I'm going to begin by making your point even stronger. So yes, the 7% is not only meaningfully higher than the 2% to 4% sequential growth that we've done in the past, but it's also the highest ever sequential guide that we've given. And when you translate it into year-over-year despite the tough compares due to the onetime COVID benefit, the 29% that results in is the highest ever guide for Q4 and our second highest guide ever for any quarter and year-over-year. So the situation is that it's an situation. There has been a strengthening in our business, they move up market and the other factors. And at the same time, no change whatsoever in the guidance philosophy that we so successfully followed for many years. That's not changed either. So the growth rates have gone up as we've gone more and more upmarket. And also we had the onetime benefit from COVID, which we are retaining. And so that's the answer. It's an situation.

Got it. And then maybe as a second part to that question. Curious like as you move upmarket, more of the base becomes enterprise, subscription becomes like a greater portion of the mix, does that potentially reduce some of the like variability, right, that you might have seen in outcomes in the past? Meaning that you could have potentially like greater comfort in some of the assumptions provided in the guide. Or how you think about growth as you look ahead?

Yes. As the years unfold, that would certainly make it somewhat easier. But we have a good degree of predictability in our business where it broke down fairly substantially was during COVID. We were one of the companies that managed to retain our guidance during COVID. We also one of the few companies that went out and quantified the benefit overall in our corporate growth rate to be somewhere in the mid-single digits. But we -- in the uncertainty, we became more conservative and perhaps overconservative and resulted in unnaturally large bleeds. But aside from that interlude, the predictability is good and will likely, to your point, get better. But as they -- as a business -- as an installed base gets bigger.

Got it. And then maybe turning to margins. So we estimate that the 2022 guide implies an adjusted EBITDA margin of I think 15% to 16% based on what you guys have said, which is down from this year, but the 2026 guide implies an adjusted EBITDA margin of 23% plus. So I guess to get there, Five9 would have to add a decent level of margin expansion each year beyond 2022. So can you maybe talk about some of the drivers and the risks there? And what are the areas of potential efficiency gains that we could see over time and where you have more or less conviction?

Yes. Thank you. So a major driver and a minor driver. The major driver to the improvement is gross margin expansion from current 64% to 70% plus. And I just want to remind the investors on the call that this is not a voyage of discovery for Five9. Of the -- we've expanded EBITDA margin many -- by some -- I think it's 46 points since we went public, it is, I just checked. And out of that, 13% came from gross margin expansion. And so -- and what it is it's mainly the scaling of our subscription revenue against fixed and semi-fixed costs. Those costs are largely fixed. There is an exception for the carrier cost, but that's what's been happening in the past and likely to happen in the future. We also expect to get a slight improvement in our professional services. Nothing dramatic and is only a small part of our overall business. And also, the usage part of our business, which is a little under 30% of our recurring revenue, which is 92% of the total, that does happen to grow slower than our subscription and has lower gross margins in the 50s rather than currently in the 70s. So gross margin expansion, 6 percentage points repeating what we've done in the past. And then the other one is G&A leverage. We've now gone 28 consecutive quarters with year-over-year declines. And that is something that may not happen every quarter going forward, but it's likely to continue as a trend.

Perfect. Awesome. Well, with that, I know we're slightly over time. So maybe we'll just stop it there. Barry and Dan, thank you so much for joining. I learned a lot and great insights today, and to everyone listening in on the line, thanks for joining as well.

Thank you. Thanks a lot.

Thank you.