All right. Welcome, everybody, to the fourth day, the afternoon. We're winding down, but I'm very excited to have Five9 with us here today. For any important disclosures, please see the Morgan Stanley website at morganstanley.com/resources disclosures, or talk to your sales representative. I'm Meta Marshall, for those who don't know me. I cover the communication software space here at Morgan Stanley. We're delighted to have Barry Zwarenstein, CFO of Five9, here with us today. Barry, I think you need to start with your own safe harbor, so I'll hand it off to you to start before we take questions.
Great. Great, dueling safe harbors. Before we start, we'd like to remind you that the forward-looking statements made during today's discussion, including those regarding future events, expectations, projections, beliefs, et cetera, of our company, product developments, AI and automation, and potential growth drivers, are such statements or predictions should not be unduly relied upon by investors, may differ materially, and please refer to our most recent Forms 10-K and 10-Q for factors that could cause such a change. And just one thing, if I could just interject. Some of you may be wondering why I'm sitting up here alone rather than with Mike and/or Dan. It's not that they've lost interest, far from it. It's just that we have a long-planned customer advisory board and analyst relations meeting down in Mexico, and I left early to come here, and they stayed on. So, that's why they are not here.
I'm sorry about that.
I'm delighted to be here.
All right, so from an investor perspective, we saw multiple swings in sentiment around Five9 and contact center throughout 2023 as to whether or not kind of CCaaS players would be winners or losers as AI is adopted. How have your conversations with investors changed relative to the time, this time last year? And what do you think is still the most misunderstood piece of the Five9 story today?
Yes, it was a very hectic time towards the end of 2022 and the first part of 2023, when, you know, the thing that we had been beating the drum on and expounding upon, namely the benefits of AI and automation, all of a sudden, we were a loser. Despite the fact that, you know, our head of AI, Jonathan Rosenberg, a luminary in the area, had had the title Head of AI and CTO for a number of years, despite the fact that we'd shelled out serious money to buy the leader in IVAs back in November of 2020. So all of a sudden, we're a loser. You know, somebody must have pulled a Moses on us, so suddenly we can't go into this promised land of automation.
Fortunately, on June 5th, 2023, we believe the team flipped the script. We flipped the script by having this webinar, which had three of our key customers there, who said, "You know, agents are not going to go away. In fact, we continue hiring. If there is a diminution, it's not going to be that material. And we're really excited about the benefit that you can get from AI and automation." Our 2 internal experts, namely Jonathan Rosenberg, who I just referred to, our CTO/Head of AI, and Callan Schebella, a guy who founded the IVA business and was the CEO before we bought it, the two of them explained very clearly to the investors the importance of our platform.
The platform that has the integrations into all the back office systems, the platform that has all that reporting for consolidated across all the multi-channels, that has the security, that has the compliance. You don't just put in a little point solution and expect that that's going to work in an environment that people want to have omni-channel at the end of the day. So, right now, the AI and automation is on the march. Real results, real revenue, it's beyond the hype. Just quickly to throw out some pointers and some of these we'll come back to later on, I'm sure.
Yeah.
On the IVAs, if you look at a chart of the IVA ports in Five9, it looks like a hockey stick. If you look at Agent Assist, that's our fastest-growing booking SKU in Five9. Admittedly, from a small base, but still the fastest growing. If you look at Agent Assist, one person sitting next to me at our customer advisory board, exact verbatim quote: "It was like night and day, the difference after we implemented the summaries in terms of the improvement." So, we're seeing a really good traction. In terms of misunderstood, I think post that event, we've been pretty clear, consistent, and steady in our communication, so I don't really have anything to add there.
Okay. So I mean, in terms of maybe given you already had a head of AI, you knew you had hired somebody who had a lot of experience in this space, maybe the answer to this is, it hasn't, but it seems as if the market has changed their perception on how big this opportunity could be over the last 12-18 months. And so just maybe contextualize in terms of ARPU or cloud penetration or overall seat count, just how you see AI impacting kind of your opportunity set.
Fair enough. So You mentioned the word contextualize, and, and let me make it, just pick it up before we go into AI in particular, because it's, it goes together. So the three drivers now as to why the shift is taking from on-premise to cloud is taking place. In fact, it's probably at an inflection point, and it's still early days. The first of the three is the fact that we've got the legacy vendors no longer coming up with, in most cases, with new on-prem platforms. So the people who've got these mission-critical systems, they interact with consumers, have to do something or pay a tax in terms of becoming increasingly obsolete. The second is the importance of customer experience.
The metrics that things like customer satisfaction and propensity to leave and NPS scores, people are valuing those, even if it means a higher cost than the traditional metrics, like average handle time, first call resolution, and so on. And then the third one is, importantly, is the AI and automation, which you theoretically can do some aspects of it in the cloud, but you really need to on-premise, but you really need to be in the cloud to be able to do it. So in terms of the traction that we're seeing in terms of seats and so, I'll come to seats at the end. But in terms of the metrics that we shared on our last quarter earnings call, 17% of our enterprise bookings came from separately priced SKUs in the AI and automation.
By the way, virtually every single, if not every single enterprise deal, had some discussion about AI and automation in it. I mentioned Agent Assist. Each quarter in 2023, it was doubling, growing progressively, and by the end of the year, in the fourth quarter, it was 6x what it was in the prior quarter, Q4 2022. I mentioned already that Summary is the fastest growing booking SKU within Five9. So we're getting a lot of traction. The ARPU on this is pretty high. Commonly, we would sell just the subscription part of a regular non-AI seat at about maybe $130 per seat per month, and then the usage will be on top of that. Here we're seeing regular handles of 300, 400, even 500 per seat with the AI automation included in there.
Okay, perfect. And so how are you thinking about end market demand in 2024 in the context of these AI tailwinds and then lingering pressures with consumer discretionary customers? You know, within your kind of 16% revenue growth guidance, just what are the, some of the key areas that could either drive upside or downside to your expectation?
Important question, Meta.
Yeah.
And very topical. So in order to answer that, there's level set in terms of Five9 having two parts of the business: the install base part of the business and the net new part of the business. In terms of the install base part of the business, it's been a reasonably good situation with 16 of the 17 verticals that we track. The Q3 to Q4 growth rate in 2023 was similar to what it was in 2022, although that was subdued versus some of the prior year histories, given the macro. There is, however, the consumer, which is a big part of our overall mix, the third biggest vertical. It's the most, by far, the most seasonal in the fourth quarter. And we saw headwinds there, pockets of weakness, where we're not like some of our other competitors in the industry who are living in this, in this world where they're impervious to what's happening in the macro.
We saw consumer discretionary, subset of the consumer for the first time since coming onto the platform, the bigger accounts, actually contracting from Q3 to Q4. Never happened before. So we had some heavy sailing there. The good news, just by the way, before I come to the net new side of it, is that the logo retention has been excellent, in the mid-nineties for the enterprise part of the business. And so when inevitably the American economy turns around, we're gonna take, we're gonna benefit from that, in a serious way.
On the net new side, much brighter story currently. We currently with reference to not net new, was to reference to what's happening in the install base. In terms of the net new, the numbers speak for themselves. RFPs doubling year-over-year in the fourth quarter. Pipeline for the enterprise and strategic at record levels. The best quarter ever for enterprise bookings. So, clearly something really positive is happening there, and it is this backlog that we've built up because these things on the commercial side, they're going pretty quickly, but on the enterprise side, it can take, seven months is what we model, but it could take much longer if it's an international deal or a big deal. And so, we have this backlog plus some what we call go-gets.
That'll be orders, that'll be imminent, and that'll be delivered before the end of the year, that allow us to have high visibility into what's gonna happen later in the year and which gives us the confidence that allowed us to give the guidance we did for the 16% growth, which implies, given the H1 guidance that we gave, a revenue re-acceleration in the second half of 2024. If you would indulge me just a moment longer.
Go for it.
There's another way of looking at it, just to simplify it arithmetically. In surveys, we'll talk about Dollar-Based Retention Rate, I'm sure, later on. But for easier communication arithmetic, if you assume that it's 110%, that 10% growth versus 2023 equates to about $83 million based on the recurring revenue, which is what the Dollar-Based Retention Rate is calculated upon. The other $62 million comes from that backlog I was just talking about. And between those two, that gives us the $145 million dollars that is the year-to-year growth implied by the 16%.
Okay, that's helpful. And so maybe on that DB&E question for a minute, you know, you just, in your previous answer, spoke a lot about, you know, net new adds traction is incredibly strong, but you have some kind of headwinds within the install base, which makes a lot of sense. But, you know, you've called for DB&E to inflect positively in the second half of the year. And so, you know, just what is it that kind of gives you that confidence that the install base can kind of come back in that second half, or is it kind of layering in some of these, this pipeline?
Yes. So it's a layering in the pipeline that is the main thing
Okay
Not the upsell so much. So, let's real quick go to it. We gave guidance when we gave the third quarter call that the Dollar-Based Retention Rate in the fourth quarter would be flat or very slightly down, and that's what happened, despite the somewhat weaker install base. And the reason is that we have great visibility into the spot rates and into the stratification in terms of big, medium, small businesses. And we knew that our second biggest customer, which was basically a zero a year ago, it's a healthcare industry, healthcare industry company, would be ramping. And that's indeed what happened and allowed us to come in at 110, just as we guided to. We're expecting Q1 to be lower, slightly.
For H2, we've said explicitly, we expect an inflection point based upon what we know is in the backlog.
Okay
For it to go up. Before we leave this, the key thing to keep in mind over here is the bigger the company, our data shows, is the higher the Dollar-Based Retention Rate, and the bigger companies are growing faster. And this also gives us the idea or the plan to have the Dollar-Based Retention Rate go up to high 120s by 2027.
Okay, perfect. We just kind of talked about you're seeing AI in almost all of your new deals, and I think we can understand that. Just what are some of the efforts or initiatives to kind of upsell your installed base?
Yeah. So, we've not had, frankly, as much success here as we initially would have expected. We've done two key things, which we believe will make that be much better in 2024. The first of the two, and I can't help but smiling when I think about it, is that we brought in a new head of this business who is tested and proven and done huge things for us in the past. So he's back now and doing this, and we've infused into his organization not just the farmers, if you will, but people who wake up each morning with a quota mindset, the hunters, if you will, and the working together into that organization. The second of the two things is we have something called try and buy type program.
The try and buy type program is we're going to... Well, you can imagine what it is. We've come out with three different trials. The first one was Five9 Analytics, which is focused upon the operational side of the business, somebody who's trying to see what handle times are, resolution times are, et cetera, what it can done to improve it, so it helps with the analysis of that. The second one is post-call summaries to increase the quality of those summaries and reduce the time needed to do that. Very strong results there.
And the third one is something called Five9 Insights, using LLMs to address the more business-oriented part of the company, management, where they want to be able to see where, you know, which are our longest calls and persistent, the most numerous calls, to see how we can improve that. Now, all three of these products require only minimal PS or maybe no PS, which means they can go in pretty quickly. So we're kind of hopeful that we'll see some results on that.
Okay. Is there, you know, you just broke down kind of the multiple different products. You gave that 7% of subscription revenue, and 17% of enterprise bookings is attributable to AI. Is there just a way to kind of break down or contextually kind of give a sense of how much of it is some of these kind of IVAs versus Agent Assist technologies and, you know, any margin impact we should be mindful of as these products grow as a percentage of revenue?
Yeah. So the leader in terms of revenue, Meta, is the IVAs. It was the company we acquired. We had the, the leading product and, and so on. The Agent Assist is gaining ground very rapidly. It's becoming more and more important. When you think about it, if you're going to be offloading the more mundane, repetitive inquiries off into the IVA or into the chatbot, what's left? What's left is for the agent is the more difficult and challenging cases, so they need more assistance. So Agent Assist is really showing some really solid traction. It's things like real-time transcriptions. Think about maybe an agent who's hard of hearing, real-time guidance to do the coaching and the training, and the post-call summaries that I mentioned a moment ago, and those are all seeing pretty good traction.
Okay. And then how are you seeing the conversation around pricing evolve in this market? You know, a lot of your competitors doing a kind of combination of seat-based and transactional models. You know, I know there's also some customers who want something that's, where they know what the price of it is going to be. Just how are you seeing kind of that conversation, evolve with customers?
Yeah. So, we try and accommodate and be responsive to what the customer want wherever we can. Right now, we have three different basic structures in terms of for the AI and automation products. For the IVA products and the chatbot, it's based upon ports, essentially capacity type- pricing. For Agent Assist, which goes to the agent, it's therefore linked to the number of seats. It's more uplift to the seat price. And for WFA, workforce automation, the automatic surveys and so on is based upon usage. Now, having said all of that, we've got an active program to bring in some more options. It could be more usage-based, although we would end up back solving to get to a similar situation to what we now have. Bundles, if you're a new company starting out, say, a mid-market company, which ones do I buy? Well, we can give you a collective. But all of these will take some time to go through all the legal and educational and systems type influences to.
Any margin impact to kind of be made.
No , the key thing is the one there about if we're selling it on a usage basis or per seat basis, we will back solve to come up with a similar thing.
All right, perfect. You know, on past earnings calls, Dan has noted that kind of virtual agent solutions have about a 50% completion rate. And most recently, kind of noted that there's a lot of volatility to that number based on kind of what tasks customers are giving, kind of these AI agents. What has the past year and kind of the boom of AI tools taught you about how to guide customers as to where AI can be most effective?
Dan must have had two Red Bulls that day with a 50%, because that's pretty high. But it would be delightful because we make more money the more we automate.
Yeah.
It varies tremendously by customer. If the customer comes to us and says: We would like to, we'd like to have one of your bright, shiny IVAs. So what - how much are we gonna save? And the answer is, we don't know. We need to go in and figure out what the best applications are and build, you know, heuristically better and better solutions for them. So the, it's, it varies so widely. I mean if I was pinned down to what I've heard anecdotally, it would be more in the single digits-
Okay
Rather than it would be in the 20s or 30s.
Okay.
Or even fifties.
But do you feel like you guys are now kind of getting a context of, if a customer asks you to automate a certain use case, to say, "Hey, that's not the best use case," or they're willing to try it?
Oh, no, they, they all want to try it.
Okay.
It's just what they have, which particular products and what the expectations are in terms of the return upfront.
Okay. So you can better guide them to what the return might be.
Yeah.
All right. So professional services has been a bottleneck to customer onboarding. You know, it's also been kind of a chokehold on kind of margin accretion. Just how has thinking evolved on keeping professional services in-house versus leveraging partners?
Yeah, there's been a definite evolution. We used to do virtually, w ell, we did 100% of our installations in-house. And part of the reason for doing that is that we believe passionately that you only make that first impression once, and we want to make sure that we have those NPS scores in the 80s and 90s, where we know that some of the others in the industry are in the 20s and 30s. And because we're building trust and by that consistency of implementation. But under Mike's leadership, under Project Pull-Through, we've aggressively gone in and worked on certifications, maintaining the quality through the certification, and just especially internationally, outsourcing some of that implementation to outside parties. And the metrics speak for themselves.
This last year, certifications tripled year over year, and the number of implementations worldwide doubled. Obviously, we currently have high single digits, low double digits losses on professional services. We don't mind too much on that because we will get the software quicker, because we'll have the ability to do it well and quickly. But at the end of the day, the extent that it's done more by outside will be less of a drag on the margins. In terms of our own improvements. As we scale up, we've had some really big some of the biggest contact center cloud contact center deals out there, and some of the ones that are almost the biggest, and it just takes a lot of scaling to be able to get past that.
This acquisition we made of Aceyus, which at its core, is helping to move efficiently disparate systems on-prem to the cloud, so that the client doesn't even know that there's a change taking place. That too, is gonna help us improve our professional services efficiency.
Okay. You know, now that you're bringing in these partners to kind of help with professionals or help with the implementations, just is the makeup of some of those deals different than o r are they bringing you different kind of customer types, than you had access to previously?
The partners, the only thing I would jump out there is, the increasing importance of systems integrators in terms of, these bigger deals. They tend to be involved a lot more.
Okay, got it. And then just how do you see the competitive environment today, both just from CCaaS vendors, but from just the number of new entrants in this space, both from larger platform vendors to kind of smaller conversational AI specialists?
Yeah, it's become very popular to become a contact center service, cloud contact center provider. I mean, the market's big. It's attractive on the top and bottom line. It responds to the CX focus of companies. But let me go into the more specifics of... I'll start with the main landscape, the donors and the competitors. In terms of the donors, traditionally it was Cisco and Avaya. The top position within Five9 would be changing year on year, depending upon that particular year. Genesys now, with the end of life of their product as of October 2022, November 2022, is now also a donor. The competitors are NICE, Genesys and ourselves. We're not involved in all of their deals because some of those are involved in conversion of their base, which sometimes we don't see those deals at all.
When we do see them, as best we can tell, and we track it very closely and report to our board name by name, we have win rates of over 75% in the deals that we are in with either or both of them in. Off on the side is Amazon, which doesn't have a customized, ready, it doesn't have a product per se. What they do is they involve an SI or their own resources to help customers come up with a highly customized solution over a period in two or three years. The way we talk about it is that if we and Amazon are on the same deal beyond the first, at the most second round, one of us is making a mistake because we're offering a different type of solution.
But they are a very credible competitor. Then if you think about prospective new entrants in the traditional sense, not a point solution, but an end-to-end contact center, Microsoft, you know, makes, has, has made announcements about wanting to get into it. They have two of the key elements, both to the extent that is linked with UC, they've got Teams, and then they've also got the Dynamics side of it with the CRM. In the meantime, though, we don't see any... There's no product that we're competing against, so that might come later. In the meantime, our relationship with them, both on the Teams side and the Dynamics side, is excellent, really excellent. You've got Zoom as well.
Zoom has got a lot of talent, a lot of resources, a lot of determination, but doing what we do is not, it's pretty difficult. There's been no new entrant of scale into the industry in the last 15, 20 years, and that's for a reason. It requires very, very challenging development both on the voice side and the data side and on the software side. Time will tell. And then finally, on the point solutions, you know, at the end of the day, we offer our own chatbot. It's got all the advantages of being integrated with IVA because it's built on the same studio platform.
But if, nevertheless, somebody wants to go with a point solution, they will, they will do so, but they're gonna need to come to us if they wanna be part of that platform. And at the end of the day, most people want to have an omni-channel solution.
Okay. Before I keep going with questions, are there any questions from the audience? It's a quiet crowd. It's after lunch. We've lulled them into silence. So, you know, Five9 has been active in the M&A market with a lot of smaller tuck-in acquisitions. You know, can you just walk through some of the skill sets the last few have brought you and just how you think about organic and inorganic development, particularly kind of at a time of inflection within the space?
Yes. So let me start with the second part first. We've got a preference for organic, but as we demonstrated, as you alluded to, we've made several tuck-ins. We've got a really high-stepping corporate development function, headed up by Jim Doran from Cisco, who they know how to do acquisitions. And so that's our mindset, organic, but we'll do the, we will continue to do tuck-ins. We made four, and arguably each one has been more successful than the prior one. The first one is, we made a company, bought a company called Whendu. What they did was workflow automation, and that, as I mentioned earlier, is now one of our eight key products in the AI and automation front. The second one then was, a company called Virtual Observer.
This is our internal WEM, Workforce Engagement Management solution, that we're constantly improving and, is doing, very well. And then, potentially the star, up until recently, of the show would be the IVA acquisition, which was, a company called Inference, that we bought in November of 2020, and that sort of demonstrated our commitment to this AI and automation going all the way back to that. And it remains one of the leaders. And then, finally, most recently, last August, I think it was, we purchased a company called Aceyus, which I made reference to. And at its core, they help companies make the migration. These are major companies, as we got to know them, as we were doing some of these big deals, make the migration seamlessly from, in terms of data visibility, and non-siloing from the on-prem to the cloud.
Okay. Maybe something that I find when I talk to investors kind of misunderstood is kind of the usage piece of Five9. And just, I know that that was kind of a headwind to growth over the last year, and so maybe just kind of giving some contextualization of how you see it and how it fits into Five9.
Yes. So it's important to be very clear when we're talking about something that we understand what we're saying to each other. No, I wasn't I just want to be sure that everybody understood. When we talk about usage within Five9, we're talking about the long-distance charges. And, and just to make it crisp, if you look at Q4 2023, the breakout between subscription, recurring subscription services, was 77% of the total. Those long-distance charges, plus some other small charges, was 15%, and the complementary 8% from professional services. Now, as you alluded to, Meta, that usage component declines. When we went public, some 10 years ago, it was 36% of the total.
Now it's down to 15, and it'll continue going down by 2-3 percentage points per year as a proportion of the total. Now, there's a number of smaller reasons why this is happening, which I'm not going to take the liberty of going through with you right now. But there's one big one, and that is that some of these bigger companies that we're landing are bringing e ither come to us through partners like AT&T or British Telecom, or major corporations that have their own unexpired contracts with the Verizons of the world or whomever. For example, again, that healthcare company, our second biggest company, we don't get any of the use of that company's usage. So that day will continue to go down over time.
Okay, perfect. And then just maybe as a last question, just how you're kind of thinking about operating leverage within this business, and just of what you're holding yourself to in terms of kind of threshold for investment?
Yeah. W e've got so many really, really good opportunities particularly in R&D. Going, in no particular order, further innovations on AI and automation, more going more globally. It's been a real bear in terms of India, but we're pretty close. But there's other parts, other localizations and other advances. FedRAMP, massive investment. It's gonna take a couple of years before we can really see the fruits of that, but it's a big market. Think of all the hundreds of millions of Social Security recipients and veterans and so on. But at the end of the day, so we want to make those investments. In 2024, I wouldn't see very much leverage at all. We are trying really hard not to beat the $2.16 of EPS guidance that we gave for 2024.
We've tried in the past, but we've always beaten on the bottom line. We don't want to. We've got too many good opportunities. Longer term, 2027, our model out there, which has become increasingly aspirational because of the macro, but is out there, we want to be at 22% EBITDA, which we've been at, flirted with in the past.
Great. Well, I'm excited to keep following this story, and thank you so much for being here today.
Thank you, Meta.
Sorry to pull you out of Mexico.
All right. Well, thank you very much. Thanks a lot.
All right, thank you.