Thanks, everyone, for attending this session. For those that don't know me, my name is Taylor McGinnis. I'm on the software equity research team here at UBS, and in this session, we have Five9. So we have Barry, the CFO, and Dan, the President. So thank you guys so much for joining.
Thank you.
Barry, I didn't forget. Over to you.
Thank you. Before we start, I want to remind you that we've been making forward-looking statements about future events, trends, expectations about the industry and our company, including product developments, AI and automation, and potential growth drivers. Such statements are predictions and should not be unduly relied upon by investors, and actual results, and it may be materially different, and we undertake no obligation to update. Please refer to our 10-Qs and 10-Ks with the SEC for further elaboration.
Perfect. With that, let's kick it off. So maybe a good place to start would just be in terms of what you are seeing in the macro. So I know you flagged some softness and seasonality among some verticals, so maybe you could just elaborate in terms of, you know, what you saw throughout the quarter and into the beginning of Q4 there. And then as a second part of the question, I know, Barry, you flagged the consumer vertical as being part of the prudence and the guide for Q4 So was that just more of a reflection of things that you guys are, you know, hearing in the environment? I know you mentioned some statistics, or are you actually seeing more of that incremental weakness in the consumer vertical?
Fair enough. There are two questions, but they are very tightly integrated. So we track 17 different verticals within the company. The biggest is healthcare, the second biggest is financial services, and the third biggest is consumer, which is the most seasonally volatile part of the business. Going back first to Q3 of 2023, for the quarter as a whole, we saw the consumer actually pretty normal, for the taking the whole quarter at its entirety. However, we did see a continual deterioration within the quarter. Let me be more specific. One way to externally, w e have our internal data, of course, and which tracks the external data very closely. The way you can track it is through JPMorgan Chase credit and debit card data on consumer discretionary spending. Why is this important?
Because at the end of the day, our people are looking at transactions that come into the contact center, and, if they're less spending, there's certainly going to be less, events and less transactions and less need for agents, and hence our subscriptions. And, we saw within the quarter that it went 5%, 3%, and 1%, July, August, and September. And at 1%, given that those are nominal numbers, it's actually negative. So for the quarter as a whole, it was okay, but we started getting nervous because things were getting progressively deteriorating. And then we looked at other things. Credit card data, credit card delinquencies at 20-year highs, automobile delinquencies at 20-year highs.
JPMorgan Chase review of intended spending, 42% of consumers said they plan to spend less this year than they did last year. And so the debate is raging about what is the consumer, is it fine or not fine? But those were telling us, "Be careful." And what we did is, we decided that we were not going to put through any of the $6 million approximate beat that we had in Q3, and take a more prudent stance with respect to Q4. I'll make two more comments. Since then, two other pieces of data that really resonate with us. The Bureau of Labor Statistics does its own definition of retail. It's not identical to ours.
It shows the hiring that took place in October of 2023, for November and December of 2023, that those retail hiring was lower than even the anemic October 2022. So we're thinking, you know, maybe we were actually pretty smart to do what we did. And then, just in The Wall Street Journal this last week, Deloitte came out with four charts that show how the consumer is getting tapped out. So this is not an environment we feel to be, brilliant and effervescent about consumer spending. There are pockets, for example, airlines are doing very strong, but we only have one airline in our, currently in our, portfolios.
Perfect. Oh, go ahead, Dan.
What's up?
Oh, I thought you were gonna add on to it.
Okay.
Oh, okay. Well, on to the next. Well, maybe turning to, like, despite the macro, the $1 million ACV plus deal momentum has been really strong. And I know one pushback that sometimes we hear is, "Okay, all of your bookings activity looks great. Why, why is Five9 not growing faster?" Right? And I think even some of the math that we've played with, you know, depending on when you assume these deals ramp, like, you could get to a scenario where Five9 x some of these deals or ramps are growing in the single digits. Now, I don't, like, think that's probably the case, but I'd love to hear from you guys in terms of how should we think about those deal ramps? Are those linear? You know, when from deal signing do you start to see those flow into to revenue?
Maybe you could just give us a little bit more color on that trajectory.
Yeah. So that really, we know we talk about the big mega deals and the, the multimillion and tens of thousands of agents, and that those are the shiny objects, and everybody wants to know how those roll out and turn to revenue. But really, the bread and butter for the handful of those that we've had over the last 18 months or two years, there's probably 10x that many that we've had in that $1 million-$5 million range, and those tend to turn over quickly, more quickly to revenue and then ramp up to full strength. And so that momentum is strong, and it continues to be a nice, steady flow of that business. And they typically will have a sales cycle of anywhere from nine to 18 months.
They'll ramp up and start to deploy within about four months, is what we model, and then they have about a three to four -month ramp from there. There's exceptions to that, of course, but that's the average that we use for the model.
Yeah, and then maybe-
Let me just add to that as well, because you make me nervous with that, when I comment about the potential growth rate and you playing around with the numbers.
Yeah.
When you look at the recurring revenue in the Q3, it was 18%. PS is a little bit weaker. And by the way, it was also 18% in Q2. And frankly, when you get, so when you think about Five9, think about businesses that come from new logos coming on the platform and ramping, and the other is installed base. That one's doing well. We'll get into that, or very well, and the other one is the installed base, which is having these headwinds that we talked about earlier. But generating 18% in that environment, we feel actually pretty chuffed about it.
Yeah. Awesome. And then maybe on, like, the new logo ramp, just to add an additional question on what you said, Dan. In terms of the adoption, I'm sure it varies for every customer, but does that tend to be linear? Does it tend to be more back-end loaded? How do those, i s there any average in terms of how that looks?
There are. I wish there was a pattern.
Yeah.
Every customer is different, depending on their, you know, do they have three large contact centers? Do they have 30 smaller, spread out ones? Do they have - do they want to roll out department by department or location by location? Or do they want to set the whole thing up and do a flash cut? Yes, all the above occur.
Yeah
We have to be prepared and work with the customer. The good news is, our professional services team works with the customer to figure out what does that forecast look like. So we get visibility into it pretty early on.
Yeah
as to what we expect to see, and then try to keep them on track to be able to roll out at that, at that pace.
Yeah.
So it's a comfort to give us visibility into that backlog as to when that's going to convert to revenue. And so that's one benefit we gain from it.
Yeah. On that last point, on the professional services side, so I know that's been a big area of investment for you. When you think about that, combined with the acquisition that you just had, does that have the possibility maybe to accelerate some of the ramp of these deals? And then on the flip side, is macro at all having any, you know, impact, where are customers hesitating, maybe pushing things out a little bit? Would be curious if you're seeing anything on that front.
I'll answer the second part first which is no, no hesitation there at all. The macro doesn't impact. Once they purchase, they want this installed, and so they can take advantage of the efficiencies they're going to gain from the new platform. And so they're trying to go faster. We'll go as quickly as they can, so we're never constrained by our professional services not being able to handle the load, if you will.
But back to the Aceyus acquisition, if you look at that, that was one of the primary reasons for the acquisition, was to be able to help customers make that transition more smoothly. And it's really from the standpoint that they would put up the objection on occasion, that if you think about what Aceyus does for us, it takes information, metrics from the legacy systems. So if you think about our industry, you have the legacy, the Cisco, Avaya, Genesys, and others that are out in the market.
In these large companies, the, you know, the largest of the big companies, they tend to have all three or four platforms, right? They've done M&A of their own. They've acquired different platforms. And believe it or not, all of us measure all of our metrics differently. So even something as simple as average handle time, you'd think that's straightforward. But we all— Do you include the ring cycles? Do you include the hold time? Do you include the transfer time? There's all sorts of— so everybody measures things differently.
The most difficult thing for an enterprise, if somebody's running the operation and they're trying to assess the performance, to measure Topeka, Kansas, contact center on a Cisco system with a Bakersfield, California, on a Avaya system, you can't measure performance very well at all with the data that comes out of those systems, 'cause it's, it's apples and oranges. What Aceyus does is it takes that information and normalizes it and centralizes the data. So when I've gone into customers using Aceyus, they say, "Aha!" I said, "Is that coming off of a Cisco or an Avaya?" And they said, "I don't know, and I don't care.
I'm running my business." The reason that's important is when they then make a decision to go with Five9, they can make the migration, include Five9 into that equation, and still operate the business as usual, with consistent apples-to-apples comparison of performance. If we didn't integrate with Aceyus and thus acquire them, they would be, "Whoa, whoa, whoa, don't... You can't touch these sites, because then I won't be able to run my business and compare what's happening in them," and that would cause delays.
So this is actually a way to help them transition much more smoothly from their legacy systems over to the cloud. And it also gives us a leg up with those large enterprises that are already Aceyus customers, 'cause they have multiple systems. Because when we go in, we're in the pole position, not only do we have a contractual arrangement already with them for that piece of business, but we also, there's a fear of, "Gee, are they gonna continue to integrate with a different cloud system if I was to choose somebody other than Five9?" Well, there's no guarantees there, so that's a risk factor that we take care of for them.
Yeah. All makes sense. Barry, I'm gonna go back to something that you talked about earlier. So you mentioned recurring revenue growth in Q3. And I think on a sequential basis, that was actually pretty much in line—sequential growth was pretty much in line with what we saw last year, and it seemed like the drag, really, on revenue came from professional services. So when we think about the implied guide down that we saw in Q4, can you maybe give more color in terms of what the pieces and contributors were to that? Was it truly just being more prudent on the consumer side? Were there any deal slippage? Doesn't sound like there were any deal delays in terms of the roadmap of some of those companies. But maybe you could just break that down for us.
It was the consumer and the macro impact, and that's why we need to be careful over there. The professional services is very lumpy. There's. Think of it as just literally scores, if not hundreds, of individual sales orders that have to be negotiated as the scope of the project changes, as new projects come on.
It's purely a matter of timing. If you look back X years, you'll find that there's always one quarter where the PS goes down.
Got it. I'm gonna ask a last question on the guide, a pesky analyst question. But when we look at the cadence of upside last year, it seemed in the H1 was, you know, where you saw a lot of the, a lot of the upside, and now when we look into this year, you're kind of seeing a similar cadence. So maybe that's just by coincidence, but can you comment on that? Is it as you move, you know, throughout the year, you get more visibility to the year-end, maybe the range in terms of, you know, variability, where numbers could go, you know, it starts to narrow? I'd love to get your, your thoughts on that.
Yeah. A few comments. First of all, we're not about to give quantification to upside.
Yeah.
But I will say this, that, it's only in. If you go back further, you'll see that there was variability by quarter.
In fact, if we were to pick one quarter that is the most difficult to forecast, it is the Q4.
Yeah.
Because that is the biggest seasonality.
Got it.
By far.
Got it.
I mean, if we go back pre-macro and pre-COVID, it was the sequential changes. Q3 to Q4 was like, for 8%-12% sequential.
If you go back to 2021 and look at the healthcare. This was a public number, so this is not new information. Healthcare and, consumer, the two, most seasonal, the sequential growth was 24% in those two verticals. So, it takes all of our vaunted FP&A team's, ability to be able to get that close to the pin.
Perfect. Thanks for that. And maybe let's look. Now, enough with the near term. Let's look at the outer years. So you have that $2.4 billion 2027 target. So if you run the math on that based on where you put the guide for next year, I think it implies a 27%-28% CAGR, which is obviously a big acceleration from the growth that you're seeing today. So can you just talk about your comfort level with that? Obviously, you're seeing a lot of momentum on the large-scale activity and, you know, how you think about the drivers of that.
Yeah. I will concede upfront that, given what's happened this year that $2.4 billion in 2027 has become more aspirational. But, there's still a framework where we think it's, certainly achievable. Bear in mind a few things. When the, inevitably, the U.S. economy reverses and picks up real steam unequivocally, we spring-loaded to take advantage of that. Our logo retention on our enterprise business has been excellent, mid-90s%.
We can provision those seats overnight. So when the transactions come in, there's gonna be more agents, and then the hold times will increase, and they want to bring it down, and they'll have more agents. And the second thing is that 2/3 of our revenue comes from enterprise subscription. Not the commercial side, just the enterprise and just the, the subscription. 2/3s. And we have, since time immemorial, said that that business can grow with a three handle. Sometimes it's grown with a four handle and a five handle, but certainly with a three handle.
We did caution several quarters ago that it may, during this macro environment, dip into the high 120s, and this last quarter, it did indeed go into 28%. But when that thing turns, it'll be a large part of our business growing in the 30s.
Got it.
And one more thing to add, as you alluded to on the net new logo side of the business, we talked about our strategic pipeline doubling year-over-year, the number of RFPs up 66% year-over-year, 21% sequential Q2 to Q3. And so those are early indicators of longer term. It takes a while to go through the sales cycle, bring those online, get them ramped up, and have them contribute to revenue. But, if all the other factors remain equal, you know, close ratios and so forth, then, that certainly would fuel additional growth.
Got it. And then on the softer seat seasonality that you're seeing, when we look at NRR, that downticked two points or so this, this last quarter. So where are we in terms of maybe some of that optimization or softer activity, and could you start to see more stabilization in that, in that metric?
Yeah, I'm so glad you asked that particular question, because it really is key to the 2024 guide and 2027. So, the meetings are beginning to run into one another, so please stop me if I've said this already. But, through the Q2, we were saying consistently that the smart money was betting that it would go down. This quarter, Q3, we changed our tune. We said, either it's gonna be flat or down very slightly, and we added, secondly, that you'll see a positive inflection in 2024. And we don't make those statements lightly.
Yeah.
The reason that we see that happening in the Q4, the flat versus or slightly down, is two: one big, one small. The big one is that we're getting easier compares, and the smaller one is that some of the bigger customers are ramping, and they can have a material impact because they need to be in our install base for 12 months, and some of them now are, and then we're getting a tailwind from that.
That's helpful.
Overall tailwinds from that.
Got it. Makes sense. Okay, now let's switch gears to Generative AI. So obviously, that's been, you know, a source of some of the softer sentiment, right, on the stock.
So maybe you could just tell the group why is Five9, you know, the best positioned in order to monetize a lot of these Generative AI offerings. You have the CRM players, you know, having a little bit of overlap functionality, as well as the cloud infrastructure players. So maybe you could just elaborate for people there.
Sure. Yeah. And the AI, you know, when you look at our AI and automation portfolio, we have eight different modules that we have brought to market and that we provide to our customers. And we really pioneered this from a contact center perspective. We went out and we saw this and have messaged about it for five years now, that automation is coming. It arrived. We went out and made two strategic acquisitions, Inference from an IVA perspective, Whendu from a workflow automation perspective. So we kinda hit the ground running with some momentum. And I think our approach is slightly different than others.
Well, we're all checking similar boxes to say, "Oh, yeah, we do summaries, we do agent assist, and we do IVAs and so forth." But the approach we've taken gives great comfort to customers to make a Five9 decision over our competitors. And what that is, is that we've taken that agnostic approach, where we can plug in the different engines. This is really critical, and it's not just to say, "Oh, we're gonna stay abreast of the changes that occur." We can always be putting in new underlying engines, whether they be NLU engines, whether they be speech engines, and so forth, or the large language models like ChatGPT. Who's gonna be the winner in that? We don't know. We probably don't know in five years from now, and our customers don't.
But the comfort we give them is whoever is, we'll be able to leverage and bring them to bear. Examples of that include things like we have a new speech engine, Deepgram, that we just announced last quarter into the platform, that does alphanumeric extremely well. Why is alphanumeric hard for a speech engine to recognize? It understands sentences and long words, but when you have one-syllable digits, letters, or numbers, it starts to listen on that first syllable and has trouble. So most of the speech engines aren't very accurate with alphanumeric. Well, they've mastered it, so when we have an alphanumeric application, we plug that in. That's just one of many examples. So we have many speech engines. Some do foreign languages better than others, some do alphanumeric, and so forth.
And the same thing goes on the large language models. If you look at ChatGPT right now, that's great. We're also finding that enterprises will start to standardize on their preferred engine, right? If they have a preferred NLU that they wanna use, great, we'll work with it. You know, we announced an IBM Watson increased partnership and integration with their solution. We work with Salesforce very closely, the CRM comment you made. We're very complementary. We wanna take our real-time speech, convert it into text, and feed that text, real-time text over to Einstein so that Salesforce can do their magic with it and come back. So it's very important for us to be complementary and help our partners.
The CRM partners leverage their technologies, and they can leverage our technologies. And, and there's always gonna be some overlap.
Yeah.
The key there is it's a small revenue portion, but it pulls through a lot of the core revenue for us and makes a Five9 decision possible.
Yeah. And Dan, I think you brought up an interesting point, 'cause I think a lot of the investor focus tends to be on chatbots, right, and first-role agents, but there's lots of other ways for you to monetize generative AI, right? So you have agent assist, there's sentiment analysis. So can you maybe talk about those opportunities? You know, how big of opportunities could those be for Five9? What are you hearing in terms of adoption and different use cases?
Yeah. So we talked last quarter about an 80% attach rate for AI and automation onto our million-dollar-plus deals, and that's very important. That means they purchased one or more of the 8 modules. But to your point, you've got applications, and for the most part, we've talked a lot about in our industry of, oh, it's either a human agent taking a call or it's a fully automated self-service. And, you know, if you bell curve this out, most applications or most interactions are a combination. You've got some automation at the front end of the call, but during the call, you might also have some automation to help the agent be more efficient and effective, fetching data for them and delivering it back to them, so they don't have to go hunt for the answers, and so forth.
So we're all about, and our industry has been about this long before AI came to the forefront, which is helping clients do more with less, meaning more interactions and delivering a better experience with less humans, but more automation. The more we can add automation to the mix, the more revenue we achieve and we receive from that client. So as we look through the years, there are certain applications that are good candidates for automation, and there's others that are very poor candidates. Straightforward transactions where I simply want an answer to a question that's very binary and very factual, that's great. I may wanna go use an IVA or a chatbot to do that. Others, I want consultation, and I wanna do that. There's plenty of self-service tools that we've used for years.
The website, you know, we all go on a website to book our travel and book our hotels. We don't call the contact center anymore. Well, those contact centers are still, thousands of agents in them. They haven't seen a reduction. So that's that. But when we go to make a change, something's happened to my flight, I've missed it, I'm gonna get on with somebody and figure out all my options and figure out the best option. It's probably better done with consultation rather than trying to do that with a bot. So it runs the gamut. The key there is also helping brands recognize that they don't wanna over-automate or force people to go into the automation. That creates frustration and true, a bad experience, which is what most brands are trying to avoid at all costs.
So they wanna give the choice to the customer. If you wanna use the bot, if you wanna use an IVA, that's great. If you wanna use the website, that's wonderful, but we're here to help you, and, and most brands figure out that the human interaction is very hard to replace.
Got it. No, it's perfect. And on that, so two-part question, but the first one would be, you mentioned the 80% attach rate. Any high-level color you can give us to get a sense of what that might be contributing to the business today, right?
Where that could go in the future, and what that means for ARPU? Then the second question would be, as you're having these conversations, and these customers are adopting and buying some of these generative AI solutions, what is the trade-off that you're hearing them make on the seat and ARPU side?
Yeah. Great question. So think of the as we're implementing the, the AI applications onto the platform, it's typically an uplift, a slight uplift. And so if we're getting $200 per seat per agent, we add in Agent Assist to summarize calls, to provide guidance to the agent. That might be a 10% or 20% uplift to that seat, but it's not, you know, replacing the whole seat count. So the contribution, then when you look across our nearly 3,000 customers, the contribution from a revenue perspective still isn't material enough to break out. Someday we'll be breaking that out as it becomes more, but it's not something that we report on today. But to give you an idea, the 80% attach rate is important because...
A lot of those are just dipping their toe in the water because they're purchasing Five9 for the first time. Their decision was driven by the AI story and the AI solutions that we have. But oftentimes, they're finding a single use case or series of use cases, so it's starting off very small. So not material enough to really impact the ARPU across the whole install base. What we are seeing in our new, net new bookings, when we book new logos, is the ARPU there is much higher. But again, it hasn't penetrated and gotten into the base to the degree where it makes a material difference yet.
Perfect.
We're seeing a very slow creep up from the $210 range to $215 range, as far as to $215 per agent, and it's slowly creeping up. So just the opposite of, you know, we're not seeing any price erosion, we're actually seeing more, more software that we can sell to each customer.
Perfect. And then the last minute, Barry, I'll ask you a question on margins. So you've been in investment mode ahead of some of the demand that you were seeing on the booking side. So I think some of the margin improvements, especially on the gross margin line, you've seen, you know, some deterioration, them are maybe flattish on the EBIT line. Can you talk about how we should think about that trajectory going forward when you start to see, you know, some of these revenue tailwinds materialize? Is that the catalyst for more margin improvement? Maybe you could talk about that.
Yeah, you put your finger right on it when you talked about revenue tailwind. So there's two blades to this. We are making massive investments across the world, in India, South Africa, and putting up instances in other parts of Europe, FedRAMP, you name it. And despite that, we've kept our margins there, and we're pretty proud of that. What would be transformative is when the revenue picks up, and that's why we've hesitated to give, because the uncertainty around the revenue, to give any indication of what we expected for Q4, unlike last year, where we had more visibility. 75% of our revenue comes from subscription. Now, that includes the commercial versus the two-thirds I talked about earlier.
Those margins go up when the revenue goes up. For the last 10 years, nine of those years, special case in one year, it's the strongest in the Q4 because that's when the subscription revenues are strongest. It's the leverage against the fixed and semi-fixed costs.
Perfect. Well, that's all we have time for. Dan, Barry, thank you so much for your time, and thanks to everyone listening in as well, too.
Thank you.