Roper Technologies, Inc. (ROP) Earnings Call Transcript & Summary

Awesome. So let's get started. Thank you, everyone, for joining us this morning. Kicking off day 3 of the Morgan Stanley TMT Conference for me. Very pleased to be hosting from Roper Technologies, Neil Hunn, President and CEO; and Jason Conley, EVP and CFO. So gentlemen, thank you so much for joining.

So maybe just as sort of to open up the conversation for maybe some people in the room maybe not as familiar with Roper Technologies, a little bit of a kind of one-on-one. This is a little bit of a different story from a lot of the [indiscernible]. What is the Roper story all about?

Yes. Thanks for having us. It's always great to be here, Keith. So yes, those that are newer to our story, Roper is this year, roughly $8.5 billion business, 40% EBITDA margins, 30-plus percent free cash flow margins. And we're a durable, steady, vertical market software compounder. I think that's what's unique about sort of our model versus others. So really a dual thread offense sort of steady organic cash flow generation and then we take all the cash flow generation and deploy it offensively. Most of the time, historically through M&A, more recently through buyback. When you put it together, sort of we think about how durably and consistently can we compound our free cash flow per share. That's been in the mid-teens area, and we have the strategies and operational rigor in place to try to push that to be in the high teens area. In terms of the portfolio construct itself, it's vertical markets, right? So it's -- all of our businesses are leaders in small markets. We're super vertical-oriented, application-specific, high retention, high gross margins, high cash flow dynamics attached. And we'll get into all that, but that's a flyby of Roper, durable compounding vertical market business.

Got it. And you guys often talk about edge markets as kind of where you're looking for those either market leaders or emerging market leaders. How do you define [indiscernible] market? And within that, what are the key kind of KPIs? What are the key attributes you're looking for in terms of an emerging leader or a leader in that market?

Sure. So just to double-click on the high level. So we love owning businesses that are leaders in small markets. So small market, we mean small, like the largest TAM that we have in the portfolio of 29 businesses is about $4 billion, plus or minus. Like a lot of our TAMs are $1 billion plus or minus. So that is the first thing as it's a very small TAM. So these are TAM like software for early childhood education centers, software for autism therapy clinics, software for large law firms, software for tax accounting for publicly owned utilities. I mean these are niches within niches and very, very sort of specialized. And we like these small markets for 2 reasons. One is they are protective. So very rarely do we have a new entrant because the size of the prize is quite small, and they're well served by us. But also, I think more importantly, it means we compete on this notion of customer intimacy. Like we are just ingrained in the operations of what our customer does across the totality of the portfolio, our customers can't do what they do without us, right? So we are sort of the backbone of their business, system of record or whatever network that they operate within. So that's what we mean niche and sort of edge market. That's what it is. small markets where we have a clear leadership position.

Got it. I'm going to skip around the question list because I think that brings up a really interesting point. And let's just get straight into what has been sort of the major debate point of the TMT confidence really within the software industry for the last 2 years is the -- how insulated certain markets are or how well positioned certain markets are for AI. So when you're talking about these smaller markets feels more protected, do you guys see that as a more insulated kind of market from an emerging competition from a large language model lab? Yes, that's basically the question.

So our empirical evidence to date is absolutely, right? So again, we have 29 business in the portfolio, 21 are vertical market software businesses. The balance are sort of application-specific product businesses. So relative to the 21 -- it is our -- we're not aware of a model company or an AI native start-up that's attacking the core of what we do. I think there's logic behind that, by the way, which is if you're an investor and you're starting company, do you want to really sort of attack, if you will, a system of record and a $1 billion TAM that's well served, your rate of growth is going to be sort of rate limited by dislodging of a strong incumbent that has all the incumbent advantages. And so we haven't seen any of that across the portfolio in any capacity. So it proves to be quite insular in that regard. I think more importantly, though, I'm sure in this reading your recap notes for the past couple of days, I don't think what we're about to say is like unique from what other companies are seeing, incumbency has like real tangible advantages as we're trying to develop AI tools at whatever term you want to use, the agentic layer, the task augmentation layer, right? So at the highest level, I've said this for a couple of years, we've said this for a couple of years, it's a combination of 2 things, and everybody focuses on the first one, which is like do you have the data, the knowledge graph, the context to sort of take a sort of a predictive model, make it deterministic because that's what has to happen when you're augmenting and replacing work. But I think what also is often forgotten about is you have to know where to point that capability to and what problem to solve. And these are like very esoteric "boring tasks" that are repetitive, but it's truly like magic to our customers. Like how do you automate account reconciliations that take 17 minutes into a minute? Or how do you auto verify order entry for pharmaceuticals, which is done by a pharmacist today, a high-dollar pharmacist? Or how do you take 10 phone calls out of matching a freight load in the spot market and do it with no calls? And so these are like it's truly magic, but you have the capability, the knowledge, the data, the context, but the problem to point it to. And in these little niches that we sort of tend to own and that we're the #1 player again in each of them, it's -- that's the magic of what's happening, right? So lots of opportunity.

And back to deterministic, I mean, there's a regulatory compliance overlay in almost all of our vertical market businesses, and that's very hard to replicate. That's your decades and decades of getting compliance. So our customers don't want to sort of violate any sort of regulatory hurdles, and it's harder to get in. It just increases our moat.

Yes. And that was exactly what kind of triggered me to jump down into AI. When you're talking about intimately knowing your customers, software is all about solving your customers' problems and solving the sort of the business problems and making those businesses more efficiently. And you're right, everybody just focuses on the data. And there's a little bit of an inconsistency in that argument because ultimately, it's the customer's data, right? It's not -- you're managing it for them, you have visibility into it. But what you have is that domain expertise, you understand these domains. And from my perspective, generative AI, the real opportunity is not to displace existing transactional systems is to automate workflows around those systems, and you need to know the problems to be automated. You need to know the business process to be automated. So the question in all of that is, I like it that you think about the world in the same way I do. But how do you push that down into this distributed portfolio of companies?

Yes. So very important context setting. So let me just take a minute to sort of talk about our org structure and why we operate the way we do. And then in this regard, it's -- I'll say it's better to be sort of lucky than good in the context of what we have to do with AI in the org structure. So because of the portfolio we have, like I mentioned, we're the largest player across all 29 of our markets, which means this is over a long arc of time. This has been since I've been at the company 15 years, Jason, 20. We've been in the modern day Roper for going on 25 years. And so all of our competitors, by definition, are smaller, which means probably they're generally a little more nimble. So our entire organizational structure from the dawn of the modern day Roper been super decentralized, right? To optimize on 2 things: speed coefficient and proximity to customer, right? So 20,000 people in the company, 130 in the corporate center. So every other resource in the organization is closest to the customer, organized around speed. Like that's been the design principle for 25 years. So that's why I say it's better to be lucky than good. If we happen to have been organized for like cost of development or more centralized resources and not on customer centricity and speed based on our heritage, I think we have a real problem in the AI era, but we're not. And so we're -- as soon as the sort of light bulb goes on inside of our organization -- organizations, then the speed coefficient sort of builds on itself. And so going on now 2 years ago, we said to our organization, all 29 of our leaders, like this AI thing is not a trendy paradigm. It's the next tech wave. We have to lean into it. And we've had each business now 1.5 years ago reunderwrite their entire business model, be AI native to get the thinking going. We've had them sort of change the way they do development. So smaller teams closer to the customer, collapsing product and development, sort of minimizing design, making that part of product, all the things that you do across all 21 of the software businesses. And so we really sort of have this increasing sort of flywheel speed happening at the point of impact. Now what we also identified is this is in the last maybe year, and we just started hiring the team about 6 months ago, where we were hiring sort of an AI accelerator tech team at the center. We'll be small to start, 20 or so people. We're fortunate enough to have 2 exceptional leaders come from another large application software business to sort of lead the team for us. And what we found is as our businesses were experimenting and learning how to develop an AI, it's a very different development pattern pathway than traditional development, where you can have a -- traditionally, you can have a sort of an architecture and sort of deterministically work your way through. There's a little bit of art in AI development. And so we very much wanted to sort of have a team that has sort of grew up in machine learning, trained in machine learning and trained in AI, had done this at scale in the application layer. And so their mandate in the organization is to accelerate the technical teachings to speed things along even further, but also have a build team like they're going to be a small sort of, if you will, "10x" build team that partners with our companies, and then finally, there's going to be a shared Roper repo. When you have common functions like whether it is document retrieval or document sort of parsing, that's a common AI sort of run time routine. There will be a central repo of that, so we don't have to multiply that times 21 -- times across the organization. So sort of this sort of perfect storm of our legacy close to customer, have speed built into the organization and accelerating it with the center team.

It almost reminds me a lot of companies that we're talking to have really brought forward the concept of forward deployed engineers, right? You need people who are really skilled in the AI to be able to help the businesses, right, and usually in the customer concept to understand how to utilize these technologies that really don't work very much in the same way the traditional technologies work. You guys do within your portfolio companies. You guys have talked about approximately 25 Gen AI initiatives underway within the broader organization. Any like early wins, any initiatives that are seeing significant or early traction, and on the other side of the equation, where do you guys see the white space? Like what looks most interesting in terms of new areas to go out and automate?

Yes, sure. So when I talked about that and we talked about this 25, that was probably 6 months ago. It was the culmination of the early product development work we've done in the organization. So this is product-oriented, playing offense sort of monetizing in the tech stack. That number is basically not discernible now. It explodes. So we're not going to talk to the 25, then 40 to 50 to 100, but it's much more than that now. Going back to a couple of years ago, we made the early decision that between using AI to drive play offense and put in the product stack and grow and expand versus use AI for productivity, we're tilting demonstrably more towards playing offense. And so everything that we've talked about today has been about that. So that's where the principal amount of our focus is. Obviously, we're working like every company on coding sort of efficiency and customer service and sales. But that's -- we make -- we have 40% margins. Like we're not about -- we will improve margins over time, take that productivity, put it into the product road map, but the principal of our focus is playing offense. So a couple of examples where we've seen real success there. I'll give you maybe 2 or 3, if it's okay, to make it sort of tangible. I'd start with the best story, which is our autism therapy business is called CentralReach. This is the backbone that autism therapists sort of run their practice on. To set the context here, there's something like 800 million therapy hours demanded a year in the U.S. and 300 million therapy hours supplied. So a huge supply-demand imbalance, waitlist sort of people -- these families waiting for access to care. And also inside of the care, the people that deliver the care, they turn over 85% a year. So it's a very difficult problem. And so the tools, the AI tools that Central Reach has developed and deployed into the market is a scheduling tool. It's a very complicated schedule. When you're -- you have a demand and your people are turning over all the time, you can't have a blank and schedule. It's -- and by the way, these learners are in between 2 and 8 hours a day for 6 months. So if you have one person leaving a huge block in the schedule, it's a problem. Also all the note taking and sort of medical records and then all the revenue cycle back in the billing and collections. The punchline of this is the business has increased its growth rate from the low 20s to the high 20s. Its win rate in the second half of last year from the primary competitor went from 75% to 100%. And fortunately, for our customers, the attrition rate for the therapist has gone from 85% down to 40% because it takes a lot of the bad part of the job out. So it's like this win-win-win sort of in terms of customer, the market, the learners, the families and sort of our business. Another example I would share is earlier days, but we're very excited about it. We have a business that is -- it is the technology that -- and the network that organizes the spot freight market in the U.S. So if you're a broker or a trucker, you subscribe to our network to find out where your load is. And today, traditionally, it takes about 10 phone calls for a load to be brokered. Are you available? Can you pick it up at 3? Will you do it at this rate? Oh, I need you there at 4. Can you deliver it by Tuesday at 3? They take those phone calls. And on average, that's about $100 or $200 of labor to broker a load at the broker level. We now have in place the technology to do that without a human in the loop. We charge a fraction of that cost and of the labor cost, and that's happening in the real world. It's a magic button that sort of takes that inefficient sort of human connectivity out of loop to sort of accelerate the brokerage, the brokering of loads, which we think, by the way, it's great for a broker, it's great for a trucker, but actually makes the whole spot market a more fluid market. So we actually think there could be more volume that actually comes into the spot market. The spot market won't be the market of last resort. It could be the market of sort of middle resort because it's a more fluid, transactionally sort of dynamic market because -- now this, by the way, is very expensive technology. It's a lot of machine learning. It has to be absolutely deterministic. You have to have the volume of rate data because part of this is how do you price a lane. We have happened to have in our network because of the scale of it, like 7x relative market share to our largest competitor. We know what the rates are per lane, per day by truck type, and if you don't have that, you can't do the match, right? So it's this combination of technology and data. I can give you other examples, but those are 2 that we're quite excited about.

I think what's exciting is the companies that are farthest along in AI are growing the fastest in our portfolio. So just to give you a sense of like we have this incumbency right to win. And so it gets our other businesses really excited when they see that and it fostered a lot of learning across the portfolio.

Got it. Have you guys had to adjust the pricing model? There's another big debate in software is kind of what happens with pricing models. But have you guys come up with new pricing models, more outcome-based, more consumptive-based to align to the newer technologies?

I mean we're certainly doing a lot of voice of the customer work on this right now in some of our businesses and just looking at where the puck is going and how we should be thinking, especially in the businesses where we have the highest right to sort of win that game and increase the productivity of our customers where you could see some seat degradation over time. And so I think what we're seeing is that there's just pointing it out another direction, like maybe it's a number of customers, maybe it's a number of whatever the driver of that business, maybe a number of policies, that sort of thing. And our voice of the customer work so far has said that as long as we are capturing value for them, we get our fair share of that value. Our position on this, by the way, is to get the customers really excited about the product and really using the product before we sort of capture our share of value, again, just to continue to protect our moat and then play the long game in terms of AI monetization.

The only thing I'd add on top of that, certainly on the system of work agentic layer, our clients are telling us very clearly like, yes, we'll pay you for the value, but we don't -- we're not that particularly excited about an open-ended pay as you go. So it's probably going to be some sort of subscription with overage or something. So they have some certainty of what they're going to spend.

Tiers.

Some tiers, exactly.

I think it's a part of the equation a lot of investors have over-rotated on the idea that everything needs to move to a consumption model because when you talk to customers, customers don't love...

They don't want to...

Because they look at it as a variable -- a variable cost model, they'd much rather have certainty in terms [indiscernible] and enables them to get leverage on the technology.

And our current cost as a percent of their revenue on average for enterprise software businesses is about 50 to 100 basis points. We're not -- we're not a huge part of their cost part today. So we have latitude if we're demonstrating value to capture that value over in a fixed way if it turns out that way.

And just to double-click or just to round out the answer on the subscription plus. We are fully confident we'll be able to capture like the power users that are just driving token consumption. We're very comfortable with the margin profile of the subscription and then sort of like an old school cell phone plan. You get this many units of work. And if you go over, then you're going to pay, but there's some predictability for the customer.

Got it. Got it. I want to switch gears and talk about sort of recent results and how the company has been operating. We've been having this conversation for a couple of years running now. And I know the center of the Roper story is consistency in results, consistency in sort of the yields that you guys are able to bring to the marketplace. So let's talk about 2025. Revenue growth came in at 12%, $7.9 billion in revenues. Adjusted EBITDA grew 11% to $3.14 billion. Can you talk us through what worked well in 2025? And maybe what are some of the areas for improvement into 2026?

Sure. Yes, I can cover that. So I think as we obviously had some great business building in the year, but our growth wasn't really where we thought it was going to be at the beginning of the year. And there's a few reasons behind that. One is Deltek. It's our biggest business. It's exposed to government contractors. And that, as you know, when Doge hit and then there was nobody to deal with at the agencies with our customers not being able to talk to anybody at the agencies. And then the shutdown, that business was severely impacted in 2025. So you look at a business that has been a steady mid-single-digit plus grower and it was in the low single digits in 2025. So a big part of our business and a big segment. And then -- so if you look at our Application Software segment, it actually organic revenue improved 70 basis points year-over-year if you sort of exclude that. Of course, you got to own all of Roper, but just to give you some context of the impact of that. Procare was a little off to a lower start. We'll talk about that a little bit later. I think Neil will touch on that. That's a business we acquired a couple of years ago, but we feel confident in the ability to inflect the growth rate there. And then Neptune, which is in our TEP segment, our products segment, second half just impacted by tariffs and just skyrocketing input costs. And so we had to sort of process that with our customers in the second half of last year. And so again, we own it. We know we're going to do better going forward, but just a lot of things happened last year. As we look forward to 2026, we're sort of -- we're guiding organic growth 5% to 6%, taking a view that at least coming out of the gate, things aren't going to get demonstrably better. But we will have, like in application software, we've got a mid-single-digit plus. We're going to have a better comp just on Neptune alone -- or not on Neptune, on Deltek alone. So we're not expecting a lot of improvement, just a little bit better comp there. And importantly, our CentralReach business becomes organic in the second half. So that will add 80 basis points or so to that segment. Network Software will be mid-single-digit plus. We feel very good about that. That's just our Foundry business, which was dealing with writers and actors strikes to media and entertainment. That will get better this year. And then we'll have our Subsplash business will also become organic in the fourth quarter. And then our TEP segment, that's sort of a little bit more wider range of outcomes as we sit today. We have guided low to mid-single digits with the first half being a little bit lower. We have Neptune down for the year, but we'll see how that plays out. Neptune, their water meter business, and we're waiting to see when the COVID super cycle sort of normalizes out. So that's sort of the one watch out that could provide some upside. But as we sit today, we're sort of -- that's the approach we're taking.

Got it. That the same conservative. Double-clicking on the Deltek side of the equation. A lot of the executives that I've talked to about sort of the federal disruptions in 2025 see the opposite side of the coin of that. On the other side of the disruptions of DOGE is also the idea that we got to run the government more efficiently and software is going to be a big part of the solution. Do you see an upside scenario in Deltek of that you guys become a bigger part of the solution in terms of how to garner efficiencies in the government?