Get started. Well, my name is Clarke Jeffries. I am with the technology research team here at Piper Sandler. Very pleased to be joined by Jason Conley, EVP and CFO of Roper Technologies.
Thanks for having me.
Yeah, absolutely. Welcome to Nashville. Let us maybe just start off with some background on Roper Technologies for those who might be not familiar. What is the history of the company? What does the business look like today?
Sure. Thanks for having us. It is the first time I think we have been in Nashville, so it is great. Roper is a free cash flow compounder. If you think about Roper, we have compounded free cash flow in the mid-teens over the last 20 years. Our roots are really in, we were a diversified industrial company, and now today we are a diversified technology company. Really since 2008, we have been acquiring vertical market software businesses. We have sort of had this industrial veneer, but a lot of software underneath it. Then our GICS code changed in 2022 when we divested our industrial businesses. The through line for Roper, over the last 20 years, is we like being leaders in niche vertical markets. Small markets, typical TAM size for us, the largest TAM that we are in is maybe $4 billion. Right?
We're leaders, so we extend our leadership. When you have market leadership, you can still continue to grow, cross-sell, get new logos. It's a good place for us to be. You get a low range of outcomes as a result of that. We take our free cash flow, we reinvest that to the next great vertical software business. Today, we're about $9 billion of revenue, 40% EBITDA margins, 30% + free cash flow margins. We have 29 businesses, 21 of them are vertical market software. Eight of them are product tech, vertical tech businesses with a lot of recurring revenue. Think of us as mission critical across the industries that we play in. System-of-record application software or critical network software businesses are primarily what we own in terms of software.
Perfect. Maybe we could just briefly talk about some of those key segments or key end markets just to give a flavor of what those TAMs look like.
Sure. Just to set the context, we love end markets that are protected, but at the end of the day, we really are thinking about the types of businesses we want to own, the business models, the high gross retention, the high recurring revenue, the ability to cross-sell, all those things I've mentioned. Good working capital efficiency and low CapEx. Those are the characteristics from a fundamental standpoint. If you think about, as we've acquired over the years, we've just sort of naturally fallen into really good kind of protected markets. We're in healthcare, we are in govtech, we're in education, institutional tech, insurtech, and then a variety of sort of, I'd call them just broadly industrial tech type spaces. Yeah.
Given that, I think one of the questions I get from investors is just how I think a lot of investors are familiar with strategic acquirers of software assets. Where there is some iteration of the code, whether implicit or explicit around a strategic strategy. Maybe talk about differentiating how a purely private equity approach would be versus how you approach M&A of software assets. That would be a great place to start.
Sure. I would say just from the outside, if you talk to sponsors or you talk to bankers, they think of us more as a private equity versus a sponsor, or versus a strategic. That can be true, but I think the way we differentiate is we think about owning a business over decades, not over five years. So we can think about a business that in 10 years is going to be better than it was in five years and the like. Because we are just continually reinvesting in that business. We are adding new capabilities, new leadership. So, I think that is a huge differentiation. The other is, if we are going to buy a business, and this is part of a change in strategy to buy a little bit earlier in their lifecycle, we are doing bolt-ons to increase the organic growth of that platform.
I would say private equity does the same thing, but they also know that they can potentially buy down a multiple by doing bolt-ons. So we will do bolt-ons really to make the platform stronger. Again, just thinking over a very long arc of time. So I would say that is a key differentiator between us and private equity.
Yeah. I am just thinking of the metaphor of a house that has been a perpetual rental versus a home that is owned by the owner. Just what kind of investments do you make to make sure that that asset is at the highest quality possible? Or do you defer things? Do you never invest in it because it is always within the horizon, so.
Sure. Part of that is, we know that when we acquire a business, we know also the kit of tools that it takes to sort of get the business
Yeah
where we need it to be. Maybe it's not the first year or the second year, but the third. We're trying to move that forward. Obviously, we're always trying to get better, but they're always really good businesses in good markets that have primarily, they started in venture, had good product market fit, then got into private equity. They had that iteration of a strategy, which is kind of shorter term, and then it's like, well, now we have a strategy where we can build on that forever. That's sort of
Yeah
that's the life cycle, and we like that space.
Yeah. Well, we'll definitely get to, I'd say, the work that you do with ongoing assets in the portfolio. I'd love to talk about at least the onboarding funnel of new assets via M&A. It's been a little bit tighter of late in terms of M&A activity. Obviously, you mentioned the bolt-ons, going back to the prior year, there was about, I think a third of the M&A activity dedicated to bolt-ons. What's your expectation right now for the M&A environment? What are the things that might unlock it in the coming quarters? What is maybe the immediate strategy for when that might unlock?
Sure.
What would you go after first?
Well, I will just start. It has been a crazy couple of years in terms of capital markets and specifically in private equity. What you have seen is businesses that were acquired in the 2021, 2022 cohort with lower cost of capital. So there is sort of that dynamic of exiting at a potential lower multiple because just the cost of capital alone. Then obviously we have seen what has happened in the public markets, with SaaSpocalypse. That is sort of gummed up the system as well because the sponsors are trying to figure out, when can I actually exit at a more reasonable multiple? There has been a bit pressure though for several years now for DPI. So LPs are pressuring to get return on capital. So all those things kind of swizzle and private equity has to continue to return capital, and that is how they fundraise.
Coupled with, you also have impending private credit cliffs in 2028 and 2029, and they are going to have to work backwards a year, kind of N -1 on that so they are not faced with that cliff, because we think that cliff is going to be pretty substantial. It is going to be hard to put more debt on a business. They are probably going to have to equitize as well. So I say all these things to say all the ingredients are there to have a much more robust market. We think it is probably going to be somewhere closer to 2027 than the end of this year. But we are encouraged by that because all those factors mean reasonable prices. But it is just hard, right?
It's the kind of stages of coping with buying something for really high and then having all these external factors sort of force the hand to sell it at a more reasonable multiple. We're going to be a buyer of choice. We're having a lot of constructive conversations. I talked about all the macro. At the micro, having really good conversations with a variety of sponsors, proprietary looks, management meetings, those sorts of things prior to any process. We're hopeful that that will provide some unique opportunities.
Well, I think a very hot topic for how investors try to view these software markets is really what's the impact of AI? I guess what I think is the presumption is that if shipping software has become less of a scarcity in the market, the full stack of software becomes a lot more open for new competition.
From your vantage point, a company that's getting their hands dirty with companies with TAMs less than $4 billion, has anything changed on the recent climate with AI in terms of business formation or new competition? As an acquirer of those potential assets, you must have, I guess, some view of what's happening on the ground. Anything you'd say on how AI has changed the climate within those sub-$4 billion TAM companies?
Yeah, I would just say, we've been at this for two and a half years. We've really pivoted to, and our M&A team primarily work with private equity sponsors, but we increasingly had a lot of constructive conversations with venture. Part of it is to see what's in their portfolio, and more so just like what are they doing and where are they going after? I would say broadly, they're not going after vertical market software. They may go after some new part of labor that needs to be automated, and that part might not even be attached to the system of record. It could be like a front desk type software at a company. We're not seeing anything sort of-- there's a couple of things that have emerged in a couple of markets, but nothing that's wedged in.
We've actually had an example in one of our businesses where someone had a little wedged product, got about 10 of our customers on it. But the customers were frustrated because of the integration was sort of wonky and they had to do a lot of manual. So they came to our CentralReach business, came to CentralReach and said, "Can you create this wedged product?" And within eight weeks we did, and that competitor is now gone. Not that that wants to be a primary strategy, but the ability to innovate much faster with AI allows us to kind of, again, our customers trust us. They'd rather consolidate with us, so we do get a lot of feedback from them on things that they want or any sort of fringe startup that might be out there. Again, most startups are not going after where we play.
They're going after other parts of the labor, which we're going after a lot of that too, but more so the things that are connected to the current workflow or things that can sort of change the workflow that we're in and then how that attaches back to the system of record. So kind of rethinking how processes work. That's part of what we've been focused on.
Yeah. Last part on the funnel of onboarding assets. You have had a strategy that shifted to earlier lifecycle companies. Maybe you can give us some background on what prompted that change and even if you are in a position of acquiring earlier lifecycle companies, are you still able to do a medium-term target of cash flow outcomes for those businesses? Or how do you have to adjust the expectations
Yeah
for those kinds of assets?
So yeah, we pivoted the strategy a little bit in 2023, and we deployed about $10 billion of capital against that. Part of that is also doing more bolt-ons, as you mentioned. It used to be 10% of our capital. It has been more like a third. And again, that is all to increase the organic growth for the platform that we are attaching it to. I think for us, there are a couple of reasons why we would want to do it. Obviously, higher organic growth in general, which equals higher year five and further cash flow. We have to underwrite a little bit higher growth. But the businesses we are looking at are the same businesses we own in our portfolio. It is just think of it as like the TAM has not been digitized yet. There is still more adoption.
There is still a lot more white space in those markets. We understand the markets. It is just we are underwriting it to a little earlier part of their life cycle. This is not venture, right? This is first-term private equity, typically not third term, second or third is where we used to play. We like that. We like the ability to capture more growth for shareholders. Then also, as I mentioned, to be able to do bolt-ons. These businesses had not had a Frankenstein bolt-on sort of scenario that we can put our print on that business and have our own platform to grow from there.
Yeah. Well, let us shift gears to maybe the strategy around commercializing AI. I think it is something I have heard from a variety of companies that are acquisition-minded, as some of the functionalities for AI can maybe supercharge integration. It can be productized across the portfolio, and these sort of individual businesses may not. They can share some learnings. So, maybe we can talk about the big pillars on how to commercialize AI, some sort of the agentic products that have been released, and maybe any specific segments that have the fastest traction so far.
Sure. That is a lot. I will try to capture it. Well, look, this has been a fun year. We kind of really got into building agentic solutions on top of the software towards the end of last year, early part of this year. But I think we have learned that the adoption can vary widely depending on what is happening in the end market, what is happening from a top-down perspective at those customers in terms of pushing change management. We are adopting as we go along. But I would say our CentralReach business was sort of first out of the gates in terms of AI adoption. If you think about it, CentralReach plays in the autism space, so you have got a huge disparity between demand and supply for autism therapy.
We were able to provide solutions right in the workflow that enabled scheduling to be faster, claims to go smoother, and be first pass through. We were right in the workflow, and it had high ROI for the autism center, immediate ROI. It made the therapist's existence better because they didn't have to spend as much time with the administration, more time with learners. In that case, what they did too is they did a freemium model for, call it six months. They could track the telemetry, they could see that there was adoption, and then said, "Okay, now we're going to start charging, and we're going to start charging above and beyond the core system of record." You had quick uptake on that.
It was good because it was seamless, and the therapist wasn't, by the way, worried about losing their job because there was just so much demand in the industry. You go all the way to another, sort of think about in our Freight Match business, our DAT business. We're automating workflow for the spot freight market, so a broker doesn't have to make 10 phone calls now to match a load. We have the technology right there. Well, it turns out that's a different app because a broker thinks, "Wait, that's my secret sauce. I can negotiate. I can get on the phone. I can do this." Which is true. For maybe 90% of what they do, they're still going to be able to do that.
But for 10%, where there's high-density loads, and it's pretty obvious what the rate's going to be because we've got all the data there, that could be automated. But that's change management, right? There's all of that sort of goes into it. Then you can paint a brush, and there's a continuum of that sort of going on throughout our businesses. But I would just say, the good thing is we've got the learnings from CentralReach in terms of the freemium. We applied that to Vertafore. It's really great to be able to kind of learn from each other, successes and failures, as we go forward.
Yeah. I think something that people may not appreciate with some end markets is just that the labor growth has been stagnant for a very long time.
Right.
And this idea of impairing the labor or replacing the labor, we needed that 10 years ago
Yeah
for a lot of these end markets. So, maybe just on the strategy of how to achieve that, of commercializing. You introduced an AI accelerator team, a part of those shared learnings, increased the velocity of tools. Maybe we could talk about how you hire that talent, the size of the team, maybe some intentions for growing that team, or targets for the coming quarters.
Sure. Just to be clear, the AI accelerator team is mostly about helping our businesses develop agents, right? For revenue, not for productivity. Having said that, they're definitely helping them with AI SDLC, right? How do you agentically code? So there's a benefit to that. So we hired a couple of leaders from an enterprise software company at the end of last year. And we're now up to, I think, between 25 and 30 folks. So think of them as just a team that can come in, so they're at, call it like four or five businesses right now. They can come in. The businesses provide the product roadmap, the agentic product roadmap, and then they help teach and then build at least the initial agents so the customers or the companies can then take that to market. So, we've had great success.
Vertafore was our first business that they went to, and we've got, I think, somewhere in the realm of 10 agents now that are out in market that are at least developed. I think four or five are out in market now. So that's something that Vertafore on their own, it probably would've taken them a lot longer to do that. And so now it's this whole thing of teaching and building the teams from within the businesses, to take the ball and run with it. But I think it's been a great experiment because if you think about Roper, and we honor the autonomy and the decentralized model that we have. But this is such a compelling opportunity for all our businesses that there's just been a pull for this. I am just delighted that we have decided to make this decision, and it is going well so far.
Yeah. In terms of this being targeted towards commercial products for the end businesses, I think one of the things that stood out to me was something around your Strata business, and they had a product release, and they said that customers should not have to recreate the financial context they have already established in Strata. I think as you have done this AI accelerator team, how often were these businesses that were under $4 billion of TAM, they were just unable to execute on obvious adjacencies?
Well, first of all, it is a different way to do software development, right? This is not linear coding. Like I said, it is a 10x AI SDLC, first of all, on the coding, and then to actually build the agents, do the evals, understand how the product works. It is a different muscle. We are teaching them. Strata is a great example, though. We have been able to elevate some folks in the organization that have said, "I want to be a part of that." The Strata, it is called the FDI, Financial Decision Intelligence. It is a great product. I can talk about it for a long time. We have a long queue. We actually have a lot of folks that are testing it out right now, a lot of large health systems, and we expect bookings this year on that.
It is essentially, if you think about health systems, so Strata, think about it as FP&A on steroids for health systems, and we are able to marry up the financial, operational, and clinical information. We have always been able to do that. But now with AI, we have got the data structure that we have built over many decades. So you think about how to split it between department centers or operating rooms. There are many different cuts, and we have that structure. There is the semantic layer in terms of how you actually calculate margins. I know that sounds trivial, but in a hospital, it is super complicated. Then the ontology, the ontology of how a patient moves through the system. So the combination of all that is super powerful.
Now with AI, it's moving from, let's just say an analyst before would have to go and pull some things together and get the report they needed for their boss. It's like now in the morning, they can get up and they have actual insights because they can prompt the Strata FDI immediately. We're getting great feedback on it. We're really excited about it.
Before I transition, any questions from the room that'd like to be asked? Okay.
Just a lot of great AI assistance for you. Anything that's being disrupted by AI?
Anything that we're seeing that's getting
Yeah.
You mean like our business is getting dis-
[inaudible]
I'm sorry?
[inaudible]
I mentioned before the AI startups, and so far it has been pretty quiet. We have had one business where it is within our Deltek business. Call it is about 10% of their revenue. What they do is they pull public RFP data off and then help the businesses prepare proposals for bidding on work. That business, because it is public data, obviously that is more commoditized. We have always had a lot of small companies on the fringes, and they will go after smaller government contractors. Our response to that has been, we also at Deltek own Costpoint, which is the ERP for project-based businesses. Now a customer can marry all of their historical Costpoint. What did I win on this project? What did I do? And incorporate that into this bidding solution. That is different. I think that will help sort of bolster that business.
That is the biggest where we are actually physically seeing competitors come in. I will say in the insurtech space, there are more startups going after the work that the labor pool is there. We are too. We have a Velocity AI platform that is going to be the orchestration layer for a lot of the tasks. Because if you think about an insurance brokerage, it is super manual processes. I used to manage insurance for Roper. It is like battle of the forms. Carriers have all their different ways of wanting to see things, different portals. We think that is a huge TAM that we can go after for Vertafore, but there are others that are trying to do the same. That is probably the two that I would call out.
Now that you have been observing, obviously, the impact of AI, now that you have been observing it for maybe a year or two years, what are the characteristics of the businesses that are most resilient and then perform with AI? Are there some commonalities you see in those businesses?
The question was, what are the biggest most
[inaudible]
Yeah, it is. For us, there's many layers to that. I think if you're critical to the workflow, if there's a high cost, if there's a risk if things go wrong, right? So when things go wrong, there's a regulatory overlay that's probably embedded in that. There's high configuration density at the customer level. So that's a huge part of how easy is it to rip out, how critical is it to interface with all the other systems, and then also just a network effect. If you think about our segments, application software is the first three of those. In our network business, our DAT business, it's all about that interconnection between participants, and that's very hard to break unless you have critical mass.
The ideal business has all of those together.
It can. Sometimes you can have three or four of them and not. Let's say in application software, you don't really need the network as much, but you do need to be super entrenched. The other part of this, and this is what's exciting about AI, is the word customization used to be a four-letter word, but now you can customize the AI on top of however their workflow is configured. So you have high entropy because that customer is different than another one, and we can actually scale now with AI across the customer base. So that's exciting.
Well, Jason, unfortunately, we're out of time.
Yeah.
But I really appreciate you making the trip to Nashville.
Yeah. Thanks for having us.
Yeah, absolutely.
All right.
Take care. Thanks.