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

All right. I think we're ready to kick it off. Stage 2 of the Communacopia Technology Conference. My name is Joe Ritchie. I cover the multi-industry sector at Goldman also Cohead, our U.S. Industrials and Materials Research. Really happy today to have Roper with us today. We have Neil Hunn, President and CEO; as well as Jason Conley, CFO. Thank you guys for joining us today.

Thanks for having, great to be here.

So Neil, at the Communacopia Technology Conference. So at this point, you would think that folks be pretty familiar with the Roper story, but for those that aren't, maybe just provide a little overview on your value creation framework, and we'll take it from there. .

Yes. Just real briefly, Think of us as a durable, steady low range of outcomes, cash flow per share compounder. That's the mindset of the organization. In terms of the profile of the assets in the business, about $8.5 billion of revenue, 40% EBITDA margins low 30% free cash flow margins. Perhaps interesting for investors is the portfolio of businesses are hard for investors to own individually because they're small. They're tiny businesses that are leaders in small markets. So we're the leader in a small TAM. So we have the scale advantage locally, but we compete on intimacy with customers. We have strong product market fit, strong gross retention. And collectively, the cash generative capacity of enterprise is quite compelling. Now stand on each of the business with each business would not be particularly uniquely productive in how to deploy the cash flow they generate to their own individual purpose. So therefore, we take all the cash flow up to the center, the free cash flow, excess free cash flow and deploy it to what is we think the next best opportunity, whether it's a new platform, a tuck-in acquisition, a buyback at the enterprise. When you put that together, the cash flow compounding per share capacity in the mid-teens. And we go to work every day trying to find a way to get that to the high teens. So the rate of double goes from every 4 years -- excuse me, every 5 years to every 4 years. So low range of outcome, durable cash flow basher compounder, one of the largest, if not the largest vertical market software businesses that's publicly traded. That's a collection of 29 businesses in earn the hood. I'll stop there. We can take where you want to go to.

Yes, super helpful. Before we get into the whole AI discussion, which I'm sure will dominate a lot of this discussion. We've talked historically about being very focused niche markets, vertical software, why is the disruption risk low when you think about the horizontal players that are out there potentially coming into your markets? Like how do you think about like the defensibility of your -- of the businesses that you're in? .

Yes. I think there are several defensive aspects to it. I mean the first, and this is proving to be the case even in the AI era is the size of the prize is quite small. When I see our TAMs are small, our largest singular TAM for one of our -- our biggest TAM is about $4 billion a year. This is pre-AI. Most of our TAMs are $1 billion or less. So the size of the prize is just quite small. It's already very, very well vended, particularly by us and a very small number of competitors we have in each one of our verticals. So that is the case. I think 18 of our 21 software businesses or work in a regulated end market. Oftentimes, the regulatory regime is why you need very specialized software in a vertical, whether it's education or insurance or government contracting or whatever it may be health care for sure. and that specialized end market, there's something unique about the way that end market operates that isn't suited particularly willful horizontal software. Every one of our businesses were created sort of in the nooks and crannies of the horizontal. So from the day that they were born, the businesses were born, they had to create differential value from the generic application of horizontal software. And those are some of the reasons why it's historically durable. Going forward, on an AI front, if you think about in our opinion, like what are the attributes for an AI winter, Well, first, you have to be able to take this probabilistic outcome to deterministic that's largely done with the context and data knowledge graphs that we have across all of our businesses. Then you have to apply that in a workflow with essentially no latency and you have to have the customer trust to be able to do the first one of those. So it's the aggregation of those that is, I believe, an note advantage in AI, which I'm sure we'll get into as we get further into the questions.

Yes. So let's start. So since last year, you have given a ton of examples on earnings calls, in particular, about how AI adoption is accelerating across the portfolio. Maybe just level set us on what parts of the portfolio are already monetizing AI and like bookings or retention ratios and how you're monitoring it?

Just to step back in a little bit of context. So when the AI gun went off a couple of years ago, our decision at the enterprise level as we want to take all of this goodness of these tools and drive into the product stack to drive customer value to drive growth and play offense. And so from the very beginning, that's been the mantra in the mindset, not how can we do what we do substantially more cost effectively and take it to margin. So we're playing sort of a growth game knowing that over time, we'll get the productivity and maybe there's some margin benefit, but we want to take all the productivity gains and go play offense. . So what that manifested is all of last year and this year, it's been how do we develop these get SKUs that are doing these task replacement sort of things across a large swath of our portfolio, not domiciled in like 1 or 2 of our businesses, but essentially all 21 of our software businesses. That's been largely successful in the first wave, and now we're in the commercialization sort of learning phase of that broadly across the portfolio. Now when you have a portfolio of our size, you definitely have companies that are that lead the pack and ones that lag. So your question, like are the early adopters and where you've had real success in autism therapy, legal, increasing success in some of our other health care IT franchises, early success in insurance, early success in pharmacy automation and just to name a few, but it's -- I call it early. Like we've got to scale this. It's got to get scaled across dozens of a dozen-plus of our software businesses, not just the product releases, but the commercialization, the retention, but signs are encouraging early, but it's too early to call like the trajectory of that.

So maybe you have a disparate group of 21 software companies, right? What is it about the ones where you've had early success versus maybe more kind of like you think the adoption is going to be a little bit more gradual? What are -- what are some of the reasons why these companies, like whether it's central reach or Aderant is having a little bit more success than others.

I'll start and Jason can add some color. And so I'll pat like 2 distinct points on this map and then I think the world will -- most of our adoption will live in the middle. So in autism therapy, there is a huge demand for the service in the United States. There's something like 800 million therapy hours that are demanded every year. There's 300 million therapy hours supplied. There is a line outside the door of every autism therapy clinic for families that need these services. . And so our 3 categories of agentic SKUs basically allow the therapist to have more hands-on physical therapy time with the learners, so instead of 15% administrative time, it goes down like 5%. So they see more learners, the line gets shorter, more families see the care, the practice makes more revenue. And so it's a win-win-win. It's just a win all the way around. So that's been a very fast adoption, faster adoption pace for us because there's -- it has been. On the other side of the coin, same conceptual setup and freight matching in the spot market. We own the business that organized as a spot freight market in North America. It's called DAT. Roughly 750,000 unique loads are advertised on our network every day between brokers and carriers and the capability that we built, bought and built now can automate entirely the matching process of a freight broker to the carrier. Today, it costs between $100 and $200 of manual labor to do that. There's about 10 phone calls that happen each one of those matches. Are you available on Thursday. Can you do it for $3,000? Can you drop it off on Friday by 3, this happens back and forth, I need your insurance. we collapsed all of that to an automated charge $40 and $50 to do it. That at rate of adoption has been slower because the person pressing the Magic buttons like that's my job, right? So there's a human adoption that sort of has to sort of go through. Now the unit economics are so compelling. It will happen, but the rate of adoption is going to be a little bit slower. I think the reality for most of our Gentex tasks are going to live in between those 2 because in the case of DAT, it's not entirely placing a freight brokers job, but a big swath of it, it is. Most of what we're doing, whether it's pharmacy automation or insurance automation, we're replacing and augmenting individual task -- and so I think the propensity to sort of -- yes, I don't like doing this reconciliation task. Of course, I don't like to have confirmed every script every day for 4 or 5 hours of my day, if software can do that. I'll let it do it. So then the people can go to higher value-added test. So I'd expect the rate of adoption for the bulk of what we're doing to be -- between the 2, not super fast, not super slow.

The only thing I'd add is we are in small markets. And so what we've seen at least some cloud adoption as an example, I'll pick on legal. We thought it was going to be the slowest to adopt cloud. And then when COVID hit, you had this tipping point. And then when law firms knew that others were implementing it. You had -- it's sort of an acceleration. I do see parallels here where you're going to come back a year from now at these user conferences you're going to have reference customers. And they're going to show the benefits of it from a productivity standpoint and say, I got to -- so there's a lot of sort of conversations that happen within the industries. And so I think we just need to see enough of those to get that tipping point.

Super, super helpful example. So last year at this conference, I asked you about disruption risk. You mentioned foundry as a potential portfolio company that could be disrupted maybe sometime down the road. And then this year, foundry return to growth you launch more AI tools. So has anything changed about the long-term risk with foundry? And then as you think about the broader portfolio, any further thoughts on whether there are other parts of the portfolio that could be disrupted?

On foundry specific, the context here this survey has at foundries less than 1% of our revenue. So if we're wrong in this statement, it's not a hugely impactful to Roper. What foundry does, by the way, is it has a very clear market leadership position in post production, high dollar per minute Netflix streaming, theatrical releases that takes a live action and the computer-generated graphics and composites that into a single frame. So I think the Game of Thrones dragon and the humans putting it into a single screen. It all happens in post production called composting and we have a very, very strong market position there. If you believe that all high-dollar content is going to be completely generative, then there's not a post-production step to do. So that's like the existential threat for that business. The current course of speed is that is not what's going to happen. AI is being used to generate more CG and more CG means more compositing and what foundry is doing is making the composite step even more automated and so in the intermediate term, multiyear term, it feels like there's wind at the back of foundry, but we can all draw individual conclusion on where the future 10, 15, 20 years of high end, like $1 million plus per minute sort of releases will be in terms of content creation. So that's sort of the frame the existential risk. For the balance of the portfolio, we feel very comfortable. Like if you think that's ultimately with foundries like what is the -- does the end customer exist, yes or no. This compositor exists. For the other 20 software businesses, we feel very convicted that the end customer exists. There's going to be an autism therapists. People are going good at church. Insurance is going to continue to be sold for complex sort of business insurance, like the customer exists and it's in our job to automate as much of that customer's activity to sort of service that end market as possible. So we feel very comfortable with that opportunity. Its very in favor of growth versus contraction.

Very helpful. Think about -- maybe talk a little bit about the pricing models that exist, right? So I'm sure they're disparate across the different portfolio companies. So.

Yes. No. And it is disparate. Some of that is inherent with whatever the business model was before. So we have a couple of businesses like parts of DAT where it's automated already. It's based on per transaction or a soft writers based on per script. So the genic charge meter is going to be similar to that, and the customers are used to that consumption. But that's more of an exception. I think what it's starting to be apparent, and I think everyone's heard this is that as a CFO, especially, I want to have certainty on my budget. So something like a credit system is what we think is going to be probably the preponderance of what's going to be out there. It's something that Vertafore is put in place in the market. So essentially, if you're a customer, you can draw down credits based on the task that whatever agent you want to choose, there's different meters based on different tasks. And so that's good because it's a user loses a type policy. We get recurring revenue out of that, more certainty to the customer. And so that's where we see things headed.

Okay. Great. And then we talked a little bit about some of these efficiency gains. You talked about DAT and maybe some of the slower adoption. One of the things that we've talked about historically has been that AI could reduce some seat and some of the workflows. Have you seen any kind of seat compression today across your portfolio or no?

We've not. Gross retention is -- continues to be a very strong mid-90s at the enterprise level, the number we sort of talked about earnings calls. We obviously monitor that. We also inventory like the pricing model this goes back 1.5 years, 2 years ago. And there's only 3 of our businesses that have a pricing leader that we might potentially have to adjust the meter over the media mark. Nothing we have to do like -- we change it like in a nanosecond. And the reason there's only 3 is think like health care IT, to the extent we have seats is like the total number of health care workers. So we're not -- we don't need to change that. In insurance, is the one that we have a very clear line of sight, property and casualty insurance on how we need to modify the pricing meter. We have the plans in place. We have the telemetries being built. The first wave that will go out in the first part of next year, it's completely aligned with the customer. We're not trying to price grab or change total price. We're just changing the underlying meter from a seat to something that's more indicative of the growth and the unit economics of the insurance agency itself. So that's the one that is sort of front and center. The next one will be our legal business, which we charge on the number of fee earners, but this is not -- sort of once you hit a fee tier, then you basically can't drop down. So we might sort of introduce a secondary pricing meter with that. We may or may not. And those are the 2 that we have to sort of work on in, I'll call it, the shorter-term and then and possibly over a multiyear period of time, maybe modify what we do with Deltek, but not something in the short term.

Got it. Is your customer changing? Who you're actually targeting within each organization?

So interesting on this is it's hard to generalize across all 21 software businesses, but I'll take a stab.

Please.

Our customer profile is typically a medium-sized business, think a 3-facility daycare center or a 4 facility closed door pharmacy or a 500-person insurance agency -- we certainly sell to large enterprises in their government contracting business, and there's very large insurance companies for sure. But the preponderance of the number of customers that we deal with, we are the IT budget. Like they don't have Salesforce administrators, they don't have big IT departments. We kind of like we are the IT department, if you will. And so in that regard, the customer is not really changing because we're generally selling to the principal or a very close proxy to the principal. Maybe the CFO of a large insurer say, maybe not the owner or the CEO. And so in that regard, no, it doesn't change that much. At the enterprise level, like we do sell to larger customers, I would say that we have typically sold and that's where it might change a little bit more to the end user than the IT department where we're sort of selling like a tool to make the user more efficient than they have to take it and sort of we will work with the IT department. But that's like a minority of the sales motion that we have.

That's helpful. You mentioned earlier, Neil, that you're trying to take all your productivity gains to play offense, right? So maybe discuss like where you're seeing the biggest kind of efficiency gains within your own portfolio? .

Okay. So I guess said earlier, a couple of years ago, when the bottle went off is like how do we go drive product velocity and play offense and get first products to market and sort of accrue the AI-agentic learnings at an accelerated pace and compound that. And so without question, where we're seeing the most productivity in the organization is where the emphasis has been in the software development life cycle. And it's all -- it's direct -- it's straight up offense. As we get -- we have more product velocity, and this is not just using the models to code faster. There's a whole workflow coatings, but a small portion of the software development life cycle, it's the whole software development life cycle that we've got a very clear model on what it looks like, the way you organize your people, smaller teams, closer to customers, solve problems versus build features to drive product velocity. And we've seen really encouraging, super encouraging results in pockets of enterprise and are on track to sort of be fully agentic SDLC across all 21 software businesses by the end of this year. So like not 30% to 50% improvement, but multiple fold like 3x to 5x sort of product sort of velocity sort of releases. So that's great. We're just now have the organizational sort of mandate to get other functions together to sort of share best practices about what they're doing on the AI productivity. Finance team, sales team, support team, services team, implementation teams. There's pockets of real goodness because the incentive is like, hey, if you can drive productivity to go put more in the go-to-market or product and every company's incentive to do that is just how do we sort of get the competitive juices flowing inside the portfolio like, oh, that company is doing it. I want to do it better. I want to at least adopt what they're doing. And we're just like literally this month, getting the other functions together sort of tap into that.

Great. I'm going to ask one more question, and I'll turn it over to the audience before you move beyond this topic. But -- just in terms of contribution, right? So I think you've said AI is not going to be a material revenue contributor this year. You're going to see it flow in bookings first and then into some of your recurring revenue streams. How are you thinking about potentially the 27 contribution? And when would you expect it to really kind of inflect for you? .

Well, I think the second half will be telling in terms of the early signs of bookings. And then from there, I think it's just adoption. So I think adoption is the ultimate question. I think we're getting signal from customers like, wow, this is really going to kind of revolutionize what I'm doing? Or -- so I think the signal is really strong, but in terms of like when it's actually going to materialize is the $100 billion question, and we're just not there yet.