Tyler Technologies, Inc. (TYL) Earnings Call Transcript & Summary

Hello, everyone, and thanks for joining the 2021 UBS TMT Virtual Conference. My name is Taylor McGinnis, and I'm an analyst on the software equity research team here. And in this session, we are joined by the CFO, Brian Miller, of Tyler Technologies. Brian, thanks so much for taking the time today.

Yes, you bet. Good to be here.

Perfect. And before we get started, I would like to remind everyone listening in that you can submit a question via the Ask a Question tab in the upper right-hand corner of the screen, and we'll make sure to save some time at the end to address any of these questions.

So with that, let's dive in. Maybe to start, Brian, can you talk about the impact that Tyler Technologies might be seeing from the current public sector spending environment? I guess what trends are you seeing in terms of demand? And what are your expectations for budget growth next year? Have recent developments with the new variants, supply chain issues, labor shortages and others may cause any material changes in terms of what you've been seeing?

Yes. I'll start kind of with the last and work backwards. But the recent developments, we really haven't seen the trends change much recently with new variants or supply chain. Those haven't really been factors. What we have said and emphasized it in our most recent earnings call was that, in general, the activity in the market is pretty much back to or, in some cases, ahead of pre-COVID levels. So what we saw -- and just to step back a little bit, the biggest driver of demand for us is that governments have very old systems that are reaching end of life and almost have to be replaced. Everything we provide, all the enterprise applications we provide for government of power essential functions, whether it's property taxes, public safety, ERP, licensing and permitting, courts and justice, so all these things that we automate are essential functions. And governments, because they are not run for profit, they're not ROI-driven, and they don't have competition, they typically use systems as long as they possibly can until they're really at end of life. And when they get to that point, then replacing them is a fairly nondiscretionary decision. They do have some discretionary -- discretion around the timing, so in more difficult economic conditions or more challenging environments like we've seen in the last couple of years, there have been some delays and people that have said, "I would buy a new system now, but I'm going to wait until things are a little better," that doesn't mean they can wait 5 or 10 years. And so what we've seen is that those pauses that we saw early in COVID around sales processes, around new RFPs coming out, around existing processes slowing down, whether it's because they were distracted and focused on other priorities related to COVID or they had concerns around the impact that COVID might have on their budgets, we clearly saw delays. We saw a lower level of RFPs, less activity in the market in 2020, started to see those processes start to resume, and RFP starting to return to more normal levels as we moved into 2021. And now, like I said, we're about generally kind of back to pre-COVID levels. So the demand didn't go away, but there were delays. What we do think is that -- well, there's a couple of things. One, with regard to funding, government budgets are really now in pretty good shape. The impacts of COVID on their revenues were not nearly as significant as they feared early on. Revenues were minimally impacted. Local governments, which make up the biggest part of our customer base, about 85%, are primarily funded by property taxes. And property values, as you know, are very strong in most places. And so there's not a lot of pressure on property taxes. And a lot of the sort of onetime cost impacts around COVID are largely behind them. And now we've got this massive stimulus coming from the federal government, which is providing $350 billion of direct aid to state and local governments, another $167 billion to schools that can be spent between now and the end of 2024. So we think that in the short term provides them with a lot of confidence in their budgets, but in the longer term, gives them a bunch of incremental money that can be potentially used to accelerate the replacement of systems or advance their systems upgrades and IT infrastructure faster than they would have otherwise done.

Got it. And I'd love to dive in maybe a little bit deeper to something you mentioned in the first part of your answer, which is just on some of these end-of-life dates and that really being a catalyst for a lot of the activity that you guys see. So can you maybe talk about any key end-of-life dates that are important drivers for growth for you guys? If some of those pauses that like happened, which ones are those where you might have like a backlog coming, right, of some of these migrations to your systems, mind just providing more color there?

Sure. Yes. It tends to spread itself out. I mean there's not like one catalyst or one sort of time point where a lot of systems get replaced. Some governments will skip entire generations of technology. I'd say the average system we replace today is probably 20 years or more old. There's a steady flow of systems that were bought in 1998, 1999 because of Y2K. If you probably weren't old enough to remember Y2K to have been there, but when systems were replaced because they thought they wouldn't work when the dates change. So those systems are now 20-plus years old, and a lot of those are ripe for replacement. But it's not uncommon that we replace systems that are 40 or 45 years old. We're replacing -- for example, one of our big projects right now is the court system in Cook County, Illinois. It's the second largest court system in the country. It is a 45-year-old mainframe system written in COBOL that's been used since then. And now literally, there aren't COBOL programmers around. So they -- one of their fears is that something can break that they can't fix anymore and they can't have a court system that stops functioning. So these things happen over a long period of time. It's sometimes difficult to accelerate that demand within a given jurisdiction, but it creates sort of a really steady market. We've had low double-digit growth, compound annual growth rate of low double digits for 15 years. So it creates a really nice compounding, but it's never an explosive market. We do have long sales cycles as well. So once they decide to replace the system, even in a normal midsize deal, it might be a year from the time they make that decision to the time they get to a contract. But in something like a large statewide court system, it could be 3 or 4 or 5 years. So we have a lot of patience for our market. Competitors or especially horizontal companies that don't just focus on the public sector often don't have that same patience. So it's kind of all of our business revolves around a market that moves slowly and makes decisions slowly but creates this really nice, steady, sticky market.

Got it. And then I know like 3Q typically is a big quarter for government or public sector spending. So could you maybe talk about in the conversations that Tyler Technologies had with some of the organizations there, I guess, anything relative to your expectations that you found interesting? You talked a lot about like the funding that could be coming over the next couple of years. What does that mean? You could potentially see almost like a pull forward, right, in some of these like modernization efforts that maybe you didn't historically have a catalyst for. Any thoughts there?

Yes. Our sense from talking to clients and prospects is there will be some replacements that are pulled forward. So someone that might have thought that their system had -- they know it's not state-of-the-art. They know it doesn't have a lot of features and functionality that would be considered kind of table stakes today, but they might have thought it has 5 years or 10 years of life left in it. But particularly around the demands that came out of COVID and needing to work remotely, something that governments struggled with like a lot of private sector did, but probably more so, that they found that those systems really didn't function well, if at all, in a remote environment. So again, going back to like a mainframe court system, it only works effectively if employees are at their desk, at a terminal that's wired to the mainframe. And they don't have remote access capabilities. They don't have the ability to conduct a virtual court trial like you can with our system or to select a jury in an entirely virtual process. So I think with the -- what some governments are increasingly saying is that we don't have as much life left in these systems as we thought we had, and we'd like to replace them sooner. And now the availability of this money coming from the federal government may help us do that, depending on how we allocate it. Our sense is that very, very little of the American Rescue Plan money has been allocated or committed yet because they've got until the end of 2024. They're kind of in the early stages of figuring out where they're going to spend it. We're certainly doing our best to identify areas where they can spend it with Tyler and be engaged with them. But I think this plays out over a longer period of time. And even if they've decided to pull something forward, again, it still may be a long sales process. So it does feel like a tailwind to us but kind of more of a midterm tailwind than next quarter or the quarter after. So there is a little bit -- and I think there are things that governments might have previously thought of as sort of nice to haves but not crucial. And in particular, we look at our Socrata data and analytics platform. So it's a business intelligence analytics platform that interfaces with our systems or other systems to make data available and usable for decision-making. And I think governments are increasingly thinking of those kinds of solutions as more have to haves than something that they might have previously thought was just nice to have. And so the ability of additional funding to potentially pay for those is certainly helpful.

Yes. And then maybe diving into the American Rescue Plan a little bit deeper. So you mentioned early stages. It sounds like there's still things that need to be figured out, but any early trends that you guys are seeing or that find interesting in terms of how some of this could get allocated in what areas or how discussions on this have trended over time?

Yes. There's -- again, it's pretty new. The CARES Act, which was the first round of stimulus, was a smaller amount. It was just really for state governments and large local governments, so a smaller subset of government entities. And it was really meant to replace or to compensate for expenses associated with COVID. And so we have a shorter time period for using it, and it was much more restrictive in terms of what it could be used for. The American Rescue Plan is much less restrictive on what they can use it for. So almost anything that could come from Tyler qualifies. We do believe that -- and they've got until the end of 2024, as I said, to spend the money. They get half of it this year, half of it next year. And it's $350 billion for state and local governments. It's spread across virtually every local government or every government entity gets something and then another $167 billion for schools. So it's a massive amount of money. I'd say that -- like I said, not much of it we can tell has been allocated yet. We're certainly engaged in a lot of discussions with both existing customers and new customers around that. We believe that with the focus and the sort of the spotlight that's been placed on a lot of government systems in the last couple of years and the increasing awareness around some of these aging systems, we believe that IT, IT infrastructure will get sort of more than its proportional share of that spending, but we really don't have hard numbers or much actual experience you have to point to. We're starting to track actual opportunities that are identified as specifically being funded by the stimulus. And I think we said last quarter, we had about 40 of those in our pipeline. I don't know, somewhere between a couple dozen and 40 that were deals that were specifically identified as stimulus funding-related, but we expect that to grow significantly over the coming quarters.

Got it. And then maybe digging on some of like those deals. Do you think that -- and this gets back to something that you talked about earlier, but do you think this has the potential to decrease maybe some of like the uncertainty on like funding these projects and the impact there being maybe shortening even at like traditional sales cycles where you talked a lot about how those have been longer in the past, more providing durable growth? Are you guys seeing early signs that some of that [ traditional sales cycles are shortening ]?

I'd say that is a possibility because there are a number of factors that go into making these long decisions, long processes. One of them is funding and when -- what budget that comes out of and when. So giving them a lot more flexibility around their budgets or some extra money there logically would seem to take that part out and maybe potentially accelerate them. We have, over recent years, been able to move from significantly in our processes from a few years ago almost every process was a fully competitive RFP full-blown competitive procurement process. And today, about 1/3 of our deals take place as a sole source, so without a competitive process. Sometimes that's through a joint purchasing alliance or a state contract. But often, because they already have a Tyler product, they can acquire other Tyler products sort of as a piggyback without going through a full-blown process if they know that they want to buy a Tyler product. So that's been able to shorten some of our cycles. And given the size of our customer base that as that continues to grow, that gives us more and more opportunities. And we believe that that's an important aspect of the NIC acquisition as well because they have very broad contracts at the state level through which Tyler software can potentially be sold as well.

Got it. Yes. I want to touch on the NIC acquisition. But before we go there, maybe a broader software question for you. So it seems a lot of the investment over the last 1.5 years was really focused on front office type initiatives or moving infrastructure to the cloud. Tyler has a lot of back-office ERP-related products, too. So curious, did you see like some of those delays? And what are your thoughts -- or what are you seeing in terms of a recovery here and an appetite to modernize some of those back-office systems?

Yes. When you think about where Tyler is, again, excluding NIC's business, but Tyler's traditional business is really in back-office software, running central back-office functions, whether it's ERP and accounting, human resources and payroll or licensing or property taxes or managing the courts, that's really kind of what we do. I would say most of the -- some of those areas are affected -- their individual dynamics around some of those sub-verticals. So public safety, for example, I'd say, was generally less affected over the last couple of years. Public safety agencies never went remote. They were always there. So they weren't able to go work-from-home or shut down for a while. So mostly those processes after some sort of brief pauses continued normally. ERP was probably the most affected in terms of delays and pauses and processes. And that's where we've really seen the volumes over the last 2 or 3 quarters kind of come back to normal. So most of our products we're seeing operate in a more normal demand environment now. NIC is more on the front end. So they're providing the digital access to these back-office systems, mostly at the state level and facilitating payments. So they've seen nice growth in their business because people are being, in some cases, forced to; in almost all cases, want to do more business digitally with government rather than going down to the DMV office or sending in a check. So they've really benefited from the trend towards more online interaction with government. And so now we really kind of have both the front end and the back end through the combination with NIC.

Yes. And I'd be curious for your thoughts on how you expect to see now with the NIC acquisition and your historic back-office type offerings. How do you expect that mix to evolve in the near term as we get further away from some of the pandemic impact? So is it that you might see more of a rebound like you're talking about in terms of the back-office stuff and that could be an increasing portion of the mix? How do you think about the durability, right, of some of NIC's front-office offerings? How should we think about that over time?

Yes. I mean the 2 businesses are just really highly complementary. So we do both in terms of what we do for government and the services and products we provide and, really, the levels of government that we serve. So Tyler is, in our software business, about 85% local government, cities and counties to school districts, local agencies; 10%, 12% state; and less than 5% federal. NIC is 95% state government and maybe 5% federal, almost no local government presence. And NIC is providing the digital front end, the interface to these back-office systems to facilitate transactions, the state governments or payments in the state governments or accessing information. And so we believe there's a tremendous opportunity for us to leverage their state relationships, their intimate knowledge of what's going on in the state, what the initiatives are in -- at the CIO level, agency level and surface opportunities to sell Tyler software products into state governments through those relationships and through those contract vehicles. So that's a market where we really -- we've got a number of products that are applicable at the state level, but we haven't really had the sales channels or those long-term relationships. So really, NIC brings that to us, and we think there's a tremendous opportunity to move more back-office software systems into state government and expand our presence there. And then on the payments side, NIC has a really robust payments platform and really strong payments capabilities, mostly at the state level, where if you're renewing a driver's license or hunting and fishing license or applying for a professional certification, and you pay a fee to the state, NIC would handle that transaction but also do the payment processing. And Tyler has, for some time, had a goal of trying to be more involved in payments streams around our clients' businesses at the local level but have been really kind of in the early stages of that. NIC brings us that technology, that experience, that infrastructure that we believe now we can leverage our local government relationships to push that payment processing business down to the local level and significantly expanded. Just sort of as a point of comparison, in 2020, Tyler basically as a reseller of payment processing services was involved in about $3 billion of payments through our clients. That's a really small portion of the payments that are flowing into our customer base. NIC processed about $24 billion worth of payments. And now they have a new contract with the state of Florida to do all this payment processing for Florida, which will add another $50 billion of payment volume. So their scale is significantly different from where Tyler has historically been. So we really kind of see both areas as significant growth areas, and the combination provides an opportunity for both of us to have incremental sales in markets that we wouldn't have had on our own or wouldn't have had for a long time on our own.

And can you perhaps provide a little bit of an update even on the time line of some of these things? When do you expect to start to see some of this -- some of these synergies that you're talking about materialize? How is integration going so far? Can you provide a little bit more color there?

Yes. Sure. So we closed the deal in late April, by far, the largest deal in the company's history, a $2.3 billion purchase price. First time Tyler has acquired another public company. So they were in a different set of challenges and things to address around that, but also much more sophisticated than our typical tuck-in acquisitions. So we've had it since April. Just last quarter was our first full quarter with NIC. And really, I'd say what we've been doing in these first few months, we talked about kind of laying out what our objectives were for the -- for this year. The first key objective was don't mess up the business. Don't let the integration, don't let the combination distract from either of the 2 businesses achieving their plans for this year, continuing on the paths that were very successful already. But secondly, the next priority was really to start to integrate and mostly around the go-to-market. We don't have a lot of technology integration. It's not like we're integrating software products or platforms. They're a front end, we're a back end. And it's really about how we go to market together to realize those sales opportunities. And the third objective was to identify -- start to identify a pipeline of those sales opportunities and start to execute. So what we've been doing in these first few months was really educating each business on what the other has to offer and what we do at a deeper level so that they can help us sell into the state market and we can help sell into the local market their products or their services. So we've done that over the last few months, educating the Tyler teams on the NIC contracts, the relationships, the initiatives in each state and where there are opportunities for Tyler and then educating the NIC teams on all of the Tyler products, and we have a lot of products. And so they understand what we have, what our strengths are, where we have customers, where we're pursuing customers and so that they can help identify those opportunities and ultimately help make those introductions or help lead their sales into the state market. So we feel really good about what we've accomplished in these first few months there. We have a cohesive one payments organization now and a go-to-market strategy that combines those -- the strengths of both sides. And we're starting to surface opportunities. I think we said we've got 40-plus concrete actual opportunities we've identified to sell Tyler products into NIC states that are currently in our pipeline. We've had a few successes early. We had a deal with Colorado in Q2, which is NIC's biggest revenue-generating state where we sold a licensing system into the state with the Tyler MicroPact platform combined with NIC Payments. We had a deal last quarter in West Virginia for a motor vehicle titling system that includes some Tyler components. So we're in the very early stages of starting to build a pipeline. Again, given the length of some of these sales cycles, I'd expect that we'll see continued success next year. I don't think it's going to be a big impact on growth rates next year, but we feel really good about where we are in these first 6 months or so down the road. And as we've gone through these processes, our enthusiasm around the business and the opportunities that can come out of it have expanded significantly for both teams. So it's a bigger opportunity than we, I think, we fully appreciated when we did the deal.

Yes. So maybe like a different angle to this. So you talked about state and federal, that being a small portion of Tyler Technologies today. And the NIC acquisition clearly gives you inroads to the state piece. But I guess is there also significant focus on the federal opportunity or potential opportunities there? I guess how should we think about the federal piece relative to state and local governments? And are there any exciting areas that you feel Tyler really hasn't addressed that could be significant opportunities?

Yes. Yes. We've always been -- since we started in this space in 1998, we've always been laser-focused on the public sector. And for the first 15 or so years, that really mostly meant local government, but we've certainly broadened our approach and started to have some Tyler business from the state level. NIC certainly brings us more there. But the federal government, obviously, is a big piece of public sector. Tyler also has a pretty clear focus on providing off-the-shelf software. Now our software is very scalable. It's very configurable and often is modified or configured for very large customers differently, but it's all around one core set of code in a given product. And so we don't want to be a project company or a custom development shop. And so historically, at the federal level, particularly somewhat at the state level, there have been a lot more custom development projects. And so that really didn't fit with our model as well. But increasingly, governments at all levels, including state and federal, have -- I think, are more of the mind that custom software development is something to be avoided if it can be. If there's a good off-the-shelf product like the Tyler product that addresses the needs of a large government, that that's a more reliable path than building a one-off custom system. And so to the extent we can address the federal market in a model that's compatible with Tyler's business model, it's a big untapped market for us. We did an acquisition about 3 years ago with a company called MicroPact, which really was our first big step into federal government. MicroPact is a low-code development platform like a ServiceNow or a Pegasystems, but only for government. And it is a platform that's used for a wide range of business process management or case management type applications and government. And there's a number of prebuilt accelerators or prebuilt modules. So things like we do all the background checks for the Department of Justice or workers' comp appeals or EEOC claims or licensing processes. So those types of things. And about half of MicroPact's business was with the federal government and about half with state and local. So through that experience we've had over the last couple of years, we've continued to uncover opportunities both with the MicroPact platform or with other products that could apply in the federal government. We also -- MicroPact brought us a pretty well-developed and extensive partner network in the federal space, which is something we haven't had in the past. And so they do a lot of work, either sales or services, through partners. And so that's been helpful as well. So we do have an appetite for more business at a federal level. And whether that growth comes from organic growth with our existing products or more acquisitions in the federal space, we just barely tapped -- or dipped our toe into the federal space, but we would expect to continue to grow there, both through MicroPact, through NIC and potentially through other Tyler products.

Got it. And then I'd love to ask you a question on the competitive landscape. So I think as like a lot of even larger SaaS companies mature, they're starting to get more of like a vertical focus. And you hear a lot of those companies talk about this. So do you -- like, I guess, who do you typically go head to head with? Do you see any risk from some of these broader ERP software companies or others that are starting to create a more vertical approach?

Yes. We have different competition in each of our sort of sub-verticals or product areas. Tyler is definitely unique in terms of the breadth of products we have. And in most of our products, we just compete with companies that generally are narrowly focused from a product perspective and often narrowly focused even from a geographic perspective. So a company that does court systems only, they don't do public safety and they don't do licensing and permitting. And it may be a company that just does court systems in California. That's just kind of where their market was and they're very specialized or property tax systems in New York and New Jersey. We're unique in terms of our geographic -- the breadth of our geographic focus and the breadth of our products. And that gives us a couple of advantages, one that -- a lot of scale and a lot of development resources. We're able to leverage a lot of development. So we have common foundational elements, one workflow engine, one payment platform, one security and sign-on [ and ID ]. So those create advantages from having more Tyler products because they work together in a suite. But -- so each of our verticals have different top competitors, and they're good competitors in each of those. The one area where we do tend to see horizontally focused companies would be ERP, so accounting, human resource, payroll. There are differences in public sector. They use fund accounting. It's a different basis of accounting. So every ERP competitor doesn't necessarily work in public sector, but those that you'd expect, Oracle, SAP, Workday, Infor and a number of others do. I think our -- and we've competed against those businesses for a long, long time. The difference is our vertical focus. And even as some of them may talk about having an increased vertical focus, we tend to see their interest in public sector often sort of ebb and flow. It doesn't seem to be a consistent priority. I think part of that may have to do with the pace at which our market moves. So it's going to be a much slower-moving market than almost any private sector market. And so I think their patience for our space sometimes isn't there. And so they may be in our space when other markets move more slowly, but they turn their attention elsewhere. And so we haven't seen that same sort of consistent investment from a vertical focus and public sector. And our advantage is that we not only have like accounting, human resource, payroll, but we have dozens of other applications that are government specific that are all integrated, so a utility billing system, a licensing and permitting systems, land development systems, even down to things like parks and recreation or cemetery management, things that are very government-specific that those companies don't have and we've had for years or decades. So we've got this deep functionality that it takes a long time to replicate. We've got a very clear cloud strategy. So traditionally on-premises products, migrating them in an accelerated process to the cloud. So we compete very effectively there. So yes, it's a limited part of our business where we see horizontal flares, and they've always been competitors, and we've always competed effectively against them. But we pay a lot of attention to them. I mean, Workday is a great company, great products, again, not nearly the depth that we have in the public sector. And again, they don't have all of those other products that we believe create a -- not at all at once, but over time, a real value to a public sector entity to have a court system, a public safety system, a land record system, a tax system, all of those things are multiple things from Tyler because the more you have, the more they work together.

Got it. That's helpful. And then maybe turning back to the NIC acquisition. But given that most of your business is from local governments and their budgets are largely made up of property taxes that is considered highly stable, as you work with the NIC business to expand in the state and federal use cases, are there any potential risks we should be cognizant of given the greater volatility of funding in some of those other areas?

Yes. You're right. I mean, local government -- probably, I think most people don't appreciate the stability of most local government budgets. Certainly, they vary somewhat from place to place. So New York City -- what makes up New York City's budget is very different than what makes up Plano, Texas' budget, but most local governments are primarily funded by property taxes, and in most cases, more than half of their revenues. And over a long period of time, those are generally pretty stable, but particularly now where in most places, property values are very strong and there's not a lot of pressure on property taxes. So that impact, as I talked about earlier, has not been as significant. State governments tend to be more reliant on things like sales tax or, in some cases, income tax or different kinds of revenue streams. But generally, I think the overriding thing is the products that we provide are all mission-critical. And so regardless of where the budgets come from, the things we provide tend to fall very high in terms of the priority of spending. NIC's business itself, their portal business is self-funding. So they really aren't a line item in a budget or they don't get paid -- they don't get checks from the state. It's funded by -- all the services they provide are funded by user fees, transaction fees and payment processing fees that the users or the citizens of the businesses that interact pay. So you pay $3 to renew your driver's license, and that goes to NIC, and that helps fund this overall relationship. Insurance companies pay NIC to access driver history records to make -- to set rates and make decisions around ensuring car insurance policies. So that funds a lot of the portals. So NIC's business tends to be self-funding, but each of the -- each level of government and each government within those levels may have slightly different funding models. But generally, we fall in a pretty high priority in terms of what things that we provide are.

Got it. And then moving to your cloud mix and your subscription business. I'd be curious like in the pandemic and in terms of like trends and what you're seeing today, how would you characterize the appetite to migrate to the cloud and some of your subscription SaaS offerings? Maybe we can get an update there. And then in terms of how to think about this transition and the potential financial impacts there, any updated thoughts there?

Yes. Like a lot of things, government has been slower to move to the cloud than the private sector, but there's definitely been an acceleration in the last few years. And our cloud strategy really goes back almost 20 years to when we started offering a hosted model, really taking the same systems that were installed on-premises and hosting them in a Tyler data center than customers paying on a subscription basis. And so we've had a very gradual evolution from on-prem to the cloud. And until really a couple of years ago, we described ourselves as cloud-neutral or cloud-agnostic. So we let customers decide which way they wanted to acquire the software and how they wanted to run it. And we didn't have a strong preference. And so over time, our business gradually shifted to more customers choosing to be in the cloud. A lot of that driven by their challenges around running their infrastructures, so their ability to hire and retain and pay market rates to highly skilled IT people, the challenges of an aging workforce retiring and having to replace those people, concerns around security. So a number of things that have continued to make the cloud more attractive to them. And I think 2019 was the first year that more than half of our new business came to us through the subscription model or through the cloud. Last quarter, it was 75%. Some of our products is -- some of our products are already cloud-only. But a couple of years ago, we signaled a shift to really from cloud-neutral to cloud-first. We entered into our partnership with AWS. So we're starting to migrate -- or starting to put new customers in the public cloud at AWS and have a strategy over time to migrate our existing customers that are in our data centers also to AWS and get out of the data center business. But we still got -- even though the majority of our new customers are coming through the cloud, we still have -- of our 27,000 software installations, only about 4,000 are in the cloud today, and we've got 23,000 on-premises customers. So the opportunity over the next several years to migrate those -- to accelerate the migration of those from on-prem to the cloud and effectively double the revenue stream from those, which is typically kind of a 2x relationship when we see customers in the cloud versus what they would be paying us from a maintenance perspective. So we've got $475 million a year of maintenance revenues that can potentially double as they move to the cloud. We're working through processes of optimizing our products to be efficiently deployed at AWS and making changes to the architecture of some of our major products. We said we expect by the end of 2023 for that process, all of our major products, to be cloud optimized to be deployed with AWS. And so that really will kind of be the -- one of the gating items for us to start the acceleration of the move of on-prem customers. But we would expect that over the next couple of years, we'll move from even the last 25% of our new business that's bringing -- coming to us on-prem will move to -- coming to us virtually 100% of our new customers coming in the cloud and then to start accelerating the migration of our on-prem customers. We'll still support our on-prem customers for a long, long time, but there'll be new features that will only be available in the cloud. We do have a number of products that will only be available to new customers in the cloud starting as early as this year -- this time of year. So doing a number of things to kind of accelerate that, along with the general market trend of public sector embracing the cloud much more so. The one area that probably lags the most is public safety given the critical nature of those products like a 911 system, concerns around reliability of networks in some places. There's a little bit more reluctance in that space, but I think that will come around over time as well. But ERP, 90% of our new business is cloud, even some of the bigger opportunities like large port systems. We signed our biggest cloud deal ever about 1.5 years ago with state of North Carolina for a $90 million cloud-based court system. So definitely strong trend there and something that we're excited about in terms of providing greater revenue growth and certainly more predictability. So today, Tyler is about 80% recurring revenues, and that's streaming up as well.

Yes. And then maybe a last question for you on this. When you think about those 23,000 on-premise customers that will be -- or could be shifting to the cloud over time, any sense for trajectory there? You talked about, hey, that you're only going to be giving certain features and functionality in the cloud offering. If some of this additional funding could accelerate some of those migrations, is there any areas that you're pointing to in terms of how the trajectory of this could look or anything you would caution people on in terms of modeling this shift?

Yes. I just -- I would say like we're probably a couple of years away from when it really starts to accelerate. Today -- like last quarter, I think we migrated 70 or so customers. So at that rate, 23,000 takes a really, really long time. But we really haven't opened floodgates for that because we're still running pilots of each product in AWS. We're still working through this product optimization, which will make -- lower our hosting costs and make it more efficient. And as I said, we expect to kind of finish that by the end of 2023. And that's really, I think, when we'll start to accelerate that and go from less than 100 customers a quarter to hundreds of customers a quarter. So I wouldn't -- I keep it on the more gradual side for the next couple of years, then I think there's a point at which it significantly speeds up.

Perfect. Awesome. We will end it there. Brian, thanks so much for all of your insights and for joining us today. And to everyone listening in, thank you for joining as well.

Absolutely. Thanks for having me.