Okay, awesome. Thank you everybody for joining today. I'm Dylan Becker, the research analyst here at William Blair that covers Bentley. For all of the necessary disclosures, you can find those at williamblair.com. It's a pleasure to have Greg Bentley, the CEO of Bentley Systems, here with us today. Thank you for taking the time. Maybe as a way to start, Greg, there's a lot of generalists here. Maybe some are more familiar with Bentley than others, but as a way to level set, give us a backdrop, a perspective on the evolution of Bentley and what you guys really do for infrastructure.
Dylan, thank you very much for inviting us, and appreciate everyone's interest. As I'm reminded by our company's 40th anniversary in the coming quarter, we began with 5 Bentley brothers. The other four are proper engineers and worked from the start on a consistent and continuous software platform for infrastructure engineering applications. And these 40 years on, our family still represents the majority of the board and the majority of the economics in the company. So I'd like to think that we have the same long-term investment alignment as you would like us to have. We generated the most, and developed and acquired the most comprehensive, integrated set of applications for infrastructure engineering.
We also focused from the start on a subscription commercial model, which today is 91% of our revenues, and we have had a long-standing and direct relationship with most of the major infrastructure engineering organizations in the world. They are doing what I consider to be the consensus most important work in the world. Infrastructure underlies the quality of life, both through our economies and our environment, and that's the work of civil and structural and geotechnical engineers. Our company is not a household name in general, but it is to those folks who do that important work in the world, Dylan.
And it's a fair point as well, Greg, because there's a lot of complexity. There's a lot of stakeholders that are trying to solve this problem. Is there a fair way to think about what that typical customer looks like for you? What kind of types of projects are they working on?
Well, when we break down our ARR by end market, the largest infrastructure engineering sector is for, we say, public works and utilities. And that would include roads and bridges, rail and transit, grid, water and waste, municipal applications, and most of those we consider we're the leader in the world. We do also participate in, we say, the resources sector. That would be mining and renewables, for instance, where we're also the leader in terms of these tools by which the engineers do their work. They capture their work, they express it, and they deliver it through our software. And then there's the privately funded sectors, if you like, of industrial and commercial and facilities, and those are much smaller for us.
60% for us, public works and utilities, 25% resources, and the balance, industrial, 10% commercial and facilities, that is, vertical buildings, is about 6% of our footprint. Another way to characterize our users is that 60% of our revenue comes from accounts who spend $250,000 a year or more with us. We have accumulated, become, in many cases, their most comprehensive supplier. There's no account that accounts for more than 2% of our revenue. And then geographically, we've scaled up everywhere as well. North America is the minority of our revenue at this point in time.
And we'll get into maybe why that may be around global mandates and things like that. But what about from a financial perspective? You noted it's not a household name, but maybe it should be, given the quality of the financial profile here.
Well, we've become what I suppose would be called a classic compounder for some time now. Our... The bulk of our revenue, as I mentioned, is recurring subscription, and we endeavor to become 100 basis points more efficient every year. We can accomplish most of that through our 92% direct sales model. If you think about it, that allows us to become the spend, you know, if our, if our NRR, our net revenue retention, is 108%, and it costs us 4% more each year to support our direct sales force, that falls into our margins. We measure our margins net of including stock-based compensation. Stock-based compensation is important for a software company, but we are measured and disciplined in how we do that.
So we're up to about 26% operating margins after stock-based compensation. And we have considerable visibility, as I say, into what we do. Our subscription revenues, the bulk are recognized ratably throughout the year, 75% of our subscription revenues. We have no multi-year billings or multi-year revenue recognition, so it's all rather straightforward. And we're pretty efficient at conversion of our adjusted EBITDA to cash at about 80%. And over the period of time since 2018, our operating margin dollars measured that way with stock-based compensation have compounded at 16% per year, which is a combination of our low double-digit ARR growth and 100 basis points per year, which is our internal commitment to getting more efficient each year.
Very, very helpful. So we've kind of gone over the broad overview of the business. Maybe let's switch to the industry perspective. There's a lot of emphasis on infrastructure globally. There's a lot of efficiency needs associated with that as well. I guess, what, in your mind, is incentivizing kind of the most change from an infrastructure development perspective?
Well, it is the world's converging priorities, and I don't even think this is a matter of debate. Energy transition, climate resilience, and importantly, adaptation, generally maintaining the fitness for purpose of our infrastructure as we use it differently through and after a pandemic and so forth. That is all the work. It's the, it's the concern of everyone and the priority of everyone, but it's the work of civil and structural and geotechnical engineers, to the point where they are at the limit of their capacity, as more retire each year than are graduated.
Going digital has become, in the pandemic, it was a necessity, and since then, by virtue of the resource capacity constraints, has become quite the priority across the world. Everyone is at the limit of what they can accomplish. Going digital is the way to get more done, and has improved our growth rate over the period we're talking about.
And you see this, right? It's very apparent globally. There's multiple trillion-dollar initiatives out there around kind of investment, and spend. I would assume that that takes some time to flow through into the model, too, and those projects to get executed. Again, these are very large-scale assets. I guess, what's the right way of thinking about global initiatives on investment towards infrastructure and the durability of how that plays out over time, if that makes sense?
Well, I think the right way of thinking about those infrastructure expenditures is as investment. We don't tend to put the word stimulus after infrastructure now as once we did. And even in the United States, infrastructure is now a matter of not only daily discussion, but bipartisan support, since we see the benefit of expenditures. And it's not as if we have a choice about those expenditures. The programs you're referring to are multi-year programs, but every year, our infrastructure is a year older, and fitness for purpose changes. Resilience requires us and rewards us for expenditure on infrastructure across the world. And I think there really is a change in that thinking.
I might say that one can imagine over a long period of time with government fiscal situations, that most infrastructure will need to be privately funded, but that will be yet a further acceleration in going digital because all of the world's P3s are design, build, and operate concessions, and the software for digital twins is even more valuable over a evergreen lifecycle perspective. So I think there's lots of consistent and predictable work ahead.
I do want to get to the asset analytics point, but before we get there, you've kind of called out that the business is at about $1 billion in ARR today. There's a lot of incremental spend to come. What's the right way of thinking about the overall kind of market opportunity? I know you guys have this slide of engineering product engineering software moving towards the spend levels within infrastructure engineering, but how should we think about the opportunity?
Well, when we think of our addressable market, we compare infrastructure engineering to product engineering, for instance. The engineers cost the same. They're in the same place. But in product engineering, what's spent on software for each engineer is a multiple of about 3 times what's spent in infrastructure engineering. If we were to look at electronics engineering, it's another comparable multiple of product engineering. And you might say those engineers are in R&D functions, and their companies and organizations are, you know, gonna fully spend on what they're worth. In infrastructure engineering, engineering firms, and let me back up and say that half of our revenue are from owner-operators of infrastructure. So these are largely public owners and are all asset-intensive organizations to own infrastructure.
The other half of our business, literally half, is from their supply chain, the project delivery firms, the engineering and construction firms. Those have primarily had the commercial model of billing for the hours that they spend, and that's not something that would incent them to get more efficient over time. But in the current environment, they're as busy as they could possibly be. They're facing these resource constraints. It's not possible to hire more or new engineers, and the backlogs are higher than ever, and going digital, using more specialized software products of ours, is the priority now to start to catch up. So in infrastructure engineering, our accounts, the engineering firms, spend on average $1.41 per hour on our software, and that's an hour of an engineer for which they bill on average $150.
So 1% substitution of technology for labor, and it's the software that makes that a valuable hour and can make it an even more valuable hour. So I think there's lots of headroom to catch up with the rest of the economy and the other categories of engineering in what our users do by using more and more specialized software.
And we'll get into this with E365, because I think that's a part of the initiative of accretion within those customers as well, too. But where can that percent of spend end up going to, within that kind of billable hour framework, if that makes sense?
Well, I guess you could ask in other parts of the economy. Probably our, our group here is familiar with Fintech. What is the percentage of software spend in relation to labor spend? And I, and I suspect we're nowhere near diminishing returns. I, I don't think so.
Sure. Okay, going to the Asset Analytics side, you—since you called it out, the ability for you guys to provide kind of that audit log between the design stage, the execution stage, and then moving into the operational stage. You aggregate, you collect a lot of information and data. You touched on digital twins, but what's the right way of thinking about the leverage ability of your data asset over time?
Well, Dylan, here's the digital twin opportunity. We know that the modeling and simulation done with our tools by the world's civil and structural and geotechnical engineers is only used once during the CapEx delivery stage, the design and construction stage of the asset. We know that because the data is captured in the format that can only be opened in by our applications, and we know when our applications are used. But the modeling and simulation logic, the work of the civil and structural and geotechnical engineers, the geotechnical engineers do the foundations below the ground and model the environmental impacts of water permeation and vulnerability. That can be used and should be used throughout the life cycle.
First of all, construction should start with the 3D design models and be done in 4D to simulate the occupancy of space and time during construction. It shouldn't be dumbed down to 2D digital paper. And then, the modeling and simulation logic can be used to maintain the fitness for purpose of the transit station or the water treatment plant or the airport terminal over its lifetime, and in particular, in a digital twin, which would bring together, if you like, the digital context, the surrounding environs through drone surveying and continuous surveying, so that you can always have an as-operated reality model. And then bring in. That's the OT. Bring in the ET, the engineering technology, the digital components, the design logic, and the simulation, and finally, consider the IT, the maintenance records.
If you put that all together and apply AI to it, you can operate the asset to optimize the maintenance so that you don't do unnecessary maintenance. You're only doing that maintenance which your digital twin shows you was effective, and when it was effective, and look across your fleet to see that, and monitor so that you continually rerun the structural and other simulation engines to make sure it's safe and see when an intervention is needed. That's what we can do to extend the lifetime and maintain the resilience of our existing infrastructure assets. But because we can do that, and at the same time spend less on maintenance because of the digital twin insights I'm describing, it can actually cost less than nothing to go from using the engineer's work only once to using it continually in a digital twin. Tremendous opportunity.
Sure. Sure, maybe it's a good segue into the commercial model, too. You called out the fact that the business is largely recurring in nature, but how do you actually charge for a lot of these kind of types of solutions? What does the push towards E365 mean for kind of more sophisticated application use cases and the incentivization of usage?
Well, traditionally, we charged per year for a subscription. And we realized that the half of our users whom are engineering firms could often recover their software spending from the owner-operator, their client. The owner-operator understands they get a better quality of work and fewer hours are spent if more software is used. But the engineering firm would need to substantiate what software they're using to recover the cost at a particular time. So our E365 commercial program, which we introduced for our largest accounts and have been continuously expanding, it's, it's up to about 42% of our ARR at the moment, charges literally on a consumption basis per application per day. And then its, its cost can be literally attributed to the project and often recovered by the engineering firm.
When we worked out what to charge per application per day, we marked that up in relation to our history of what you would pay per year and the number of days on average used, so that it covered our cost of assigning our engineering experts to each account, our success team. We have 1,000 of our 5,000 colleagues whom are civil and structural and geotechnical engineers, and they were previously available to help our users if they could engage them and hire them, and we could propose it, friction that prevented that from happening much.
So now, under the E365 program, each account is assigned a team that helps them introduce new digital workflows because they, they want to use more software, and we can help them identify how especially to go from more generic software of ours to the more specialized software of ours. Because no, no one, no engineer is working on generic infrastructure. They're actually working on a roadway or a bridge or a tunnel or a floating wind turbine, for which we have ever more specialized software that costs more per day. But we can introduce those workflows and help with competitive displacements and so forth. So since we have changed to a consumption model for our enterprise accounts, it's been to everyone's benefit in what we call our application mix accretion, the upsell to their...
where they spend their hours and days using more specialized and more expensive software, getting to that, increase of a $1.41 per hour to a much higher number over time progressively, 'cause we don't see any limit to that. We help them do that in the E365 program, and last year, that application mix accretion alone accounted for 6% of ARR growth of our total of about 13% last year.
Mm-hmm. And so it sounds like there's a lot of exciting initiatives underway with the business and a lot of kind of competitive differentiation overall. If we were to flip to the competitive side of the equation, you guys are very in the weeds on kind of particular infrastructure domains, and you can address the end-to-end life cycle of a project. But maybe what's the easiest way to contextualize the competitive landscape and how your kind of differentiation has driven maybe a more formidable moat over time?
Well, we are most comprehensive in what we do. Another way to describe the sectors I talked about earlier is infrastructure can be thought of as vertical infrastructure, and that's buildings. And that is only, that's the commercial and facilities. That's 6% of our business. The rest of infrastructure is horizontal. It's networks of road and rail and grid and water and the things that connect it all together. Our principal competitor is Autodesk, and Autodesk is most concentrated in vertical infrastructure, in buildings. And that's because Autodesk started on the PC, and PC had small address space, and buildings are smaller projects that don't last as long. In horizontal infrastructure, we started on the Unix workstations. Those are more challenging in terms of data requirements and so forth.
Each of we and Autodesk overlap between horizontal and vertical to some degree, but that's kind of how things break down in terms of competitive landscape for us.
Sure, sure. You did talk about as well, too, how the platform has evolved over the last 40+ years or so, give or take here. A function of that has been M&A, right? There's been an organic component and an inorganic component, but how do you think about running an M&A playbook, the kind of types of assets that you look for, adding value across the existing ecosystem?
Ecosystem is the right way to look at it, Dylan. We have done over 100 acquisitions over our 40 years, and I would worry about a software company that would acquire dissimilar code bases, and it's like throwing a boat anchor out the back. But from the very start, what distinguishes us is continuity and consistency, if you like, and my four software developer brothers... By the way, that's the only thing I've ever done for a living also. But they have made sure that we have a platform fixation over time, that we only develop something once and use it multiple times.
So our platform, first of all, MicroStation, is the platform for our vertical applications and our iTwin platform, our cloud platform, that has underlies the Bentley Infrastructure Cloud, that's now grown to be about a third of our revenues. Each of those have attracted an ecosystem of specialized developers, companies that develop specialized vertical applications for our platform, and we have, in the main, acquired the best of those over time so that we cover all disciplines and life cycle aspects. And so they have been pre-integrated, if you like, and have reinforced each other and have helped our profitability model over time.
Going forward, and by the way, that program has accounted for about 100 basis points of ARR growth over our history, which has come down in the past couple of years, and we have a particular interest now in acquiring earlier-stage companies in an opportunity I'm going to describe as asset analytics. The change is that we had, for a while, a venture investing program, and we would invest in fractional shares of companies investing in digital twin opportunities. But they would spend all their time raising venture capital, and that became less desi- less...
That became harder over the past couple of years, and we saw the opportunistic time had come where we could acquire the entirety of these companies, and we focused on those with AI computer vision to apply to our platform, so that instead of digital twins being a big opportunity with a long sales cycle, we can have instant-on digital twins. For instance, an example that we developed ourselves was for the cell towers of the world, where the tower cos, instead of rolling a truck and have a drone fly, creating overlapping imagery, which video is, and using our software to resolve that into an engineering accurate representation and identifying what's on their tower and the opportunities for additional revenue. Whenever you add something to a tower, you need to resimulate the wind resistance, the structural sufficiency, electromagnetic interference, and so forth.
And then over time, refly that drone flight program, do it autonomously, and use AI to track the corrosion and so forth. So that's our an asset analytics opportunity where we charge, not per user, which is the mainstream of our business today, but incrementally per asset. And we can get 3 digits per year per cell tower, and there are 3 million in the world outside China. And the most recent acquisitions to draw this back is our company named Blyncsy last fall, which does roadway operations and maintenance asset analytics from dash cams, and where we can turn that on the next day with our AI.
So we would like to be the processing, back office for the engineering firms to be able to add to these insights, their own proprietary analytics, and with their business development force covering all the owner-operators of the world and taking responsibility for that data as it becomes more valuable, compounded with AI, and enable the digital twin opportunity, instead of being a long lead evangelism opportunity, to be an instant-on, immediate, asset-light opportunity. And that, this year, we hope, will inflect our ARR growth rate, upward.
Perfect. We've got a couple of minutes left here, maybe two more to touch on. M&A is always one component of the broader capital allocation strategy, but there is deleveraging across the business. You have a dividend, you do proactively repurchase shares. I guess, what's the, the high-level thematic on, on your broader kind of capital allocation efforts, given you generate, as healthy of, of cash as you do?
Well, it was great to be a public company and be able to do the two platform acquisitions we have done. One is Seequent, a company we still operate under that name, which does everything we do for 3D and 4D modeling above the ground, they do below the ground. And below the ground is really different because you can't see and survey. You have to sample with drill holes and boreholes and use mathematics to figure out what's under there 3D. And that's largely been used in mining, but applies as well to every aspect of infrastructure, where most of the risk is below the ground.
So that company has, which was about 10% of ours when we acquired it for over $1 billion, has grown by more than twice our growth rate over time and a little bit less now with the delay in mining exploration, but that will come back, and it's still growing faster. The other platform acquisition we wouldn't have been able to do without being a public company was Power Line Systems, which is needed in the world for the engineering of transmission capacity. The single biggest springboard yet to come when the permitting delays get solved in the world to connect up the renewables capacity that's waiting to be able to be distributed and transmitted to where electricity is needed.
At any rate, we extended ourselves, issued convertible debt for that, and have paid down the credit line draw we used to acquire Power Line Systems. Our leverage would now be, on the order of 3x if you count the convertible debt as debt. And I think it's more likely to be converted over the next, couple of years. But we're able to be, flexible now to take, again, advantage of, any platform acquisition opportunity that would come along, but we're not working on any, and they can't be anticipated and, and planned for. In a year, we think of spending on the order of $100 million on the programmatic acquisitions. As I described, we also endeavor to repurchase sufficient equity between shares and convertibles to offset the $60 million-$70 million a year of stock-based compensation.
We pay a dividend that will, you know, approximate 0.5% yield or so, and feel that we should have some leverage in order to take advantage of our predictable cash flow generation.
Sure. And I think we're pretty much coming up on time here, but it was recently announced as well that you will be transitioning into an Executive Chair role with Nicholas stepping into the CEO seat. Don't think that this will be the last that we'll be seeing of you, but maybe give us a sense of kind of where you're thinking about allocating your time going forward here.
Well, as Executive Chair, my priority will be to support Nicholas. He's currently our Chief Operating Officer since the beginning of 2022, and he'll now have responsibility for our corporate functions of finance and legal. However, I have asked to be able to continue the responsibility for investor relations because I like doing this, and I'll continue to be our spokesperson for that purpose, continue to look after our capital allocation decisions, and for now, I'm continuing to shepherd this asset analytics incremental business opportunity to charge per asset and monetize that. Exciting time. It enables- Being an Executive Chair has enabled us to bring in, for our first non-family CEO, someone who's considerably younger at age 47.
We think continuity and consistency has been a useful differentiator for us, and that we can now repeat that cycle by bringing in someone younger. I'm excited about it myself.
Perfect. Thank you, everybody, for listening. Thank you, Greg, for taking the time.
Cheers.
We will have a breakout session in Gurney A, for those that are interested in joining. Thank you.