Ep 7: Mark Leonard: Quiet Compounder, Constellation Software

Mark Leonard, the founder of Constellation Software is a private man. He doesn't give interviews or speak at conferences but for twelve years he wrote annual shareholder letters that practitioners described as among the best capital allocation writing since Warren Buffett. Mr. Leonard founded Constellation Software in 1995 with CAD $25 million in seed capital. By the time he stepped down as President in September 2025, the company had completed over 1,000 acquisitions across more than 100 industry verticals, employed more than 50,000 people worldwide, and generated $11.6 billion in annual revenue, compounding shareholder returns at approximately 30% per year since its 2006 IPO on the Toronto Stock Exchange. This episode of Operational Velocity isn't about the stock price. We go inside the Constellation Operating System and the main interlocking pillars that Mark Leonard built over thirty years and break each one down at the mechanism level:

• Decentralization as Architecture 
• Operating Autonomy
• The Acquisition Playbook 
• Incentive Wiring 

References

• Full letter archive (primary PDF): sorfis.com/wp-content/uploads/2025/09/Mark-Leonards-Letters-to-Shareholders.pdf
• Letter excerpts with context: bauva.com/book-summaries/constellation-software
• Quarter letter collection: quartr.com/insights/business-philosophy/collection-mark-leonards-shareholder-letters
• FY2024 results (GlobeNewswire, March 2025): globenewswire.com/news-release/2025/03/07/3039269/0/en/Constellation-Software-Inc-Announces-Results-for-the-Fourth-Quarter-and-Year-Ended-December-31-2024
• FY2025 results (CSI official, March 2026): csisoftware.com/constellation-software-inc-announces-results-for-the-fourth-quarter-and-year-ended-december-31-2025
• Q4 2024 Shareholder Report (PDF): csisoftware.com/docs/default-source/press-releases/q4-2024-shareholder-report.pdf
• Q4 2025 Shareholder Report (PDF): csisoftware.com/wp-content/uploads/2026/04/Q4-2025-Shareholder-Report.pdf
• Leonard resignation — official press release (GlobeNewswire, September 25 2025): globenewswire.com — search: 'Mark Leonard Constellation Software President'
• Globe and Mail — Leonard profile (2014): theglobeandmail.com — search: 'Mark Leonard Constellation Software'
  Speedwell Research — Constellation deep-dive memo: speedwellresearch.com
• Colin Keeley — Mark Leonard Operating Manual: colinkeeley.com/blog/mark-leonard-constellation-software-operating-manual
• SBO Financial — Constellation teardown (hurdle rate tiers, acquisition criteria): sbofinancial.com
• The Pursuit of Compounding — FY2025 update (reinvestment rates, acquisition data): thepursuitofcompounding.substack.com/p/constellation-software-inc-fiscal
• Rational Thinking — letter takeaways: rationalthinking.net/constellation-software-takeaways-from-mark-leonards-shareholder-letters
  
  

Operational Velocity is for education and general information only and is not investment, financial, legal, or tax advice, and nothing in it is a recommendation to buy or sell any security. The views expressed are the host's own, the company and figures discussed are drawn from public sources believed reliable but not guaranteed, and you should do your own research and consult a qualified professional before making any decision.

Transcript

SPEAKER_00 0:10

Great returns aren't luck. They're built on the floor. Operators and investors always pushing for more from the deal sheets signed to the systems in place. Speed without precision is just running a race. Operational velocity where execution sets the pace.

SPEAKER_01 0:25

Hey, welcome to Operational Velocity. Here's a business model that should not work. A company acquires another company almost every business day of the year. It has bought more than a thousand of them. And in 30 years it has sold essentially one. The man who ran it calls that single sale his biggest regret. It also refuses to integrate the things it buys. No synergies, no let's consolidate the back office. The company who sell keeps its name, its team, its office, its ways of doing things. It has a head office, about 20 people, running an operation that generates over $10 billion in revenue. And it stopped holding earnings calls. It gives no guidance. And its founder gave, as far as anyone can tell, zero press interviews in almost three decades, and is so rarely photographed that finding a clear picture of them is genuinely hard. Every one of these choices violates something that they teach you in business school or in corporate life. And yet, since it went public in 2006, this company has generated roughly 35% a year for almost 20 years, and the stock is up more than 260 times in 30 years. People call the founder Canada's Warren Buffett, which actually undersell him because Buffett didn't compound at 35%. So here's the question running underneath this episode. And almost nobody in the software or corporate industry has really answered it. How? How do you build one of the greatest track records in the history of public markets by breaking nearly every rule of how you're supposed to acquire and run businesses? And then there's a harder version of that question, one that only became askable a few months ago. Because in October 2025, the architect of the whole thing, Mark Leonard, stepped down. And this spring he told the board he won't stand for re-election. So the most careful builder of a compounding machine in modern business just walked out of the building. The stock, of course, dropped after the announcement, and Harvard Business School wrote a case study asking whether it's time to sell. So this is what we're actually doing today. We're going to reverse engineer the machine, how it really works at the level of the mechanism, not the myth. And then we're going to ask the only question that matters now. Was this a system or was it a man? Because your answer decides whether Constellation Software is the greatest compounder of the next 30 years or a cathedral that has just lost its architect. I'm Gotham Basu, this is Operational Velocity. Let's get into it. To understand why Constellation Software is strange, you have to understand what Mark Leonard walked away from. In the early 1990s, Leonard was a venture capitalist, 11 years in the business. And in venture capital, then as in now, it's a game of swinging for the fences. You fund 20 companies, expecting most to die. Betting one becomes big enough to pay for all of the failures. It hits business, it runs on exits, buys in, grow it, sell, or IPO, return the money, and do it again. Mark Leonard looked at that game and quit it, not because he failed, but because he saw something the industry was not trained to see. While everyone hunted the X big platform, the next huge market, Mark got interested in the companies nobody wanted. And these were small software firms serving one narrow industry. Think about software running a public transit schedule or managing a marina or a funeral home or a library. Boring, tiny, unglamorous businesses in markets so small, no venture capitalist would waste a Tuesday on them. But look at what these boring little companies had. They were usually the number one or number two player in their niche. A niche so small, no giant would ever bother to compete for it. And their customers were so deeply locked in and stuck when your whole operation runs on one piece of industry software, it's very hard to casually just rip it out. And this meant that recurring revenue, year after year, higher margins, and customers who almost never left. In venture terms, these were nothing. In cash flow terms, they were quietly magnificent. So in 1995, Leonard raised about 25 million Canadian dollars from a pension fund and some former VC associates and started Constellation Software. Even the name is a tell, not a Cal Ever brand, just a description of what the thing would literally be. Many small, independent points of light held together under a single roof. And the very first thing he bought that same year tells you everything: a company called Trapeze, scheduling software, actually for public transit buzzes. Not exactly the stuff of a Silicon Valley keynote, and that was the whole point. Here's a choice underneath the choice, the one that shaped the next 30 years. Mark Leonard didn't set out to build a company he would sell. He set out to build a permanent home, a place these little software businesses could be acquired and never flipped. That single decision, buy to hold, forever, never sell, is the seed of everything strange and everything successful that follows. Because when you decide up front that you will never sell, you start making completely different choices from everyone else in the acquisition game. You stop caring about dressing a company up for the next buyer, and you start caring about exactly one thing. Will this business still be quietly generating cash in the next 20 years? And there's a telling detail in how the company went public 11 years later in 2006. Mark didn't take Constellation public to raise a war chest or hand himself a payday. The early venture packers needed liquidity, and eventually that's how their funds work. So going public was a way to let them out. And the larger patient bunny, a pension fund that never needed an exit in the first place, could have stayed private forever. And even the IPO was in service of the buy and hold philosophy, not an exception to it. So now most versions of this story cut straight to the happy ending. And then they compounded at 35%, the end. But that skips the part that should have killed them. Because a strategy that works beautifully when you own 10 companies runs straight into a wall that has destroyed nearly every rollup in history the moment you try to own 500. Let me describe that wall. Because if you've ever worked in corporate development or private equity, you already know it. And you may have even seen it happen. The roll-up is one of the oldest ideas in business. Buy a bunch of small companies in a fragmented industry, bolt them together, and combine things worth more than the pieces. On a spreadsheet, it's beautiful. In reality, the roll-up graveyard is enormous. And here's why they die. As you buy more and more companies, the center, the headquarters, has to make more and more decisions. It isn't close enough to make them well. Which of these 200 businesses gets the next dollar? Is it the transit software company in a market that's growing or dying? Should we buy this marina software competitor at this price? Multiply that across hundreds of businesses in dozens of industries, and headquarters becomes a bottleneck. Decisions slow, and the people who actually understand each little business are three layers away from anyone with authority. And capital gets allocated by people reading summaries of summaries, and the machine seizes. So here's a question that should have broken constellation. How do you allocate capital wisely across hundreds of tiny businesses in more than a hundred different markets? And without a bloated all-knowing head office, that becomes the very bottleneck that kills you. And this is where you see the real genius. And it is not mainly the acquisitions, it's the answer to that question. Leonard's answer was to refuse to build the center at all. Instead of pulling authority up to the headquarters, he pushed it down. Further than almost any company its size. Remember, over 10 billion in revenue and a head office of only 20 people. That number isn't a cost-saving measure. It's a philosophy made visible. It is physically impossible for 20 people to micromanage a thousand companies, and the smallness of the center forces the decision out to the edges. Constellation is built as a nested structure. At the bottom is the business unit, each company they buy, keeping its name and its team. The business units group into operating groups with names most people have never heard of. Voleris, Harris, Ionis, Vella, Perseus. And above that, almost nothing. The operating groups run themselves, each one, in a phrase people inside the company actually use. It's roughly the size of Constellation itself, was about 10 years ago. And they're not divisions, they're younger versions of the whole animal nested inside it. But decentralization alone isn't the mechanism because you can push decisions down and still have people make bad ones. You need the edges to allocate capital as well as a brilliant central investor would. So Leonard installed two things that make the whole thing hold. The first is a number, a hurdle rate. Every potential investment, every acquisition, every project has to clear a required rate of return before anyone's allowed to do it. And these hurdles are demanding on the order of 25, 35% for smaller deals, and set by Leonard himself, personally like scripture. This is the invisible fence. And the discipline isn't a person watching over them, it's a number baked into the culture. And here's what actually buys you. Picture a manager running one of these tiny software units. They find a competitor down the road they can acquire, let's say a $2 million deal. At a normal company, that decision climbs the chain: a memo, a committee, a VP, corporate development, weeks of waiting, and probably a no. At Constellation, if that manager has earned the track record and the deal clears the hurdle rate, they can just do it. Green light it themselves. The company has deliberately pushed the authority to buy other companies down to people running the small units because those are the people who actually understand whether that little transit software business is worth owning. And then there's a second quieter piece of discipline underneath the hurdle rate that's worth naming. Because it's what keeps hundreds of independent deal makers from drifting into wishful thinking. Constellation keeps what amounts to an internal library of base rates, benchmarks for how much revenue growth and margin a deal like this one in this range and this industry has actually delivered across their 500 plus past acquisitions. So when a manager pitches a deal, they can just assert an optimistic growth number, the empirical record of everything Constellation has already bought argues back. And it's that exact discipline that's enforced by data, not by a boss in a room. And that's how you buy a thousand companies, not with a giant MA team at headquarters, but with hundreds of trained allocators at the edge, each governed by the same unforgiving number and the same unforgiving history. And the second thing Mark Leonard installed is the one I'd argue is the real secret. And it's the one the spreadsheet copiers always miss. But see to see it clearly, you have to understand something most coverage of Constellation software skips entirely. How a person actually becomes one of these trusted allocators in the first place. And there's a whole ladder for that. Constellation has its own private vocabulary for how someone rises inside the company. And once you understand it, the whole ownership culture stops being an abstraction and becomes a career path, a specific, visible ladder that hundreds of people are climbing right now. It starts with what they call a player coach. That's someone running one business unit, hands-on, still doing the operating work themselves, still coaching their own small team. Most people who join a constellation company live their whole career at this level, and that's fine. It's a real respected job. But some player coaches do something specific. They find a second business to acquire, and if they're good at it, and if they can find deals, and more importantly, teach the operating discipline to the people actually running that second business, they'll eventually have to give up running their original unit day to day and become something new, a portfolio manager. And a portfolio manager's job isn't running one company anymore. It's finding good acquisitions and training the operators underneath them. And if a portfolio manager keeps doing that, keeps finding good deals, keeps redeploying the cash, those businesses throw off into more acquisitions indefinitely, the company has a final word for what they've become. They call it simply a compounder. Sit with that word for a second, because it's not just a job title, it's a whole philosophy compressed into one noun. Consolation doesn't just call it senior people executives or directors or operating partners. It calls them compounders. People whose entire function is to take cash and turn it into more cash-generating businesses. The org chart is the philosophy. Now, watch how this latter connects to the ownership mechanism, because this is the part that actually explains why the discipline holds at scale. Once someone crosses a certain compensation threshold at constellation, which happens right around when they become a portfolio manager or a compounder, a meaningful chunk of their bonus stops being cash, they can spend. They are structurally not employees pretending to think like owners. They are mechanically owners. And the scale of this, what makes it culture and not a perk. And by the middle of the last decade, Constellation Software reportedly had more than 100 employees who had become millionaires purely through this forced ownership plan. With Mark Leonard saying openly he wanted that number in the hundreds more within another decade. So that's not one star performer getting rich. That's a deliberately engineered population of owner operators growing each year, each one trained on the same ladder, each one holding the same kind of skin in the same game. So here's the thesis starting to sharpen already. Before we even get to the argument proper, the ladder is the mechanism. Decentralization tells you where the decisions get made. The hurdle rate and the base rate library tell you what bar those decisions have to clear. But the compound ladder is what tells you who ends up making them. And it's specifically built so that by the time someone has the authority to green light a multimillion dollar acquisitional loan, they've already spent years being trained, tested, and financially bound to the outcome. Nobody gets handed the keys. They earn them rung by rung with real money on the line at every step. I think there's one more piece of machinery that would be good to understand. And it's kind of a strange one because it isn't a policy or an org chart. It's a stack of documents. And for years, if you wanted to understand how Mark Leonard actually thought, there was exactly one way in. And that's his shareholder letters. In the early years, he wrote one every quarter. And if you think about the discipline that implies, a founder running a company with hundreds of small moving parts, sitting down four times a year and forcing his own thinking into writing in enough detail that outside investors could follow the logic, that's quite interesting. And around 2010, the cadence changed to once a year, by which point he had his letters become something closer to a doctrine than an update. And people who follow uh Constellation closely talk about these letters the way value investors talk about Warren Buffett's. Not as investor relations copy, but as a genuine record of how a serious operator thinks in public, year after year, including the years he got things wrong. And here's why the letters matter to this episode, not just to constellation shareholders. Remember the problem that a 20-person head office cannot personally train a thousand managers scattered across 100 verticals. So how does the judgment, not just the hurdle rate, the actual texture of how Mark Leonard thinks about a deal, a market, a mistake, get transmitted to someone running a small transit software business who has never met him? The letters are the answer. They are the culture transmission mechanism. A hurdle rate tells a manager what number to clear. It can't teach them how Leonard reasons about a deal that's borderline or a market that's slowing. Or an acquisition that's quietly failed. The letters did that instead, published, read, and reread by exactly the population of portfolio managers and compounders who needed to internalize it, without Mark Leonard ever having to sit in the room. And one detail from inside the letters captures the whole spirit. Leonard has written about the discipline of using words, something like the Mies method. Breaking a decision into pieces that don't overlap and don't leave gaps applied to every acquisition and every internal investment, compared against similarly rigorous hurdle rate math across the board. And that's not folksy wisdom. That's an analytical habit written down often enough and clearly enough that it became how everyone at Constellation is expected to think, whether or not they ever exchange a word with the man himself. And this gives you a fuller answer to the question hanging over the compound ladder and who gets trusted with authority. Forced ownership tells you why they'll use it carefully. And the letters are how the thinking itself, the specific texture of Mark Leonard's judgment, got copied into hundreds of heads before he ever needed to be in the room to enforce it. That's three separate mechanisms, each solving a different piece of the same problem. How do you scale judgment and not just capital? Plenty of people have looked at Constellation and tried to copy it. At least a handful of serious firms are running some version of the playbook right now. And the mechanics aren't even secret. Leonard published them for years in shareholder letters passed around by investors the way an earlier generation passed around Warren Buffett's. So if the playbook is public, why can't people just reproduce the results? And here's my argument. Everyone thinks Constellation's genius is capital allocation. The hurdle rates, the acquisitions, the decentralized buying, and that's real, but it's the visible half, the copyable half. The half you can put on a slide. The part almost nobody successfully copies is everything we just talked and walked through. A ladder that trains and tests people for years before handing them authority, a forced ownership plan that makes hundreds of managers into actual owners, and a decade of published letters that copied the founder's judgment into their heads one year at a time. And if you put all those things together, that's not a policy, it's actually a culture. And cultures are not something you install by Monday's all hands meetings. Think about what actually would it take to copy constellation. You could announce the hurdle rate tomorrow, you could publish in an annual letter next January, you could even mandate that bonuses convert to stock. But you cannot compress 30 years of a compensation plan quietly minting owner operators, or a decade of letters training judgment into hundreds of managers who now seen the discipline payoff across multiple market cycles. The tools are copiable in an afternoon. The trust, the track record, and the deeply held belief that this is simply how things are done here, that only comes from having actually lived it for years and watched it work. That's the moat. Not a formula, a compounding culture, and cultures compound on the same clock as the balance sheet does. Slowly, then suddenly, over decades, nobody can rush. Which brings us to the part of the story that until recently was only theoretical. Here's something that's important that most of the retellings of the story skip. The model didn't start straining after Leonard Lyft. It starts straining years before, while he was still very much in the building. And how Constellation handled that strain is the clearest evidence we have for what happens next. Remember the hurdle rate? That demanding 25-30% bar for smaller deals? It's a wonderful discipline when you're buying $2 million transit software companies. But it becomes brutally hard to clear once the only Dells left on the table are worth hundreds of millions because bigger, more visible companies get bit up by more competitors, and high prices are exactly what make a hurdle rate impossible to hit. So Constellation faced a real fork in the road. Lower the hurdle rate to compete for bigger deals and quietly abandon the discipline that built the whole track record, or stay disciplined and get shut out of the larger deals the business would eventually need as the easy, cheap, small acquisitions dried up. And Leonard's answer is the single best piece of evidence in this whole episode for system over man, because it's a structural fix, not a personal one. In 2021, Constellation spun out an entire operating group as its own separately listed company. Topicus. In 2023, it did it again with Lumene. And here's the mechanism, and it's genuinely elegant. These spin-outs can pay for large acquisitions, partly using their own stock as currency instead of straight cash. And that does two things at once. It lowers the cash price they actually have to hand up over up front. And because their stock and target valuation don't trade at the same multiple, the math can work out favorably almost instantly the moment the deal closes. So the same hurdle rate discipline applied to a bigger arena, solved by financial structure instead of by lowering the bar. And here's what looked like in the real world, and it's genuinely striking number. A mortgage technology company called Optimal Blue was sold in 2022 for $2.9 billion. One year later, Constellation through Lumen acquired it for $200 million up front. Not because the business had collapsed, but because the structure of the deal, using Lumen stock as part of the currency, let Constellation family of companies capture that business at a fraction of the headline price anyone else would have to pay for it in cash. That's the hurdle rate defended at a scale nobody thought was possible for this model. And notice what Constellation did not do with these spin-outs, which is itself an own small case study in discipline. When Topicus came into the family, it had built its business on strong organic growth, building software, not just buying it. And the temptation would have been to run the standard constellation playbook over the top of it. Optimize, acquire, extract. Leadership chose instead to protect what made Topicus different and let it keep growing the way it already knew how. That's company confident enough in its own principles to know when not to apply them uniformly. So here's why this matters enormously for the question we're about to ask. This wasn't Mark Leonard personally negotiating every big deal by hand as he aged. This was Leonard, years before his departure, building a structural answer to a problem the business was always going to hit. One that other people are now running. The spin-outs had their own CEOs, their own boards, their own version of discipline adapted to their own markets. The model actually already survived one real test of whether it could bend without breaking. And that happened with Leonard stepping back on purpose before he formally left. Let's shift gears for a moment. For 30 years, a quiet assumption sat in every bull case for constellation. That Mark Leonard was up there somewhere, setting the hurdle rates, guarding the discipline, writing the occasional letter to remind everyone what the whole thing was for. But in October 2025, that assumption expired when he stepped down as president, and later the spring, he told the board he's done there too. And the market flinched, the stock fell double digits on the news, and traded well off its highs. And Harvard turned the moment into a case study whose central question is, bluntly, is this the top? Do you sell? And here's the honest steel man of the bear case, because it isn't stupid. If I've just spent this whole episode telling you the real magic was a cultural one, man spent 30 years instilling the ladder, the ownership, the letters, then that same argument is the reason to worry now. Cultures drift. Founders turn out to be load-bearing in ways that don't show up until they're gone. And it gets harder from here for structural reasons that have nothing to do with Leonard personally. The easy, cheap, small deals are a finite resource. Competition for larger ones keeps rising, and even the spin-out workaround only stretches the model. It doesn't repeal the arithmetic of a shrinking pool of cheap opportunities. That's the bear case. Take it seriously. But here's a bull case, and it's really just the thesis of this episode turned around with real evidence behind it now instead of just the architecture. If Constellation were a man, a genius at the center making all the brilliant calls, and then yes, losing him is catastrophic. But everything we just walked through was Mark Leonard spending 30 years building a machine, specifically designed not to depend on him. The compounder ladder trains his replacements continuously at every level all the time. The forced ownership plan means the people running the place now have the same incentives he always had. And the letters already transmitted a decade of his judgment into writing. And the spin-outs prove the model can be adapted structurally, not personally, to a problem that Leonard saw coming and solved it before he left. The company today is being run day-to-day by the people that the latter actually produce. Leaders like Mark Millard at Valuris and Jeff Bender at Harris, career compounders who came up through exactly the system we just spent the last minutes describing. Not outside hires bought in to hold the fort. If Leonard did the job right, his departure should be the ultimate proof of the system, not the end of it. So which is it, man or machine? And the honest answer, the one I'd stake my credibility on, is we're about to find out. And that's exactly what makes this the most interesting company in the world to watch right now. Because the next few years are the live stress test of the single most important question in operations. Can a great operator build a system so good it outlives him? And Mark Leonard bet his entire career that the answer is yes. And he spent 30 years building four separate, mutually reinforcing answers to what happens when I'm not in the room, who gets trusted, why they'll be careful with it, and how they learn to think, and what happens when the old playbook stops fitting the deal in front of them. The verdict is being written right now by hundreds of people who train, but no longer overseas. All right, so let's start to bring this home. So, what can we actually learn from Constellation and Mark Leonard? I think there's four main things that we can take out of this story. The first one is deciding on your time horizon first and let everything follow from it. Because Constellation's entire strategy is downstream of one upstream decision. We never sell. And that single commitment is what makes the hurdle rates coherent, what makes founders trust them, and what makes the culture possible. Most operators never make that decision explicitly, so half their bad call comes from an unstated, inconsistent horizon. Name yours, whatever it is. Number two, install the fence before you remove the fence. People hear decentralization and they think it means letting go. It's actually the opposite. Mark Leonard could push authority to the edges only because the hurdle rate and the base rate library constrained what the edges could do. Autonomy without a hard shared standard isn't decentralization. It's chaos with a nicer name. Give managers a non-negotiable number to clear first. The freedom and the discipline gets installed together or not at all. The third main lesson is build a ladder before you need the successor. And this is the single biggest thing Constellation did that most companies don't. And that's to build a visible years-long path. Player coach, portfolio manager, compounder. That trains and financially binds people to the discipline long before they're handed real authority. If your bench only gets billed when someone announces they're leaving, you're already behind where Constellation was when Mark Leonard actually walked out the door. And the fourth lesson is write down how you think, not just what you've decided. A hurdle rate is a rule, a letter is reasoning. And reasoning is what actually transfers judgment to someone who'll never sit across the table from you. Most operators keep their best thinking in their own head or in a slide deck nobody rereads. Mark Leonard put his in writing every year in enough detail that a manager three levels down and in one continent away could learn to think the way he does. That habit, done consistently for a decade, is worth more than almost any single decision it produced. So one of the greatest compounding machines of the modern era was built by a man who spent 30 years trying to make himself unnecessary. And the exciting thing is we're about to learn whether he actually succeeded in that. So I'm Gotham Bassoo. This has been Operational Velocity. If you like this series, then please subscribe and follow the show so it lands in your feed the day it drops. And I'll see you next time. Take care.