The Wizard in the Hallway
Somewhere in the crowded corridor of a Constellation Software company event — the kind of gathering where business unit managers and portfolio managers mingle with the informality of people who share an operating manual but have never met — a general manager stopped a tall, bespectacled man with an enormous, wizard-like beard to introduce himself. The GM ran one of the company's smaller units, fifteen people, a rounding error inside an empire that by then encompassed hundreds of business units across more than a hundred markets. Mark Leonard — because of course it was Leonard — knew the GM's name. He knew the business model. He knew details about the unit that even some of the GM's direct reports might not have articulated as crisply. Then he kept walking.
This is a story about a man you cannot Google Image. A man who, for nearly three decades, has refused interviews, shunned social media, declined to appear on television, and maintained what the Globe and Mail once called "a practically unthinkable level of anonymity" — all while compounding a $25 million initial investment into a company with a market capitalization exceeding $90 billion. Mark Leonard has no TED talk. No Twitter account. No podcast appearances, save one interview conducted by his own operating group CEO, Jeff Bender, which was quickly deleted from the internet before admirers could archive it. His LinkedIn profile, which for years listed only his job title and his years at Ivey Business School, carried no photograph. Even his biographical details are contested: one source says he was born in England in 1956; another thinks South Africa; a third can't quite remember, which is itself a remarkable fact about the founder of Canada's highest-revenue software company.
And yet the anonymity is not incidental. It is architectural. It is as much a part of Constellation Software's design as the hurdle rates, the post-acquisition reviews, the fractal organizational structure that replicates itself downward into ever-smaller units of autonomous decision-making. Leonard's invisibility is not shyness. It is philosophy — the belief that an institution should outlast and outperform any individual's ego, that the system is the product, and that the system works best when the person who designed it steps back and lets the machine run.
By the Numbers
Constellation Software
$90B+Market capitalization (2025)
1,000+Companies acquired since 1995
$25MInitial investment from OMERS and Ventures West
~30%Annualized return since 2006 IPO
100+Industries served via vertical market software
75,000+Employees across all business units
<30Employees at corporate headquarters
The Education of a Gravedigger
The biography of Mark Leonard, to the extent one can be assembled, reads like a curriculum vitae composed by a man who suspects that curriculum vitae are absurd. A bio prepared for a rare speaking engagement at his alma mater, the Ivey Business School at Western University, stated that before entering venture capital, "he worked as a banker, valuator, mason, gravedigger, dog handler, bouncer, sapper, and wind energy researcher. He particularly enjoyed the bouncing, but early retirement was necessary." Whether all of these are true or merely the provocations of a man who dislikes biographical convention is itself unknowable — journalists have tried to verify the list and emerged no more certain than when they began.
What is verifiable: Leonard holds a Bachelor of Science from the University of Guelph and an MBA from Western Ontario's Ivey Business School, where he attended from 1980 to 1982. He spent seven years in college and played multiple sports, including basketball and rugby. He is described by those who know him as a devoted family man — and also, in the same breath, as someone whose biggest regret is not spending enough time with his family, a contradiction he has acknowledged publicly. Speaking at an investment conference in Toronto in December 2022, Leonard told the audience that he had studied successful business builders and found that "almost none of them had any real work-life balance." He invoked the Dunbar number — the cognitive limit of approximately fifty meaningful relationships a human can sustain — and observed, quietly, that if most of those relationships belong to colleagues, the arithmetic of family life becomes unforgiving.
After Ivey, Leonard spent eleven years in venture capital at Ventures West Capital, a Canadian firm where, by his own account, returns were mediocre. The Canadian VC industry in the 1980s and early 1990s was a backwater, and Leonard — who would later describe his temperament as someone who cannot abide being told what to do — watched the standard playbook repeat itself: raise fast, invest fast, exit fast, destroy value more often than create it. Most startups, he came to believe, were net-negative propositions. The cult of hypergrowth was toxic. Short time horizons bred undisciplined valuations.
But something nagged at him. Among Ventures West's portfolio, the best-performing investments — the ones that reliably generated cash, retained customers, and didn't blow up — tended to be small, niche software companies. Vertical market software firms. Companies that made products to run golf course tee-time reservations, or manage public housing payroll, or track hospital laundry logistics. The problem was that these companies were individually so small they barely moved the needle on a venture fund's overall returns. They were, in the language of venture capital, uninteresting. Markets too small for VCs to care about. Leonard cared.
Software Co.
In 1995, at the age of thirty-nine — late by Silicon Valley standards, though Leonard has never had any relationship with Silicon Valley — he left venture capital and pitched a new idea. The pitch deck referred to the company simply as "Software Co." The name was deliberately unmemorable, which is fitting, because the company Leonard proposed to build was designed to be the opposite of what the technology industry valorized. No platform plays. No billion-dollar TAMs. No growth-at-all-costs. Instead: an umbrella organization that would acquire, finance, and nurture — his word, nurture — small vertical market software businesses. Buy them cheaply, hold them forever, and use their cash flows to buy more.
The initial capital came from OMERS, the Ontario Municipal Employees Retirement System, and Leonard's old colleagues at Ventures West. Twenty-five million Canadian dollars. The terms of the arrangement are not public, but the underlying proposition was starkly simple: Leonard had identified a market inefficiency — the structural undervaluation of tiny, boring, cash-generative software companies — and proposed to exploit it indefinitely. Not through a fund with a ten-year life and mandatory distributions, but through a permanent capital vehicle. A holding company. A perpetual owner.
The insight was not that vertical market software was a good business — that was obvious to anyone who looked. These companies served hyper-specific niches: software for slaughterhouses, for fertility clinics, for cemetery management, for municipal water billing. Their customers had nowhere else to go. Switching costs were enormous because the software was mission-critical, deeply embedded in workflows, and written for markets too small to attract competition. Customer retention rates ran above ninety percent. Revenue was overwhelmingly recurring. And because the software typically represented less than one percent of a customer's total budget, pricing power was formidable — customers barely noticed the line item and would never risk operational disruption to save a few thousand dollars.
No, the insight was about what happened to the cash. These companies generated prodigious free cash flows relative to their size, but the founders had nowhere to reinvest. A VMS business with a million dollars in annual profit couldn't pour that money back into R&D for a market that was, by definition, small. It couldn't expand into adjacent verticals without becoming something it wasn't. So the cash sat in a bank account earning two percent, or it funded the founder's retirement, or it got distributed as dividends to a handful of owners. The business was, functionally, a bond — a very high-yield bond with no reinvestment vehicle attached.
Leonard's contribution was to attach the reinvestment vehicle. Inside Constellation, the cash flows from hundreds of these tiny businesses would be funneled into a machine that had demonstrated the ability to reinvest capital at returns exceeding twenty-five percent. The individual companies were static. The system compounded.
The Fractal
Constellation Software's organizational chart looks less like a corporation than like a biological organism, or perhaps a coastline measured at increasing resolution — the closer you look, the more detail emerges, and the shape at every level of magnification resembles the shape at every other level. This is not metaphor. Leonard designed the company as a fractal.
At the top sits Constellation Software itself, the publicly traded parent, with fewer than thirty employees at headquarters in Toronto. Beneath it are six operating groups — Harris Computer Systems, Volaris Group, Jonas Software, Perseus Operating Group, Vela Software, and Topicus.com — each of which functions as a semi-autonomous conglomerate in its own right. Each operating group contains portfolio managers who oversee clusters of business units. Each business unit is a standalone VMS company, typically employing somewhere between ten and two hundred people, with its own brand, its own customers, its own general manager, and its own profit-and-loss accountability. The name on the door is never Constellation. A former Harris employee interviewed about his experience said that in a two-hour conversation, "not once did he mention Constellation Software. He kept mentioning Harris, Harris, Harris."
We believe that small businesses should stay small. While other purchasers of business seek economies of scale through centralization, we have taken the opposite approach. We believe that potential is maximized when teams are allowed to remain small.
This is the anti-synergy thesis. Where conventional acquirers centralize operations to extract economies of scale, Leonard did the opposite. He called Constellation "the anti-economies of scale company." When a business unit grew too large, he broke it into smaller pieces — not to punish growth, but to preserve the conditions that produced it. Small teams, close to customers, empowered to make decisions without waiting for approval from Toronto. The general manager of a business unit has full authority over operations. Below a certain purchase price threshold — reportedly around $20 million — business unit leaders can even make their own acquisitions without corporate sign-off.
The logic is Hayekian, though Leonard would probably flinch at the ideological framing. Distributed knowledge, distributed decision-making. The person running a VMS company for municipal water billing in a small German town knows more about that market than anyone at headquarters ever could. The role of headquarters is not to direct but to provide capital, share best practices, and measure results. The role of the operating group is to serve as a intermediate layer — coaching, filtering acquisition proposals, conducting post-acquisition reviews. And the role of each business unit is to run itself, profitably, and send the excess cash upstream for redeployment.
It is worth pausing on how unusual this is. Most technology conglomerates — even those that claim to value autonomy — cannot resist the gravitational pull of integration. They consolidate sales teams, merge engineering departments, standardize technology stacks. The short-term cost savings are real and visible. The long-term damage — to morale, to customer relationships, to the institutional knowledge embedded in a small team that has served the same niche for decades — is real but invisible, at least until it shows up in churn rates three years later. Leonard's insight was that the invisible damage was more expensive than the visible savings. So he left the companies alone.
The Magnetic Property of Hurdle Rates
Among the sentences Mark Leonard has committed to paper in his shareholder letters — annual documents that, until he stopped writing them in 2017, were regarded by a growing cult of investors as the second-best shareholder letters in existence, behind only
Warren Buffett's — one stands out for its implications:
Hurdle rates are magnetic in nature.
The statement is deceptively compact. Here is what it means: Constellation sets internal rate of return (IRR) hurdle rates that every proposed acquisition must clear before it can proceed. Over the company's history, they raised the hurdle rate once and lowered it twice. What they discovered — empirically, through hundreds of deals — was that when they lowered the hurdle rate, the IRR of every deal they pursued declined to match it. Not just the marginal deals. Not just the ones teetering on the borderline. Everything. The entire distribution of returns collapsed toward the new, lower bar.
This was counterintuitive. You would expect that lowering the hurdle from, say, twenty-five percent to twenty percent would only affect the marginal deals — the ones that cleared twenty but not twenty-five. The high-conviction, high-return deals should be unaffected. But that's not what happened. The gravitational force of the lower bar pulled the entire field downward. People unconsciously relaxed their standards. Assumptions got rosier. Projections got more generous. The discipline that produced a twenty-five percent return was subtly, pervasively eroded by the signal that twenty percent was now acceptable.
Leonard described the temptation from the inside in his 2017 letter: "I recently worked on a large transaction. With every day that passed, I could feel my commitment to the process growing… not because the news was getting better, just because I was spending more time on the prospect." The investment didn't meet the hurdle rate. They walked away. But the experience reinforced something Leonard had been teaching his organization for two decades: the human susceptibility to commitment bias, to anchoring, to the seductive warmth of the deal that is almost good enough.
To combat this, Constellation built a data infrastructure around its acquisitions. With over a thousand completed deals, the company possesses base rates — empirical distributions of revenue growth, profitability, churn, and other key metrics across its entire acquisition history. When a new investment champion argues that a particular prospect is "special," the data allows the organization to locate that claim on the distribution. If the projected growth rate would place the acquisition at the 95th percentile of all businesses Constellation has ever bought, that is not necessarily a reason to say no — but it is a reason to pause and ask whether the champion is genuinely identifying an outlier or merely rationalizing.
Bernard Anzarouth, Constellation's Chief Investment Officer — a man described in Leonard's letters as "the last line of defense" against over-optimistic acquisition proposals — spent years as the primary filter through which borderline deals were screened. As the organization scaled, some operating groups developed their own senior M&A personnel to supplement Anzarouth's function, but the principle remained the same: institutional skepticism, backed by data, applied at every level of the hierarchy.
The Permanent Portfolio
There is a question that hovers over every serial acquirer: Why don't they sell? The private equity model — buy, optimize, flip — exists because it works, at least for the fund managers. Buy a company for six times EBITDA, lever it up, cut costs, sell it for ten times EBITDA in five years. Collect your carry. Repeat. Leonard's answer to this question is both economic and moral, and the two are, in his framework, inseparable.
The economic argument: selling a good business is selling a compounding asset. If a VMS company generates a twenty percent return on invested capital and that capital can be redeployed at similar rates, selling the company means surrendering the future stream of returns for a one-time payment. The mathematics of compounding punish early exits. Almost all of the benefit, as the Liberty RPF podcast host noted, "comes at the end of the exponential curve." Owning a great business for a year captures almost none of the value. Owning it for twenty years captures nearly all of it.
The moral argument is more subtle. When Constellation buys a company from a founder — and founder-led businesses are, by Leonard's own admission, their "favourite and most frequent acquisitions" — it is acquiring something that a human being spent the better part of a lifetime building. The founder chose Constellation over other buyers not because Constellation paid the highest price (it often didn't), but because Constellation promised permanence. The brand would survive. The team would stay intact. The customers would not be disrupted. In Leonard's own telling, this promise created a sourcing advantage: word spread that Constellation preserved what founders had built, and founders began calling Constellation directly when they were ready to sell.
Leonard has sold exactly one company in Constellation's history. He deeply regretted it.
The operational corollary is that Constellation's managers are not optimizing for exit. They are not grooming businesses for a five-year flip. They are not cutting R&D to inflate short-term margins. They are running businesses with infinite time horizons, which changes every incentive, every trade-off, every decision about how to treat customers and employees. It also means that Constellation's acquisition pipeline is not competing against private equity in most cases. PE firms target larger businesses, apply leverage, and plan to exit. Constellation targets smaller businesses, avoids leverage, and plans to hold forever. The overlap in the Venn diagram is small enough that Constellation often operates in a market with no other serious buyer — a structural advantage that has allowed the company to acquire businesses at multiples hovering around one times revenue for decades, even as the broader software market saw multiples expand five- or sixfold.
The Incentive Machine
If you want to understand Constellation Software, you must understand its compensation structure, because the compensation structure is the strategy. It is the mechanism through which Leonard's philosophy becomes behavior.
There are no stock options at Constellation. No restricted stock units. No equity-based compensation of any kind. In an industry where stock-based compensation has become, in the words of one analyst, "bread and butter for everyone," Constellation's refusal to issue equity is not merely unusual — it is a statement of principle. Leonard views stock-based compensation as dilutive, as a hidden cost that companies systematically underreport, and as an incentive that encourages employees to optimize for short-term stock price movements rather than long-term value creation.
Instead, Constellation requires its managers — above a certain salary threshold, reportedly around C$75,000 — to invest a significant portion of their annual bonuses (between twenty-five and seventy-five percent, depending on seniority) in Constellation shares purchased on the open market. The shares come with multi-year vesting periods. The effect is that every senior manager at Constellation has a substantial percentage of their personal net worth invested in the company's stock, purchased at market prices, with real downside exposure. By 2015, Leonard reported that Constellation had created over one hundred employee millionaires. He predicted five hundred by 2025, a target the company likely hit early.
The downstream effects of this structure are profound. Managers who own meaningful stakes in the company think like owners — not like employees who hold lottery-ticket options and are incentivized to maximize volatility. They are careful with capital because it is, literally, their capital. They are attentive to acquisitions because every deal they champion lands in their capital base permanently. There is no write-off for impairment, no "adjusted" return that excludes the bad deal. If you acquired a business that didn't perform, it sits in your numbers forever, a visible scar on your track record that colleagues and superiors can see during post-acquisition reviews.
There's this inside joke in Harris: "We work at Harris, but we get paid like McDonald's."
This joke, relayed by a former employee, contains an important truth. Base salaries at Constellation, particularly at junior levels, are not competitive with Silicon Valley — or even with other Canadian technology companies. The compensation gap is bridged by the bonus structure and the stock appreciation. When Constellation shares are compounding at thirty percent per year, the forced reinvestment of bonuses generates enormous wealth over time, and nobody complains. But this creates what Leonard himself has acknowledged as a risk: if the stock ever stops performing, the compensation model unravels. Employees who accepted below-market salaries in exchange for stock appreciation would have no reason to stay. The most talented operators — people with the rare combination of operational skill and capital allocation acuity — could sell their shares, take their expertise, and buy a $3 million VMS business on their own.
Leonard's response to this risk has been characteristically indirect: he has tried to keep the stock price tethered to intrinsic value, neither so high that employees are tempted to cash out nor so low that the company becomes vulnerable to an activist or hostile bid. In his 2011 letter, he wrote that he was "coming around to the belief that if our stock price strays too far (either high or low) from intrinsic value, then the business may suffer." He described Constellation's senior managers as "amongst the most talented available" at operating vertical market software businesses, adding — in what may be the only boast in the entire corpus of his shareholder letters — "and I'm uniquely qualified to be a connoisseur of such talent."
The Operating System
What happens after Constellation acquires a company? The answer is: almost nothing dramatic, and that is the drama.
The acquired business keeps its name, its brand, its management team, its customers, and its culture. There is no wave of consultants, no rebranding, no forced migration to a common technology platform. Constellation does not pursue synergies in the conventional sense. It does not merge overlapping sales teams or consolidate engineering departments. It does not even prevent two of its own subsidiaries from competing for the same customers — a fact that has struck more than one observer as either idealistic or insane.
What Constellation does do is inject two things into every acquired business: capital allocation discipline and best-practice sharing. The discipline takes the form of metrics — ROIC, organic growth, free cash flow generation — tracked at the business unit level with granularity that most corporations reserve for their largest divisions. The best-practice sharing takes the form of a learning network: when one business unit discovers an approach that improves customer retention or accelerates professional services revenue, that knowledge percolates through the organization via internal conferences, peer groups, and the coaching relationships between portfolio managers and general managers.
The post-acquisition review, or PAR, is central to this process. Approximately one year after every acquisition, the deal is revisited: were the assumptions correct? Did the business perform as projected? What was learned? PARs originated as a head-office-led process but were eventually delegated to the operating groups — consistent with Leonard's preference for pushing decision-making downward. His standard for any corporate process is characteristically blunt: "An ad-hoc analysis done to understand a problem or opportunity is more likely to translate into action than a quarterly report that gets generated because 'we've always done it that way.' The former requires curiosity and intelligence, the latter bureaucracy and compliance."
Organic growth across the portfolio has historically been modest — two to three percent in aggregate — which has been a frequent point of criticism from investors accustomed to the double-digit growth rates of SaaS companies. But the aggregate figure obscures more than it reveals. The maintenance and recurring revenue lines, where most value resides, typically grow at three to six percent. What pulls the overall number down is the mix effect: Constellation buys shrinking businesses and runoff portfolios deliberately, because they can be acquired cheaply enough to generate exceptional IRRs even if revenue declines. A business that is shrinking by ten percent per year but was purchased at half a turn of revenue can be extraordinarily profitable. The aggregate organic growth rate is, in this light, a feature of capital discipline, not a symptom of operational weakness.
The Topicus Turn
In 2020, Constellation announced the acquisition of Topicus.com, a Dutch education, finance, and healthcare software company, from a private equity fund. The purchase price was not disclosed — typical of Constellation — but the company reported that Topicus had revenues of €101 million and one thousand employees. The real significance lay in what followed: Constellation merged Topicus with its existing European operating group, Total Specific Solutions (TSS), and listed the combined entity on a stock exchange as Topicus.com.
TSS had been acquired by Constellation seven years earlier for €248 million. By 2020, it had grown to eighty business units and three thousand employees, a case study in the organic replication of Constellation's model in a new geography. But Topicus brought something different — a demonstrated ability to spin up new businesses organically from within, generating levels of organic growth that exceeded anything in Constellation's portfolio. Leonard was, by his own account, "super excited" to study how Topicus achieved this. If the techniques could be uploaded to the rest of Constellation's hundreds of business units, even a modest improvement in organic growth rates — from two percent to four percent, say — would translate into billions of dollars of incremental value.
The Topicus spin-out also served a cultural purpose. Leonard wrote in the press release: "The plan to create a publicly listed operating group made up of Topicus and TSS was a key part of our discussions with the Topicus founders. They didn't want their legacy disappearing into the craw of an omnivorous conglomerate." The public listing gave the European operations their own identity, their own stock, their own ability to attract talent and conduct acquisitions with a local currency. Constellation retained control. But the identity of the acquired company was, once again, preserved.
This was followed in subsequent years by the spin-out of Lumine, another operating group given its own public listing. The fractal was replicating. Each group — Volaris, Harris, Jonas, Perseus, Vela, Topicus — was becoming what Constellation itself had been a decade earlier, a semi-autonomous conglomerate with its own capital allocation process, its own acquisition pipeline, its own talent development ladder. A person could join Constellation as an engineer at a tiny business unit, discover an aptitude for M&A, acquire a company or two, ascend to portfolio manager, and eventually operate at the operating group level — presiding over hundreds of millions in deployed capital, functioning as a mini-CEO within the larger constellation. The career path was the retention mechanism, and the retention mechanism was the strategy.
The Evolution of the Appetite
For most of its history, Constellation's acquisitions were small. The typical deal was in the $2 million to $5 million range. The median price hovered around one times revenue. Constellation's M&A professionals — eventually numbering twenty-six operating group and portfolio managers spending more than half their time on acquisitions, plus another sixty full-time M&A specialists — maintained a database of VMS companies that grew from twenty thousand to sixty thousand over a decade, with coverage that they themselves described as "very low." The addressable market of potential targets was, by some estimates, forty thousand companies globally. Constellation had acquired roughly a thousand.
But the math of compounding eventually creates its own problem. As Constellation's free cash flow swelled to $2 billion annually, deploying that capital in $5 million increments required an increasing number of deals — from roughly forty per year in the mid-2010s to more than one hundred by 2023. At some point, the logistics of sourcing, diligencing, and integrating that many tiny deals would become a bottleneck. Leonard acknowledged this in his later letters and public remarks, with what he called "the fervor of the newly converted" — he had changed his mind about large acquisitions.
The first major test came with Constellation's acquisition of a substantial portion of Allscripts, a large healthcare software business that was widely viewed as a shrinking, lower-quality asset. The acquisition was consistent with Constellation's playbook in one sense — buying an unloved business cheaply — but inconsistent in another: it was large enough to carry meaningful integration risk. If it failed, the mistake would not be a $5 million "lesson" but a nine-figure write-down.
Leonard's letter to prospective large-deal sellers, published as part of his shareholder communication, was characteristically direct. Constellation was open for business on transactions of meaningful size. The hurdle rates would not be lowered. The operating model — decentralized, autonomous, permanent — would not change. But the appetite had grown, because the cash flows demanded it.
This evolution worried some longtime investors. The bull case for Constellation had always rested partly on the absence of large-deal risk — the company's returns were built on a portfolio so diversified that no single acquisition could materially damage the whole. If Constellation started making billion-dollar bets, it would be playing a different game, one where private equity firms and strategic acquirers competed fiercely and where the informational advantages of base rates and niche expertise might not apply. Leonard was betting that the organizational capabilities — the evaluation discipline, the operational playbook, the long-term hold philosophy — would transfer to larger deals. Whether he was right would not be knowable for years.
The Buffett Comparison, and Its Limits
The comparison to Warren Buffett is inevitable, repeated so often that it has become almost a tic among Constellation's admirers. The parallels are genuine: both men run permanent capital vehicles that acquire businesses to hold forever. Both obsess over return on invested capital. Both write shareholder letters of unusual candor and intellectual depth. Both have attracted a cult following that treats their annual communications as scripture. Both are allergic to stock-based compensation. Both prefer concentrated investments over diversification. Both understand that capital allocation is the CEO's most important job.
But the comparison obscures differences that are at least as instructive. Buffett's Berkshire Hathaway operates across wildly diverse industries — insurance, railroads, energy, consumer goods — and generates much of its competitive advantage from the float of its insurance operations, a structural source of cheap leverage that Constellation does not possess. Constellation operates within a single sector — software — and relies not on cheap leverage but on the compounding effect of high-ROIC reinvestment within an ever-expanding universe of VMS targets.
Leonard also diverges from Buffett on the question of buybacks, though he has softened his position over time. In his letters, Leonard described share repurchases as "essentially prey to the shareholders" — arguing that because management possesses inside information about the company, buying back shares from less-informed outside holders constituted a form of informational exploitation. This view, extreme by the standards of contemporary corporate finance, is less a financial argument than an ethical one. It reveals something about Leonard's character: a reflexive discomfort with any action that might benefit insiders at the expense of outsiders, even when the action is legal, conventional, and widely regarded as shareholder-friendly.
Leonard also differs from Buffett on the question of operator quality. Buffett famously quipped that he prefers to invest in businesses so good they can be run by an idiot — because eventually they will be. Leonard, whose entire model depends on the skill and motivation of hundreds of autonomous general managers, cannot afford that luxury. The businesses Constellation acquires are not invulnerable monopolies; they are small, niche software companies that require attentive management to maintain customer relationships, invest in product development, and navigate the slow-moving but real technological shifts in their verticals. The system depends on the people. The people are the moat.
A Serious Health Issue
On September 25, 2025, Constellation Software announced that Mark Leonard had resigned as president of the company "due to health reasons." The announcement was terse. Management told the Financial Post that Leonard's health issue was "serious." Mark Miller, the chief operating officer, was promoted to president. Leonard would remain on the board.
Jamal Baksh, Constellation's chief financial officer — a man who had worked with Leonard for decades and who served as one of the few public voices of a company that preferred not to have a public voice — told the Financial Post: "Mark's a brilliant guy. He still will be associated with the company, and if he's still on the board… I don't think we necessarily lose anything because the core day-to-day stuff he was doing can be replaced, and it's his mind and his vision that we still have access to."
This was, in its way, the ultimate test of the system Leonard had spent thirty years building. He had designed Constellation to operate without a charismatic founder at the center. He had pushed capital allocation authority downward, trained hundreds of managers to make acquisition decisions, codified the operating playbook, and built an incentive structure that aligned every participant's interests with the long-term health of the enterprise. He had, in effect, made himself dispensable — or at least attempted to. The market's reaction would tell him whether he had succeeded.
Baksh's distinction — between "the core day-to-day stuff" and "his mind and his vision" — is the one that matters. For the prior decade, Leonard's operational involvement had been strategic rather than transactional. He was no longer approving individual small acquisitions. He was thinking about whether the model could extend beyond VMS, studying high-performance conglomerates like TransDigm and Illinois Tool Works, and considering what Constellation might look like in 2040. Whether that vision could survive without its originator is the open question — the one that Leonard, characteristically, chose not to answer publicly.
The 1895 Thought Experiment
In one of his later shareholder letters, Leonard posed a thought experiment that reveals more about his mind than any biographical detail could. He suggested that if Constellation Software had been founded in 1895 rather than 1995, the company might have been buying newspapers. Not because newspapers in 1895 resembled vertical market software in any technical sense, but because they shared the same structural characteristics: local monopolies, loyal audiences, high switching costs, reliable subscription revenue, and total addressable markets too small to attract the attention of large capital allocators.
The implication was devastating to anyone who thought Constellation's identity was inseparable from vertical market software. Leonard was saying, in his oblique way, that VMS was not the point. VMS was the vehicle. The point was the system — the acquisition discipline, the decentralized operations, the permanent hold philosophy, the base rates, the incentive structure, the fractal organization. Apply that system to any asset class exhibiting the right structural characteristics, and it should work. Leonard was not, as one observer put it, "quite as fascinated with VMS" as many of his investors were. He was fascinated with the machine.
This explains why the most recent evolution of Constellation's strategy — the willingness to look beyond VMS, the openness to larger transactions, the interest in non-software businesses with similar economic profiles — does not represent a departure from Leonard's vision but its logical extension. The system was built to compound capital at high rates of return. If VMS targets become scarce or expensive, the system will seek other inputs. The discipline is portable. The culture is portable. The organizational design is portable.
Whether it will actually port is another matter. Constellation's entire knowledge base — the base rates, the PAR archives, the institutional intuition built over a thousand deals — is specific to vertical market software. A portfolio manager who has spent fifteen years evaluating VMS businesses knows, in her bones, what a normal churn rate looks like, what a reasonable margin trajectory is, which acquisition structures work and which invite disaster. That knowledge does not automatically transfer to industrial equipment or media companies or healthcare services. The system may be general, but the expertise is specific.
Leonard knew this. And he did not pretend to have resolved the tension. He simply observed it, filed it alongside every other paradox his company had accumulated over three decades, and kept moving.
The Constellation
The final image: a network of more than a thousand small software companies, scattered across a hundred markets in dozens of countries, each one a dim point of light in its own obscure niche — cemetery management in Ontario, marina billing in New Zealand, school bus routing in Denmark — individually unremarkable, collectively generating ten billion dollars in revenue and two billion dollars in free cash flow, all flowing through a system designed by a man with a wizard's beard who does not appear in photographs, who has never spoken to a journalist on the record, who rode his bicycle to work, who stopped taking a salary in 2016, who once listed his past occupations as including gravedigger and bouncer, who told an audience of young investors that his biggest regret was not spending enough time with his family, and who, when he walked through that crowded hallway and was stopped by a general manager from a fifteen-person business unit he had never visited, knew the man's name.
