Walk down any street and look closely at the small businesses around you — a salon booking appointments, a tailor logging alterations, a restaurant processing orders, a repair shop managing parts. They feel permanent, almost invisible, yet they're the infrastructure of everyday life.

Now think about what keeps them running — the quiet layer of software that schedules, tracks, and bills behind the scenes. You've probably never heard of the companies that build it. Neither has most of Wall Street.

That's the point.

Mark Leonard built a $90B empire by seeing what nobody else bothered to notice. Across small towns and city blocks were thousands of unglamorous software companies serving narrow markets — parking-lot management, school-bus routing, marina scheduling — each quietly indispensable, each wildly profitable, and each ignored. They were too small for venture capital, too dull for private equity, and too steady for the press. Leonard saw them for what they were: the perfect business.

From $25M in 1995, Constellation Software has compounded at roughly 33% annually since its 2006 IPO — a total return north of 11,000%.

The secret wasn't product breakthroughs or blitzscaling. It was patience and structure: buy mission-critical vertical software businesses, give them autonomy, never sell, and reinvest the cash into more of the same. That simple idea became a self-perpetuating machine — 600 acquisitions across 75 industries, all compounding quietly behind the curtain of the modern economy.

Leonard himself remains an enigma. He's six-foot-five, speaks softly, and refuses almost all publicity. His annual shareholder letters, dense, philosophical, and handwritten, read like essays from a stoic monk of capital. Among investors, his name sits in the same breath as Buffett and Munger, though he'd recoil at the comparison.

It began, as many great ideas do, in disillusionment.

In the early 1990s, Leonard was a partner at Ventures West, then Canada's leading venture firm. The job looked glamorous from the outside — term sheets, conferences, technology bets — but to Leonard it was a grind of optimism unmoored from discipline. Eleven years in, after dozens of investments in photochemicals, biotech, and remote sensing, he had little to show for it. "The returns were mediocre," he later wrote, "and most of the companies didn't make it."

But buried in the portfolio were exceptions — small vertical-market software businesses that no one paid attention to. One wrote billing software for utilities. Another helped transit agencies plan bus routes. A third sold management tools to local governments. They weren't fast-growing, but they were enduring. The economics were stunning: 95% gross margins, high recurring revenue, almost no churn. Their customers, municipalities, hospitals, small manufacturers, couldn't function without them.

Yet no one wanted to fund them. They were "too small to matter," Leonard's peers said.

He saw it differently.

If a market was too small for Microsoft to enter and too sticky for competitors to replace, it was a fortress. A piece of software that cost a customer less than 1% of revenue but controlled its entire operation would never be swapped out. Each company was a miniature monopoly hiding in plain sight.

Leonard's life up to that point had trained him to spot durability. Born around 1956, he grew up restless and self-reliant. As a teenager he built an illegal flamethrower in his backyard. "Doesn't like anyone telling me what to do," he once wrote. He took odd jobs to pay for school — gravedigger, bouncer, furniture mover, windmill researcher — taking seven years to finish a three-year degree. One summer he worked as a mason's helper for an old craftsman named Wally. Wally told him that if you build a wall right, it will last 100 years. That idea — permanence as virtue — became his philosophy.

After an MBA, Leonard joined Barclays Bank in Toronto. It lasted barely a year. "I realized quickly I would be a terrible banker," he said. A mentor advised him to read client files until something caught his curiosity. Venture capital did. He applied to every firm in Canada and was rejected by all but one: Ventures West.

There he learned by observing failure. He watched founders chase growth for growth's sake, layering complexity on top of weak economics. His mentor, Steve Scotchmer, handed him copies of Buffett: The Making of an American Capitalist and Poor Charlie's Almanack. Leonard absorbed the lessons: stay rational, buy quality, think long term.

By 1995, he'd had enough of venture capital. At 39, with four kids and a mortgage, he left to start something of his own — a holding company that would buy small, perfect software businesses and hold them forever. He raised $25M from OMERS and a few former colleagues. The working name was "Software Co."; friends convinced him "Constellation Software" sounded better.

The first acquisition was Trapeze Group, a transit-scheduling system for public bus networks. Leonard already knew it from his VC days. The second was Harris Computer, a billing platform for utilities. Both companies were small, profitable, and run by founders who wanted continuity, not an exit.

Leonard promised them permanence. No flipping in five years. No integration. No forced synergies. He'd buy the business, let them keep running it, and reinvest the cash into more of the same.

Private-equity firms didn't understand it. Why own slow-growth software forever? But that was the genius: by never selling, Constellation removed the pressure of exits and replaced it with the math of compounding.

Each acquisition became a cash-flow engine feeding the next. Over time, the machine built momentum.

Inside Constellation, autonomy is sacred. Each business stands alone — its own brand, office, and culture. There's no shared HR, no central marketing, no integration meetings. Managers are trained as investors, not operators. They decide how to reinvest their profits, price their products, or pursue acquisitions. Headquarters intervenes only to ensure capital is used rationally.

Leonard calls it "a thousand small experiments running at once." The best ideas spread horizontally; the rest quietly die. Success is measured not in growth but in cash-on-cash returns. The culture prizes humility, thrift, and precision. A manager who saves a dollar of expense is celebrated as much as one who adds a dollar of revenue.

It's a model built for compounding.

Over time, the network effects became tangible. Former founders stayed on as mentors or acquirers. Portfolio companies referred customers to one another. Lessons from a parking-management business in Sweden could inform pricing at a fisheries-tracking firm in Maine.

The results speak for themselves. Constellation has never reported a loss. Through the dot-com collapse, the 2008 financial crisis, and COVID-19, it kept buying, integrating, and compounding. In three decades it has produced hundreds of millionaires among its employees, many through required stock ownership. The share price, which debuted at $18, trades above $2,000 today.

Leonard himself still works from the same modest Toronto office, replying to emails personally and keeping overhead fanatically low. He refuses to issue stock splits, shuns media, and caps his own salary at $800,000. His letters to shareholders read more like philosophy than finance, filled with reflections on decision-making, humility, and the fallibility of human judgment.

"Luck," he once wrote, "has played a large part in our success. Our job is to make sure that, when luck arrives, we're still alive to take advantage of it."

Constellation's endurance comes from what it doesn't do. It doesn't chase hot markets. It doesn't overpay. It doesn't integrate. It simply compounds — slowly, predictably, relentlessly.

And in that slowness lies its power. The slower the machine turns, the longer it lasts. The longer it lasts, the faster compounding does its work.

In a world addicted to speed, Leonard built something designed to endure — a decentralized Berkshire for the software age, proof that greatness in business doesn't require invention or disruption, just patience and permanence.

Walk down that same street again — the salon, the tailor, the restaurant, the repair shop. Somewhere inside, quietly keeping the lights on, is a piece of software owned by Constellation.

It may not have changed the world. But it has built one that lasts.