On the desk today · Constellation Software
He has bought more than 850 companies. He has sold exactly zero.
| TSX · CSU |
$1 quadrillion would be enough to send a check for $2.8 million to every man, woman, and child in America.
That's how big this opportunity is.
And you could claim a stake today…
Before the company goes public…
Starting with just $500.
850 Companies
Zero Sold
Before Mark Leonard built one of the most successful acquisition machines in corporate history, he dug graves. He worked as a bouncer. He moved furniture. He cut stone. After college, he spent 11 years in venture capital — and came away disillusioned. The VC model demanded exits. You built a company, grew it fast, and flipped it. Leonard hated the flipping. In 1995, he raised 25 million Canadian dollars and started Constellation Software in Toronto with a single rule that has never changed: buy small software companies. Never sell them.
I ran into one of these companies last year without knowing it. I was at a municipal office renewing a permit. The clerk was using a clunky-looking program — something from the early 2000s, by the look of it. I asked what it was. She shrugged. "It's what we've always used." That program — the kind of thing you'd never hear about at a tech conference — is exactly what Constellation owns. Hundreds and hundreds of them.
Constellation Software is not a tech company. Not in the way most people mean. It doesn't build products. It doesn't chase trends. It acquires small businesses that sell mission-critical software to tiny, specific industries — transit scheduling, cemetery management, public library systems, spa booking, auto dealership accounting — and it runs them forever.
Leonard launched with C$25 million in 1995 and started buying. The targets were always the same: vertical market software companies — small firms making specialized programs for one narrow industry. Each had high retention, steady cash flow, and a market too small for big competitors to bother with. He bought them, left management in place, and moved on to the next one. By the IPO in 2006, he had dozens. Today, Constellation owns more than 850 businesses across 100 verticals and employs 64,000 people.
In 2025, Constellation reported revenue of $11.6 billion — up 15% from the year before. If you had put $10,000 into the stock at the 2006 IPO, it would be worth roughly $1.25 million today. That's 125 times your money. A 34% annual return for 19 years running.
|
That 850-plus number — acquisitions with zero dispositions — is the one I can't stop thinking about. In private equity, you buy to sell. In venture capital, you build to exit. Constellation does neither. It buys to hold. Forever. The software keeps running. The customers keep paying their annual maintenance fees. The cash flow keeps flowing back to Toronto. And Constellation uses that cash to do it again.
The pricing power hides in plain sight. Each piece of software costs its customer less than 1% of their total budget. A cemetery management system might cost a municipality $30,000 a year. A transit scheduling tool might run $80,000. These are rounding errors in a city's budget. When Constellation raises the maintenance fee by 3% or 5%, nobody launches a procurement review over an extra $1,500. The increases happen quietly, every year. Customers pay anyway.
I find the Leonard story worth sitting with. He could have started a normal tech company. He had the pedigree. Instead, he built a holding company for software that nobody talks about — programs that run bowling alleys, manage marinas, schedule school buses, and track cemetery plots. He wrote in his 2013 shareholder letter that his preferred measure of performance was something he called the "combined ratio" — the sum of return on invested capital and organic revenue growth. That year it hit 39%. He borrowed the term from the insurance industry. It was, I think, a quiet nod to Warren Buffett.
The flywheel is simple and self-feeding. Constellation's 850 businesses generate cash. That cash funds new acquisitions. Those acquisitions generate more cash. The cycle has repeated for 30 years without pause. And as the company has grown, it has spun off entire divisions — Topicus.com in 2021, Lumine Group in 2023 — to keep each unit small, hungry, and close to its customers. The machine gets bigger. The parts stay small.
WHY THIS WORKS
Markets too small to contest. No big company will spend $200 million to compete for a $30 million cemetery software market. Constellation already owns it. The size of the niche is the moat.
Mission-critical, budget-invisible. The software runs the customer's core operation — but costs less than 1% of their budget. Too important to rip out. Too cheap to fight over.
Permanent ownership kills the exit clock. No fund manager is forcing a sale in year five. No banker is running a process. The business compounds in place, undisturbed, indefinitely.
Cash recycles at speed. Every acquired company generates cash that funds the next deal. In 2022 alone, Constellation closed 134 acquisitions — roughly one every three days. The flywheel never stops turning.
What most people miss: Mark Leonard dug graves, bounced drunks, and cut stone before building one of the greatest compounding records in stock market history. He took $25 million and turned it into an $11 billion-a-year revenue machine — by buying the dullest software on the planet and refusing to sell any of it. The secret was never the software. It was the rule: hold everything.
