Constellation Software
The acquirer that never sells

Constellation Software wins auctions it does not lead. Its edge is one sentence: we will never sell your company. Founders take that over a bigger check.

The highest bid loses more often than operators expect

Price clears most markets. It does not clear this one, because the seller here is usually a person with a name on the door.

Look at who Constellation buys from. In 2016 the company made 40 acquisitions. Thirty-five had fewer than 100 employees, and 30 came from owner-managers, according to Mark Leonard's April 2017 letter to shareholders. (Constellation Software President's Letter, April 25, 2017) Those sellers spent twenty or thirty years inside the thing they are handing over. They have a second question after price, and it is some version of "What happens to this place after I leave?"

A private equity buyer cannot answer that question honestly. Its model requires an exit in three to seven years, so the honest answer is "someone else will decide." Constellation answers it in one line and then points at three decades of evidence. The answer is worth real money, which means it functions as part of the bid.

The promise only stays cheap if the asset needs no fixing

Constellation buys vertical market software: the systems that run dental practices, marinas, municipal parking, small utilities. Tiny markets. A few thousand possible customers each. Nobody at Microsoft is coming.

Two features make these businesses safe to own forever.

• Customers almost never leave, because the software holds years of records and workflow, so replacing it means retraining staff and rebuilding a database.
• And the annual cost sits low enough against the customer's revenue that nobody cuts it to save money.

This combination is why the never-sell promise is affordable. Leonard put the objective plainly in his April 2018 letter: "Our objective is to be a great perpetual owner of VMS businesses." A perpetual owner needs assets that do not require rescue. Constellation's organic revenue growth averaged 2% over the decade to 2017, by Leonard's own accounting. No buyer could build a resale story on 2%. Constellation never needs one, so it can promise permanence without lying. (Constellation Software President's Letter, April 20 2018)

Toronto, April 20, 2018

Leonard signed his letter that day and told shareholders it was the last annual one. He had been writing them since the 2006 IPO, and investors treated them like they treat Buffett's letters. His reason for stopping was blunt: Constellation had attracted imitators, and teaching them the company's tactics no longer served shareholders.

He kept the archive online. That detail matters more than the decision to stop. A decade of dated, signed letters sits on the website saying the same thing about permanence, written before the company had 1,000 businesses and before anyone would have applauded. A seller in 2026 can read what Leonard promised in 2011 and check whether it held.

The letters also show the machinery underneath. Head office overhead fell from 3.0% of net revenue in 2004 to 0.5% in 2016. The company ran 243 separately managed business units at the end of 2017, up from 193 a year earlier. (President's Letters, April 25 2017 and April 20 2018) Nobody at head office has time to integrate anything, which is precisely what makes the hands-off promise credible to the person signing it.

A promise is an asset only when breaking it would be expensive

Anyone can say they buy and hold. The words cost nothing, which is why sellers ignore them.

What Constellation built is different in ways any buyer can copy. The claim is public and dated, so it can be checked. It is repeated by the operating groups that actually sit across the table, not only by headquarters. Volaris Group, one of the six, tells sellers on its site that "Because we never sell our companies, we are committed to their long-term sustainable growth." (Volaris Group, "Constellation Software named top strategic acquirer for 2018," February 13 2019) And every year the company holds, without divesting, the cost of a future sale, because one exit would reprice every deal after it.

The last part is the mechanism worth stealing. A promise that gets more expensive to break each year turns time into a competitive asset. New entrants cannot buy it, and well-funded competitors cannot outbid it.

What the promise costs the buyer

The bill comes due in a few places.

• Underperformers stay. A fund sells its mistakes; a perpetual owner explains them at the annual meeting and keeps writing the checks. Constellation's answer is to buy small enough that any single business is a rounding error.
• Scale attacks' returns. Leonard wrote in April 2018 that he was worried about the supply of good businesses at reasonable prices, and Constellation now deploys far more capital each year than a diet of million-dollar deals can absorb. Larger deals mean lower returns, which is the price of keeping the promise while the portfolio grows. (President's Letters, April 25 2017 and April 20 2018)
• Succession tests everything. Leonard resigned in September 2025 for health reasons and did not stand for re-election to the board in March 2026. The promise now belongs to the institution, and the next decade will show whether it survived the handover.

Four things to do this week

If you are running a business, the following takeaways can help:

1. Write the promise you could keep for twenty years

Not values, but one testable commitment: we will not resell, rename, or move the team. You will know it is the right promise when you can name the specific future deal it would cost you and you would still keep it.

2. Ask your last five sellers what their second question was

One call each, twenty minutes total per person. The signal is when someone raises continuity, their team, or their name on the door before you prompt them.

3. Publish the promise with a date on it

This promise could be a page on your site, a line in your LOI, or a sentence in the offer letter. Success looks like a stranger being able to check whether you kept it without asking you.

4. Audit one promise you broke last year

It would take an hour with whoever was in the room. You want to say what breaking it bought you and whether that was worth the trust it cost.

The next founder who considers selling to you will call the last one who did. Everything you are bidding with is already in that phone call.