Constellation Software Mark Leonard has built the finest capital allocation machine we've seen in investing. Nobody - not even Buffett - has matched his consistency in deploying capital at 20%+ returns over multiple decades. His shareholder letters are required reading for our analysts.

The VMS rollup strategy is elegant in its simplicity. Buy mission-critical software businesses in tiny markets PE shops won't touch, maintain extreme pricing discipline (1-2x revenue), then let them operate autonomously. Over 700 acquisitions later, they've built an information advantage in VMS optimization that's simply insurmountable. Our findings show Constellation is the preferred acquirer in sub-$5M deals - founders actively seek them out, particularly those burned by PE's slash-and-flip playbook.

"They didn't want their legacy disappearing into the craw of an omnivorous conglomerate. While they knew that they would have autonomy within Constellation, they also wanted identity."

"I don't mind giving away our philosophy and our general approach... what I do mind is giving away our detailed practices and processes that we've learned through painful trial and error."

"We do post-acquisition reviews religiously. You learn more from studying your mistakes than your successes."

The metrics are stunning: 29% OCF/share CAGR since '08, 5-9% churn vs industry 15-35%, 51% stable public sector revenue. Maintenance and recurring revenue now at 71.5% - this is as sticky as it gets in software. Geographic diversity is improving, with 44% US, 33% Europe/UK. But what's remarkable is they've maintained ROIC at 20-30% despite massive scale. This shouldn't be possible in software M&A.

Their decentralized model is brilliant. Six operating groups (Volaris, Harris, Perseus, Jonas, Vela, Topicus) each running their own playbook while sharing best practices. We've studied lots of major software rollup - none have solved the centralization vs autonomy paradox like Constellation. The compensation structure is equally clever - requiring executives to reinvest 75% of after-tax bonuses into shares creates perfect alignment.

The moat is widening. With a database of 40,000 VMS targets and proprietary data from 700+ acquisitions, they see patterns no one else can. Their M&A machine is unmatched - 134 deals last year deploying $1.7B, up from 12 deals worth $29M in '05. Operating margins expanding from 11.2% in '13 to 15.3% TTM suggests they're getting better at integration.

Key pushback from peers centers on growth runway. At $1B annual FCF, they're reaching the limits of small-deal math. Leonard's recent pivot to larger deals with lower hurdle rates makes us nervous - we've seen this movie before with other rollups. But if anyone can navigate this transition, it's him. The Topicus/Lumine spin-offs show he's thinking creatively about the scaling challenge.

"The best businesses are often not the ones that are growing the fastest or have the highest margins... they're the ones that can generate good returns and deploy large amounts of capital at those returns."

Balance sheet remains pristine - 1.2x net debt/EBITDA, 7.8x interest coverage. At 31x forward earnings, it's not cheap. But true compounding machines never are. Our model suggests C$3,646 fair value, likely conservative given their track record of finding new avenues for high-return capital deployment.

"We buy businesses to keep them forever. We're looking for people who want to build something that's going to last."

"The trouble is, if you drop returns on a few marginal investments, you tend to drop the returns on the whole portfolio."

We will look to add on any macro-driven weakness. This is a cornerstone holding for the next decade, but position size needs to reflect the mathematical reality of their scaling constraints. Their recent exploration of opportunities outside VMS bears watching - could be either brilliant evolution or classic diworsification. Knowing Leonard's track record, we're betting on the former.

Bottom Line: This is one of the highest-quality business we've analyzed in the past decade. Leonard has built a perpetual motion machine for capital allocation that we believe will compound at 15-20% for the next decade, even with the scaling headwinds. The combination of mission-critical software, sticky customers, decentralized operations, and unmatched M&A capabilities creates a moat that's getting wider, not narrower. Current valuation offers an attractive entry point for long-term investors. Key monitoring points are the success of larger acquisitions and any ventures outside VMS. But in Leonard we trust - he's the closest thing to Buffett we've seen in software.