Speedwell Research's Drew Cohen shares thesis on FinTwit favorite, Constellation Software $CSU.TO

Episode Summary

Executive Summary: The episode examines Constellation Software’s history, culture, and valuation through a deep debate on whether its acquisition-driven compounding can continue at scale. The conversation focuses on Mark Leonard’s decentralized model, the durability of vertical market software, acquisition competition, organic growth, and whether current expectations embed too much optimism despite the business’s extraordinary track record.

Main Topics: Constellation Software’s origin and model (Priority: 5/5): The guest traces the company’s founding in 1995 by Mark Leonard, who built a Berkshire-like portfolio of vertical market software businesses through disciplined small acquisitions and decentralized capital allocation. Scalability of acquisition-led compounding (Priority: 5/5): A central debate is whether Constellation can keep deploying large amounts of free cash flow at attractive returns as its market cap and deal sizes grow, especially amid private equity and strategic competition. Reverse DCF and implied returns (Priority: 5/5): The discussion uses a reverse discounted cash flow framework to estimate what current valuation implies about future capital deployment, organic growth, and ROIC, suggesting expected returns are meaningfully lower than historical compounding. Organic growth, pricing power, and churn (Priority: 4/5): The speakers discuss how Constellation generates low-to-mid single-digit organic growth through pricing, churn management, and incremental customer modules, with recent quarters showing improved organic performance. Capital allocation discipline and incentives (Priority: 4/5): They cover Constellation’s reluctance toward buybacks and ordinary dividends, its preference for acquisitions and occasional special dividends, and the internal incentive system that forces senior managers to buy stock with cash bonuses. Defensibility of vertical market software (Priority: 4/5): The guest explains why these mission-critical, highly sticky software systems are difficult to replace, including implementation complexity, customer inertia, and service dependencies that reduce switching risk even in the face of AI concerns. Comparisons to Berkshire, TransDigm, and roll-up risks (Priority: 3/5): The episode contrasts Constellation with Berkshire Hathaway and TransDigm, while also exploring risks common to roll-ups, such as valuation compression, pricing scrutiny, and incentive misalignment.

Key Arguments: Constellation’s core edge comes from disciplined acquisition of small, sticky vertical market software businesses at low multiples with strong free cash flow conversion. The firm has proven able to scale acquisition activity despite rising competition from private equity and larger buyers. Current valuation likely assumes continued strong acquisition deployment and solid ROIC, but not necessarily the same historical 30%+ stock returns. Organic growth has recently improved to around 5% to 6%, better than prior assumptions of 0% to 2%, helping support the thesis. The company’s decentralized structure pushes capital allocation decisions closer to operating units, which helps it source and execute many small deals. Constellation’s businesses are protected by high switching costs, embedded workflows, and implementation complexity, making them resilient against AI or competitor disruption. If acquisition opportunities dry up, the company may rely more on debt, special dividends, or spin-offs to deploy capital efficiently. The biggest risk is not obvious business collapse but lower future returns from reduced acquisition opportunities, valuation compression, or changes to the incentive system.

Data Points: Founding year: 1995 - Constellation Software was officially founded in 1995 by Mark Leonard. Initial acquisition price multiple: 1.0x to 1.25x revenue - Leonard reportedly bought early vertical market software companies at very low revenue multiples. Typical free cash flow margin: ~20% - Used to illustrate attractive economics of early acquisitions. Implied free cash flow multiple: ~5x to 6x FCF - Derived from the revenue multiple and margin assumptions for early deals. Approval threshold for deals: Originally $2 million, later $20 million - Internal acquisition approval limits increased as the company scaled. Current revenue: ~$8 billion - Approximate consolidated revenue, with some caveats around partial ownership and consolidation. Current free cash flow: ~$1.1 billion - Approximate annual free cash flow mentioned during the discussion. Number of vertical market software companies: 750 to 1,000 - Estimated count of businesses in the Constellation portfolio. Hurdle rate: At least 20%+ incremental return - Leonard indicated the company seeks at least 20% incremental returns on capital. Large acquisition cadence: ~3 large deals per year - Management claims to have completed about three large acquisitions annually recently. Observed deal market size: 70+ large VMS acquisitions in recent years - Used to argue the opportunity set remains large despite competition. Known vertical market universe: 50,000+ companies - Management says it tracks a large universe of potential targets. Unknown deal visibility: ~70% of annual deals unknown to Constellation - Management claims they do not know about most deals occurring in the market. Capital deployed on acquisitions: >100% of free cash flow in past 3 years - Shows acquisition spending has exceeded internally generated free cash flow. Organic growth assumption (earlier view): 0% to 2% - Earlier base-case assumption for organic growth in the reverse DCF. Recent organic growth: ~5% to 6% - Organic growth in the last three quarters was said to be higher than expected. Churn rate: ~7% - Discussed as churn mostly offset by new modules and customer additions. Internal development payback: 5 to 7 years - Management indicated payback on internal software development can take this long. Acquisition payback comparison: ~3 years - Small acquisitions at about 1.33x free cash flow imply a faster payback than internal development. Reverse DCF implied shareholder return: ~6% to 12% - Estimated return range under different assumptions at the cited share price. Conservative reverse DCF scenario: ~7% return - A draconian scenario with 0% organic growth, 20% ROIC, and declining deployment rates. Share price referenced: $3,720 per share - Used as a benchmark for discussing the reverse DCF valuation framework. Employee stock comp rule: 75% of cash bonus must be used to buy stock - Described as a key incentive mechanism aligning employees with shareholders.

Pivotal Quotes: "The entire question with Constellation software is whether or not you believe they'll be able to continue to get these deals and scale up." — Drew Cohen: Summarizing the central investment thesis and main long-term risk. "The market could totally change. It could get more frothy. And they are very disciplined in their acquisitions and hurdle rates." — Drew Cohen: Explaining why future capital deployment may be the key variable for returns. "We have too much money. The opportunities are out there." — Andrew Walker quoting Leonard/Consolidating theme: Describing Constellation’s posture toward capital allocation and the pressure to keep deploying cash.

Implications: Constellation remains high quality, but future returns depend more on sustained deal flow and disciplined deployment than on business quality alone. Listeners should focus on valuation, scale constraints, and whether the firm can keep compounding as acquisitions get larger and scarcer.