The quest for the perfect business

Ask us to describe the perfect business and we would probably portray it like this. It dominates its market. Its customer relationships deepen over time — knowledge of their challenges and needs accumulates with intimacy, raising barriers few dare challenge. The service it provides is critical; switching is painful and rarely contemplated. Revenue is recurring, pricing power the natural reward of indispensability, with profitability following almost inevitably. Capital intensity is low. Organic growth demands little incremental investment. The business can expand within its existing customer base by broadening its offering, while gradually extending into adjacent markets over time. Cash generation is exceptional.

To be perfect, it would not stop at these organic qualities alone. The company would be able to take cash and deploy it, decade after decade, into businesses with similarly durable moats and high returns on capital — creating a scalable system in which virtually every dollar of free cash flow is continuously reinvested at attractive rates.

Studying Constellation Software businesses, financials, operating model, and acquisition pattern feels like watching that portrait come to life — as if Mark Leonard, when founding the company in 1995, set out to engineer precisely such a creature.

Looking back at Constellation’s history and reading Leonard’s letters, we see that most of the company’s defining qualities were conceived and embedded from the very beginning. Yet over time, Leonard went beyond incremental fine-tuning, periodically rearchitecting entire parts of the system through deep thinking and careful experimentation to make it more scalable, more efficient, and better able to capture the full potential of what Constellation has become.

While the perfect business does not exist in the real world, Constellation comes close in many regards.

A permanent home for vertical market software

Constellation Software is, at its core, a “great perpetual owner of VMS (Vertical Market Software) businesses.” It acquires vertical market software companies, holds them indefinitely, and reinvests the cash they generate into new acquisitions.[1]

It does not buy and sell businesses, pursue large-scale restructuring programs, or integrate companies into a centralised platform. It owns them.

Today, Constellation operates through nine autonomous operating groups, which together oversee more than 1,000 individual software businesses spanning over 100 distinct industry verticals.[2]

The diversity is remarkable: software used to design Burberry’s textile patterns; a near-monopoly platform for carpet design relied upon by every major U.S. manufacturer; point-of-sale systems for hair and beauty salons deployed across 40 countries; facilities and ticketing software that dominates university athletics management across American campuses; transit systems; municipal services; auto repair shops; and countless other niche workflows outsiders rarely notice.[3]

What unites these businesses is not the industry they serve, but the role they play within it. Each occupies a deeply embedded position inside a customer’s operations — invisible when functioning properly, yet enormously disruptive to replace. Switching costs emerge not principally from contractual lock-in, but from years of accumulated workflows, integrations, employee familiarity, and institutional dependence.

This is the essence of what makes vertical market software so attractive. These products exist to serve the highly-specific needs of a narrow customer base. The markets themselves are intentionally small — too limited to attract sustained attention from large technology companies or private equity yet sufficiently specialised to sustain dominant competitive positions with exceptional economics.

Constellation’s businesses have high recurring revenues, strong margins, low capital intensity, and pricing power that persists even as products mature. The beauty of this business is not only highly defendable ROI, but an expanding one when it can organically grow, as so much of the incremental revenue flows to the profit line.[4]

A capital allocation machine

A business like that can print cash — quietly, reliably, and with minimal interference. Leonard understood this from the start. His challenge was never how to run these businesses — their managers would handle that, under close measurement and coaching. His problem, the good kind, was figuring out how to redeploy the cash they generated at returns equally as attractive as the businesses themselves.

The key pillars of Constellation’s acquisition model are valuation conservatism and a high return hurdle well into the 20s. Constellation has continuously stuck to that disciplined approach even as rising free cash flow created growing temptation to lower the bar. Leonard noticed that “hurdle rates are magnetic” — once lowered, returns on most subsequent acquisitions drift down to meet the new threshold.

Constellation’s ability to source deals is unparalleled. Beyond the large deal pipeline — the target pool is now believed to be approaching 100,000 and growing by a few thousand every year — it has built something rarer: a reputation. Over 30 years, the company has cultivated relationships with thousands of potential targets, many of which it has been in dialogue with for years before any transaction materialises.[5]

That patient, relationship-driven approach allows Constellation to access deals that never reach a competitive auction — and transact at multiples akin to private market norms, rather than the inflated prices of contested processes.

The knowledge advantage compounds this further. With decades of post-acquisition data across hundreds of verticals, Constellation understands the economics of VMS businesses better than almost any other buyer.

Then there is the seller perspective. For a founder who has spent decades building a business they care about, Constellation offers something most acquirers cannot: the certainty the business will be respected, the team will keep its autonomy, the culture will be preserved, and that it will be encouraged to invest and thrive.

Customers understand this too. Constellation CEO Mark Miller confirmed recently that customers actively steer the companies they work with toward Constellation — drawn by the prospect of stable ownership, continued product investment, and a long-term partner that is not there simply to double its money and sell.

Every successful acquisitive compounder eventually runs into the same wall: the law of large numbers. As the capital base grows, the deals required to move the needle get larger, competition intensifies, and the discipline that built the machine comes under pressure.

Leonard studied this law carefully, spending years analysing what he called High Performance Conglomerates — Berkshire Hathaway, Danaher, TransDigm, and others — to understand how each adapted to these natural constraints.

Berkshire’s answer, for example, was fewer, larger bets concentrated in the hands of Buffett and Munger. The fragmented and niche nature of VMS made that path far less optimal for Constellation. Leonard’s answer therefore was the opposite: build a system of extreme delegation where the acquisition machine could scale without head office at the centre.

The organisation is built to compound

The idea was to build an HR system that retains the best Business Unit (BU) managers, and trains those with the right character to become competent part-time capital allocators. BU managers run their verticals independently, overseen by Portfolio Managers, who report to Operating Group leaders, who report to head office and the board. A BU manager who masters their vertical becomes a Craftsman; a Coach is one who acquires additional businesses; a Portfolio Manager allocates capital across a portfolio; and a Compounder is the rarest of all. The capital you are allowed to deploy increases as you move up that pyramid.

The organisational architecture that emerged is layered but remarkably lean. We experienced this first-hand when we met CFO Jamal Baksh in a modest Volaris office.

The machine works because it is also built on trust — and trust, as Miller put it, begins with measurement. Underperforming businesses are not abandoned or restructured from above; managers are offered time and guidance first. If improvement does not follow, more decisive measures are applied. To help, Constellation can draw on something most acquirers cannot match: a deep internal pool of proven talent, with track records across dozens of verticals, ready to step in and run a particular business or solve a particular problem.

The incentive structure has been designed with equal care, with one overriding objective: long-term value creation for shareholders. Constellation issues no stock-based compensation — Leonard considered the stock too precious to dilute. Instead, key managers are required to invest a meaningful portion of their after-tax bonus directly into Constellation shares, which are held in escrow for a prescribed period. Performance across businesses tends to be measured simply: return on invested capital and organic growth, metrics that are hard to game and that align with what actually creates value over time.

The system surpassed Leonard’s own aspirations. He talked in 2015 of training “a couple of hundred BU managers” to reach 100 acquisitions per year — a target that read almost aspirational at the time. Today, Constellation exceeds that pace, and its portfolio has grown to 1,600 businesses.[6]

The result for its long-term investors has been exceptional with a stock that has compounded at an annual rate of 28.9% since its IPO in May 2006, its stock price was multiplied by 128x.[7]

Constellation’s shares have historically traded at a premium to any reasonable estimate of its intrinsic value, with investors willingly paying for the company’s proven ability to compound capital through acquisitions year after year. That premium left no margin of error.

The latest correction in the stock price solves that issue. At well below 15 times future free cash flow,[8] we think we are no longer paying for Constellation’s ability to perfectly evade the gravity of the law of large numbers. The valuation offers a wide margin of safety even if the reinvestment possibilities narrow or returns compress somewhat from the extraordinary levels they historically recorded.

The well-oiled system Constellation has built may continue to allow it to reemploy part of its cash in the purchase of VMS could yield returns above 25%.[9] Meanwhile, Constellation has been expanding its aperture. It has selectively pursued larger horizontal software businesses where its operational expertise can be applied. Recently, it also resumed purchases of minority stakes in larger listed VMS under its PEMS (Permanent Engaged Minority Shareholder) investment program pioneered by Leonard, where Constellation will make use of both its knowledge edge spotting investment targets, and advising investees on various topics to help them lift their long-term value.

Will artificial intelligence change the game?

Constellation’s market value was pressured by the narrative that AI will allow customers to build their own software, collapsing switching costs and pricing power. We believe this is what’s on the cards for commodity software — generic tools serving multiple industries and executing tasks that can be carried out by a cheaper and more automated alternative.

Some of Constellation’s businesses may sit closer to that end of the spectrum, and we do not dismiss the risk those face. But it misreads what most of the Constellation portfolio actually is.

Their moats reside in years of accumulated workflows, embedded data, institutional familiarity, and the trust that comes from a software provider that has never let them down. And those advantages for a price that, by and large, remains under 1% of the customer’s revenue.[10]

These are not assets that a cheaper development tool could erode easily. The solutions’ mission criticality is genuine, and measured by management using the litmus test of whether the business would die if the software was shut down, or when not, the degree of impact it would have.

A telco whose network management software goes dark faces catastrophic consequences. A transit authority whose scheduling system fails brings a city to a standstill. These are not applications that a customer replaces lightly — not because the technology could not theoretically do it, but because the risk of getting it wrong is existential.

For these businesses, not only do we believe they will dispel stories around disruption, but it seems obvious to us they are best positioned to bring that innovation to their customers.

Deep expertise of the vertical they cover enabled the Constellation companies to generate myriad ideas for ways to deliver more value for the businesses they serve. The bottleneck has often been technical difficulties and the scarcity of knowledgeable workers able to develop these solutions. AI represents a historic opportunity to unlock that value.

In addition, we believe the Constellation “galaxy” has an advantage over standalone software in leveraging AI. Knowledge-sharing across Constellation has always been a defining feature. Businesses are exchanging learnings on how to use AI to serve customers better, and develop agentic solutions and custom AI models the VMS of the group can access, tune and deploy in their market.

The company is not a victim of the AI wave. It is increasingly positioned to leverage AI to expand its organic value creation.

The fear instilled by the revolution ahead and the compression in valuation multiples should also provide good acquisition opportunities. In such a context, Constellation’s privileged buyer status will give it an edge, while the private equity groups that have brought ferocious competition for acquisitions might well see their funding run dry, at least temporarily.

The architect steps back

Finally, a word on Mark Leonard. He stepped down as President last year for health reasons and subsequently announced he would not stand for re-election to the board. He is a visionary and one of the shrewdest businessman and investor of the last 30 years. His departure naturally raises questions.

Yet we believe the market is making a category error. For all his contributions to the acquisitions Constellation closed — and we do not underestimate them — Leonard’s most enduring value-add was the system he built, to ensure the company could continue to thrive with his minimal hands-on involvement, and beyond.

The culture of measurement and meritocracy, decentralisation, the hurdle rate discipline, compensation structure, knowledge-sharing networks, the acquisition frame outside VMS… these do not live in one man. They are embedded in the organisation.

Warren Buffett once observed, only half in jest, that he looked for businesses that could be run by a ham sandwich, so durable were their competitive advantages. He did not mean management was irrelevant. He meant the best businesses are engineered to outlast their founders. Leonard spent 30 years engineering precisely that.

The culture he built goes deeper than process. A couple of years ago he went so far as to surrender his own income because he felt uncomfortable flying business class while shareholders flew economy. The level of ethics and devotion to shareholder interests permeates the organisation durably. It is now part of Constellation’s identity.

In addition, Mark Miller, his successor, is anything but a caretaker. He has been at Constellation nearly 30 years, built Volaris into the group’s most successful operating unit, championed knowledge-sharing across the organisation from its earliest days, and executed the early business split that prevented bureaucracies from spreading across the group. We also believe the qualities Miller brings — being more hands-on and slightly more forceful with underperforming group businesses — could prove particularly timely as Constellation seeks to leverage its asset base with the development and deployment of AI solutions.

To conclude, we would say the architect has stepped back. The building stands, its foundations solid. And we believe the compounding will continue. Warren Buffett claimed he preferred great businesses at a fair price over fair businesses at a great price. At Constellation’s current valuation, we are convinced we have both.

Key Risks

• Information technology companies are subject to specific risks such as rapid technological changes and short product lifespans
• The value of underlying securities can be affected by daily financial market movements
• When investing your capital is at risk.