CEO Series — Mark Leonard
The manual beats the myth.
THE TRUE STORY NO ONE TELLS
Most people can name Constellation Software. They’ll call it a roll-up. A vertical-market software acquirer. A Canadian compounder. They’ll quote the stock chart, mention Topicus and Lumine, and say something about “mini-monopolies.” None of those labels are wrong. None of them explain why Constellation out-competed almost every serial acquirer of its era.
To get Constellation, you can’t stare at the assets. You have to study the operators behind the assets—and the constraints they chose. Constellation is not a fairy tale about a founder with perfect instincts. It’s a manual written in public, enforced in private, and designed to survive bad pricing, bad seasons, and bad moods.
Mark Leonard built the architecture. He wrote down a few rules and removed everything that would turn those rules into slogans. The center stays thin. Decision rights live at the edge. Cash per share beats conversation. When returns don’t clear the bar, you don’t lower the bar—you wait. Later, others—platform leaders and operating groups—took that architecture and made it breathe in Europe, in telecom, and across dozens of verticals. The story is not a cult of personality. The story is process.
This is that story—clean, straightforward, plain English.
THE PAGE THAT EXPLAINS THE SILENCE
The moment the noise stops, the work usually gets better. Leonard did something modern CEOs rarely do: he turned off the microphone. He said, in effect: We’ll write when we have something new and important. We’ll post the tables once a year. You won’t get the playbook. You’ll get the standards.
There are only a handful of standards that matter:
Free cash flow per share is the compass. Report IFRS for comparability; decide on cash per share because it naturally includes dilution, interest, and capex.
A board-set hurdle defines reality. Compare ROIC + organic growth to that bar. If it clears, deploy. If it doesn’t, build cash.
Decentralize diligence; centralize standards. The head office is a standards keeper, not a deal factory. Operators closest to customers make the calls.
Rarely sell. The value is in long holding periods of small, sticky products, not in flipping.
Speak last. Reduce disclosure that gifts tactics to imitators. Publish the math; protect the method.
This is not mystique. It’s edge protection. When you put the rules in the open and keep the tactics quiet, you make imitation look simple and execution look boring. That’s the point. Boredom is a moat when everyone else needs applause.
THE HURDLE: SIMPLE, HARD, NON-NEGOTIABLE
Most acquirers discover a religion called “deal flow.” It feels great. Then it kills them. Constellation replaced the religion with a gate.
The gate: ROIC plus organic revenue growth must meet or exceed a board-set hurdle rate. The hurdle is not a vibe. It is a line in the sand. When pricing in the market is silly, Constellation does the unintuitive thing: it lets cash pile up. Not because hoarding is heroic, but because compounding is a fragile instrument. A single bad cohort—done at the wrong price, under the wrong pressure—can flatten years of careful work. The decision to sit still is not indecision. It is design.
There’s also a small human concession in the system: borderline learning deals. A tiny sliver of capital is allowed to go to a close call—only if the team writes down the lesson and tracks whether the lesson paid for itself. That keeps intellectual humility in the machine without letting “experiments” become a back door to mediocrity.
Between the lines: “No deal” is a result. The absence of activity is not the absence of judgment. It is the presence of discipline.
THE STRUCTURE: STANDARDS AT THE CENTER, DECISIONS AT THE EDGE
Ask ten companies to describe “decentralization” and you’ll get ten posters of org charts. Constellation’s decentralization is not a poster; it’s permission. The head office is small by design. It does not pretend to know every product or every niche. It sets the hurdle. It sets the reporting discipline. It moves capital to where standards are met. Everything else—pricing, product choices, customer negotiation, integration tempo—belongs to the operating groups.
Because decisions live at the edge, diligence is local. Operators see utility before outsiders see “TAM.” They see price before bankers see comps. They see whether the software is used or just installed. And because the center doesn’t micromanage, local teams write the truth faster. You don’t need a memo contest to fix pricing on a product you speak with customers about every day.
A thin center has a second effect: it keeps the firm quiet. Quiet firms compound better because they don’t have to perform confidence for outsiders every quarter. Standards don’t need charisma. They need execution.
WHAT CONSTELLATION ACTUALLY BUYS AND WHY IT STICKS
Strip away the romance. Constellation buys vertical market software—narrow products, embedded in workflows, with high switching costs, and revenue dominated by maintenance. The beautiful, boring line called maintenance is the most honest signal in the system. When maintenance grows and new maintenance appears, customers are still choosing function over flash. When it stagnates, the filings will tell on you faster than any speech.
These products are often small in the public imagination. They sit deep in specific industries—parking systems, local government billing, lab software, hospitality back-office, telecom provisioning, education scheduling. They don’t trend on social platforms. They keep invoices flowing, doors opening, records accurate, and compliance sane. That’s why they stick.
Holding periods are long. Pricing power is gentle, cumulative. The “growth” is not fireworks; it’s accretion. You buy, you tidy, you keep, you invest with a screwdriver, not a billboard. The compounding is mostly invisible—until you zoom out a decade.
THE TWO FORKS THAT PROVED THE CULTURE (TOPICUS & LUMINE)
When the hunting ground widens, many companies centralize. Constellation does the opposite: it forks. Topicus in Europe. Lumine in telecom/media. Both came to shareholders via dividend-in-kind distributions—not as a publicity stunt, but as an engineering move. Every fork carries the same DNA: the hurdle, the center-thin ethos, the cash-per-share bias, the “rarely sell” stance. By cloning culture instead of bloating headquarters, Constellation kept velocity and standards.
Forking does something subtle: it preserves surface area (more places to hunt) without creating one giant bureaucracy that must pretend to know every alley. You extend the machine by copying rules, not by adding managers.
ENTER THE SUCCESSION: WHY THE MACHINE WAS BUILT TO OUTLIVE ITS BUILDER
Founders often fail the final exam: letting go without letting go of the rules. Constellation designed for the day when the founder would step back. The choice to publish standards and reduce commentary was not stagecraft. It was a way to institutionalize judgment.
A real succession in this model isn’t about matching a personality. It’s about honoring the gate. If the bar stays where it is. If local decision rights stay where they are. If the cash-per-share bias stays where it is. The founder can change seats and the machine still composes the same song.
This is why the press-release moments matter less than the year-end tables. The tables will tell you whether the rules lived on.
HOW THE LEONARD SYSTEM REALLY WORKS
1) Capital intake
Cash comes in from hundreds of narrow products. Each little stream is boring alone and powerful in aggregate. Because the firm rarely sells, more streams join every year.
2) Gate check
Op-groups bring forward targets. The center tests them against the hurdle: ROIC + organic versus the bar. No sermons. No hackathons. Either it clears or it waits.
3) Local diligence
Operators closest to customers run the reality check. Is this software truly embedded? Will new maintenance appear? Are we buying at a price that respects the base rate for this niche?
4) Close, tidy, normalize
The playbooks are not public, but the pattern is visible: stabilize, clean up contracts, remove friction in invoicing, rescue pricing where it’s asleep, modernize where it’s cheap, leave alone what already works.
5) Report, don’t boast
Once a year, publish the numbers that matter. The rest of the year, run the machine.
6) Sit still when it’s wrong
When markets get hot and bankers are auctioning “platforms” at thrill-seeker multiples, the center does nothing. Cash balances rise. Bidders enjoy their adrenaline. Constellation enjoys its standards.
That loop repeats. It’s not sexy. It scales.
WHY “BORING” BEATS “BRILLIANT” IN THIS GAME
Brilliant acquirers write memos about synergy. Boring acquirers ask whether customers will notice if the software’s owner changes and whether cash per share rises after the champagne goes flat. Brilliant acquirers publish thick decks about “integration roadmaps.” Boring acquirers fix the billing system, give the product manager a budget that matches reality, and leave their phone number with the one customer who always calls at 6:10 p.m.
The lore says “culture eats strategy.” In Constellation’s world, structure eats performance anxiety. The thinner the center, the fewer places insecurity can hide. The higher the bar, the less temptation to bluff. The clearer the metric, the fewer excuses. You don’t need slogans when the rule that matters is a fraction: cash / share.
WHAT CONSTELLATION ACTUALLY DOES
Buys small, sticky software in verticals most people overlook.
Keeps it. The exit is time, not a sale.
Tidies operations: billing, pricing, support discipline, measured product work.
Allocates by a bar: capital only to places that clear ROIC + organic.
Clones the culture when scope widens (Topicus, Lumine).
Publishes the receipts once a year. Stays quiet the rest of the time.
That’s it. No TED Talk. No “Day One” chant. No “move fast and break things.” Move steady and compound things.
WHY CRISES FAVOR CONSTELLATION
People say Constellation is resilient. That undersells it. The firm is built for bad weather.
Locked-in revenue from maintenance means demand wobbles less than discretionary projects.
Small, mission-critical niches break last, because ripping out the back-office system at a water utility or a clinic is not a decision people love in a panic.
Dry powder accumulates when the bar blocks deals, so when prices fall for the right reasons, you aren’t waiting on anyone’s permission.
Decentralized eyes see distress early: local operators hear the rumor before the banker files the deck.
Crises don’t make Constellation smarter. They strip competitors of their excuses. A forced seller across the table is not a market inefficiency; it’s a decision tree that just got shorter.
WHAT YOU CAN ACTUALLY LEARN FROM LEONARD
Structure beats forecast. A mediocre forecast inside a sound structure compounds better than a brilliant forecast inside a leaking one. If your hurdle is honest and your center is thin, mediocrity has fewer doors to sneak through.
Liquidity kills if you depend on it. Liquidity looks friendly until it asks for its money back. Long holding periods and small dependable cash streams produce the opposite: time.
Conditions over stories. You do not need a narrative to decide. You need a price, a product that matters, and a clear gate. Tell less. Measure more.
System thinking. Maintenance → new maintenance → retention → cash per share. Add capital deployment pace and you’ve got the dashboard of a compounding machine.
Discipline compounds faster than excitement. The market often pays you for acting cool. Cool is learned behavior. It sounds like, “We’ll wait.”
DEM — DECISIONS, EXECUTION, MINDSET
Decisions (the architecture).
Define “good” with numbers people can’t sweet-talk. Place the hurdle outside the room so the room can’t bargain with it. Decide which few metrics explain most of reality. Decide what not to collect. Decide how not to brag.
Execution (the rhythm).
Move at the speed of conditions, not headlines. Scale what works locally before you export it globally. Treat “integration” as a series of small humiliations you remove from a customer’s week. Divide work into what actually makes cash show up and what just makes a dashboard greener.
Mindset (the stance).
The operating stance is calm. Calm is not slow. Calm is the difference between watching a bid process and needing to win it. Calm writes clearly, sizes cleanly, and avoids “we deserve this” speeches. Calm is a cash posture.
COMMON MIS-READS AND HOW CONSTELLATION AVOIDED THEM
“Scale = safety.”
Wrong. Scale is a sensor network only if you build the pipes. Constellation’s scale helps because local leaders can send true signals upstream without getting translated into a slideshow.
“It’s a Canadian Berkshire.”
Lazy. Berkshire owns whole businesses and insurers and invests across instruments. Constellation is a software operator with a capital allocation loop optimized for tiny niches. The common line is discipline. The engines are different.
“It’s just financial engineering.”
If it were, maintenance wouldn’t matter. In this system, maintenance is the heart rate monitor. Engineering without pulse is a corpse.
“Topicus and Lumine were PR moves.”
No. They were forks—culture export to widen surface area without raising a bureaucracy.
“It’s all multiple expansion.”
Multiple expansion does not write maintenance checks. Customers do.
THE CEO BLUEPRINT AND HOW ONE BRAIN BUILT A MACHINE OTHERS CAN RUN
The founder’s job is not to stay founder. It’s to make staying founder unnecessary. Leonard’s blueprint did that by separating rites from rights. The rite is the story; the right is the rule. He traded rites for rights.
Rights over stories: operators own decisions; HQ owns standards.
Rights over speeches: the bar decides; not the calendar or the podium.
Rights over hype: cash per share is the applause meter; nobody cares what the room felt.
With that, anyone who respects the rules can run the machine. The successor’s talent is not to echo the founder; it is to enforce the bar, guard the calm, and keep the center thin.
WHAT CONSTELLATION LOOKS LIKE WHEN IT’S WORKING
A year passes. There are fewer headlines than you’d expect. The shareholder package arrives. You see maintenance that still prints. You see small new maintenance lines that prove the product is still solving a job. You see capital deployment paced—not manic, not idle. You see realized returns that sit near the bar the company wrote down years ago. You do not see a parade of disposals. You do not see a rationalization project masquerading as “focus.” You do not see a slide where page one says “Our Vision” and page two says “Our Values.” You see receipts.
That’s the point. The manual does not need a myth when the receipts are enough.
HOW TO READ THE MACHINE
Realized returns versus the bar
Find the annual cohort commentary. It’s never a neon sign, but it’s there. If realized returns track the stated hurdle and volume flexes with conditions, the standard is alive.
“New maintenance”
Track the small organic line where disclosed across Constellation, Topicus, and Lumine. It’s not a fireworks metric; it’s a pulse. In tiny niches, a consistent pulse beats a bold promise.
Capital velocity by platform
Watch how quickly each platform deploys cash without the bar drifting. Velocity with integrity is culture export; velocity with drift is appetite.
Hold / sell
The bias is “rarely sell.” If that bias breaks, it first shows up in footnotes—disposals, gains/losses—not on a stage.
Periods of inaction
When private valuations run hot, the tables show slower M&A cadence and higher cash. That sequence is not failure; it’s the manual acting like itself.
Disclosure cadence
Expect fewer essays and continued tables. That is not a PR strategy; that is a risk policy.
THE OPERATOR CHECKLIST
Cash per share: up across a cycle?
Hurdle: unchanged in spirit, not just wording?
Decision rights: still at the edge?
Capital velocity: rising without bending the bar?
New maintenance: small, boring, positive?
Disposals: rare, sensible, explained?
Center: thin?
Tone: calm?
If those read correctly, the founder’s seat is trivia. The manual is the story.
CLOSING: WHY THIS MATTERS NOW
Public markets taught a generation to worship narrative and next-quarter math. Constellation did something less romantic and more durable. It taught a small army of operators to prefer receipts to rhetoric, to prefer bars to debates, and to prefer quiet to applause. That’s not minimalism. That’s an operating system for compounding in the real world, where prices get silly, egos get loud, and attention gets harvested.
You don’t have to admire it. You just have to read it.
The founder changed roles. The manual didn’t. If the manual is real, you will see it in the same small places every year: free cash flow per share, a bar that does not move, and maintenance prints that keep saying the software still matters. The rest is noise.
Thanks for writing this, it clarifies alot. How sustainable, though? Great insight.