# Quality Compounders A quality compounder is a business that earns high returns on capital, is structurally protected from anyone competing those returns away, and can put next year's cash back to work at the same rate. Own it long enough and the business does the work. The share price eventually catches up to the arithmetic. The point of this note is to stop me buying good stories. A good story is a business I find interesting. A compounder is a business where I can answer one question honestly: > **Who is the alternative, and how long would it take them to become one?** If the answer is "nobody" or "twenty years", I keep reading. If the answer is "three well-funded startups", I stop. --- ## The five conditions I want all five. Four out of five is a good business, not a compounder. **1. It cannot be replicated with money alone.** Capital is the cheapest thing in the world right now. The moat has to be something capital cannot buy quickly — a permit, a certification, a regulatory approval, a physical position, an embedded workflow. Permitted reserves near a growing city take two decades to assemble. A certification tied to a piece of equipment is permanent for the life of that equipment. Software validated inside a regulated process cannot be swapped without re-validating the process. **2. The customer cannot leave, even when they want to.** Not dissatisfaction — inability. The cost of leaving has to be operational or regulatory, not just financial. See [[Switching Cost Design]] for the dimensions I check, and the test I actually use: *what happens if they stop next Monday?* **3. Pricing power that runs above inflation and ignores the cycle.** This is the engine. Volumes can be flat for a decade and margins still expand, because price goes up every year and the cost base doesn't. A business that can only raise price in a good year is a cyclical wearing a compounder's clothes. **4. Management that reinvests at high rates of return.** The rate the business earns matters less than the rate it can *redeploy* at. A 40% ROIC business returning all its cash is a bond. The best version of this is a disciplined acquirer buying more of the same asset — cash out of the tollbooth buys another tollbooth. The thing to watch is whether they keep paying the same multiple or start reaching. **5. A runway measured in decades.** Not a three-year story. Reserves that last fifty years, a customer cohort that deepens every renewal, infrastructure that grows with population. Compounding needs time more than it needs a good entry price, which is the whole argument in [[Time in the Market]]. --- ## The four shapes it takes - **Physical position** — an irreplaceable hard asset with a permit attached. The constraint is geological or legal, not commercial. Related: [[Bottleneck Business]], [[Vulcan Materials - VMC]]. - **Embedded workflow** — the software the customer's regulated or licensed process runs through. Leaving means redoing the process, not changing a vendor. Related: [[Data Moat]], [[Data Flywheel]]. - **Acquisition machine** — a serial acquirer of small, niche, sole-source businesses in a fragmented category. The moat is the discipline, not any single asset. Mapped in full: [[Acquisition-Driven Compounders MOC]]. - **Network at scale** — two-sided systems where the second side makes the first side stickier, and where the local network can't be replicated from outside. Related: [[Positive Reflexivitiy]]. Map any candidate onto [[7 Powers]] before writing a thesis. If I can't name which power it holds, I don't have a thesis — I have an opinion. --- ## Failure modes The honest half of the note. Compounders break in predictable ways: 1. **The moat was the era, not the business.** Regulation changes, a substitute arrives, the workflow the software owned stops being the workflow. 2. **Reinvestment runs out before the runway does.** The category gets picked clean, the acquirer starts overpaying, and cash starts leaving as buybacks at high multiples. 3. **I paid for twenty years of it.** A great business at a price that already assumes the next two decades is a mediocre investment. Quality is not a valuation argument — see [[Buffet Indicator]] and [[Anthony Bolton Criteria]] for the other side of the trade. 4. **I confused a toll with a tax.** A business extracting rent without improving anything invites the regulator eventually. 5. **I stopped checking.** The whole appeal is that these need little maintenance, which is exactly why the deterioration goes unnoticed. > [!warning] The check I skip most > Pricing power is the first thing to go and the last thing I notice, because it shows up in mix and margin long before it shows up in a headline. Re-read the pricing disclosure every year, not the narrative. --- ## How this sits in the system This is the core of the actively managed side of [[Investing System MoC]]. It is the answer to "what do I actually want to own" — concentrated, few positions, boring, held for a long time. That follows directly from [[Investing Learnings]] and cuts against over-diversifying ([[Investing Diversification]]). The discipline it demands is temperament, not analysis. These businesses go through long stretches where nothing happens and the market pays for something else. That is [[Investing Discipline]] and the marshmallow test in [[Four Ds of Investing]]. It also sets the shape of a thesis: most of the work is establishing the moat is structural, not modelling three years of earnings. New candidates go through [[Investment Thesis Plays]], scored against [[Technical Moat Assessment Framework]] and [[Defensibility Principles MOC]]. Where it does *not* apply: thematic bets, where I'm buying a shift rather than a position ([[Thematic Investing]]), and anything I'm holding for a re-rating rather than for the compounding. --- ## The one-line version Own businesses that own the toll roads. Pay attention to the price, but pay more attention to how long the road stays the only road. See also: [[Investing Principles]], [[Charlie Munger]], [[Mohnish Pabrai]], [[AI era Defensibility]], [[Economic Machine]] #investing #defensibility #firstprinciple