Actionable insights — Perimeter Solutions: copying a 3,000-bagger

1. Source ideas by following proven operators to their next, smaller vehicle

1. Identify people with a completed, verifiable compounding record at a company you already understand.
2. Find where they went next, and how much of the old team came with them — one person is a name, a group is a system.
3. Check the new business rhymes structurally with the old one, so the playbook actually transfers.
4. Verify their alignment: ownership, open-market buying, and the terms on which they are paid.
5. Size the new vehicle against the old to see how much runway the model has left.

• Reputation substituting for analysis. The unrelated acquisitions here — medical machinery, circuit boards, an oil additive — are excused explicitly "because of the exceptional track record of management", which is a judgement, not a fact.
• Terms that differ from the original vehicle. The people are the same; the compensation is not.

2. Test pricing power on three conditions, and insist on the third

1. Is the product mission-critical — does the customer's operation fail without it?
2. Is supply concentrated, ideally a monopoly or a tight oligopoly?
3. Is it a small share of the customer's total cost? This is the condition most screens omit and the one that makes price rises unopposed.
4. Verify with the realised price-increase history rather than the theory.
5. Then check whether the pricing power is being used — modest increases against a monopoly position may mean regulatory or reputational limits.

• A monopoly whose only customer is a government agency. Pricing power against the US Forest Service has a political ceiling, and 3–4% annual increases suggest it is being felt.
• The "small share of cost" argument used for products that are actually substitutable.

3. For a physical business, enumerate the assets a competitor would have to duplicate

1. List the tangible infrastructure required to serve the customer at the required speed — plants, depots, fleet, inventory.
2. Attach a cost to each so the replication bill is a number rather than an adjective.
3. Identify the time barrier separately: approvals, testing, certification, permitting.
4. Look for a technical constraint that prevents gradual switching, which is worth more than any cost barrier.
5. State the resulting market position and check it against a verifiable customer relationship.

• The honest calibration the post itself makes: "With $100 billion, I think Perimeter could be disrupted" — the moat is expensive, not impossible.
• Inventory-heavy moats in a demand downturn. $100m of stock is a barrier in a busy season and a cost in a quiet one.

4. Check whether a diversifying segment actually broke the cycle correlation

1. Identify the external variable the core business is correlated to, and find the historical chart of results against it.
2. Look for a period where the variable was weak and results were not — that is the test.
3. Decompose the outperformance: how much came from the new segment, how much from acquisitions, how much from price.
4. Repeat over more than one cycle before accepting the diversification claim.

• Acquisitions masquerading as diversification. IMS was bought in December 2024, so 2025 growth is partly inorganic by construction.
• Segments that are unrelated to the core and to each other. Diversification of revenue is not diversification of management attention.

5. When compensation distorts the income statement, rebuild earnings from the bottom and show every assumption

1. Identify the distorting item and its size relative to revenue, not just to profit.
2. Add it back to get an operating result, then rebuild forward: organic growth, acquisitions, financing and tax, each stated.
3. Deduct a normalised estimate of the distorting item rather than pretending it is zero.
4. Publish every input so a reader can re-run it with different assumptions.
5. Cross-check against an undistortable measure — gross margin — to confirm the underlying business quality.

• The single soft assumption carrying the answer. The $100m normal-year fee is an estimate that "depends on how well Perimeter's stock performs" — and it is 55% of the final earnings figure.
• The other method disagreeing. Forward price-to-sales says the shares trade "significantly above" their historical multiple; only the normalised-earnings route produces "fair".

6. Judge capital allocation on what was done in the drawdown, not on the policy statement

1. Find the worst drawdown in the company's listed history and look at what management did with capital during it.
2. Compare the average repurchase price then with the price now — that comparison is the entire evidence.
3. Read the stated buyback policy: systematic programmes are neutral, discretionary ones tied to intrinsic value are informative.
4. Check whether an explicit return hurdle governs all uses of capital, not just buybacks.

• Buying back stock while issuing 1.5% of it a year to the founders. The two flows must be netted before calling the company a cannibal.
• Product-line acquisitions, which are harder to verify than whole-company deals — there is no separate set of accounts to check.

7. Price management compensation as a permanent claim on the equity, and compare it to a benchmark

1. Read the actual terms: fixed share issuance as a percentage of shares outstanding, and any share of appreciation, with their end dates.
2. Express the cost as a percentage of revenue in a good year, not as a percentage of profit — profit can be small or negative.
3. Note that a fee on market-cap appreciation makes management's incentive the share price, not the business, and pays out in the years you do best.
4. Compare against a reference point you respect, so the number has a scale.
5. Decide whether the quality of the people is worth the claim — and say so either way.

• A fee that rewards multiple expansion. 18% of market-cap appreciation pays for a re-rating you did not receive as a shareholder until you sell.
• The end dates. The fixed component runs to 2027 and the variable to 2031 — both are inside a long-term holding period.