Actionable insights — Who is Chuck Akre?

1. Apply the three-legged stool as three separate pass/fail tests

1. Business: establish the return the company earns on the capital already in it. High and stable ROIC is the entry ticket, not the conclusion.
2. Management: require skin in the game — a meaningful personal stake, not a pay package. Where no founder stake exists, note that the leg is weak rather than substituting something else.
3. Reinvestment: ask how much of this year's cash flow can be put back into the business at a similar rate. This is a capacity question about the opportunity set, not about willingness.
4. Fail any leg and the stool falls over: a high-ROIC business with nowhere to reinvest is a dividend or buyback story, which is a different and lower-return proposition.
5. Only when all three hold do you have what Akre calls a Compounding Machine, and only then does a long holding period do the work.

• Reinvestment capacity that exists on paper but has never been demonstrated — check the actual reinvestment rate over five years.
• A management stake that is large in dollars but tiny as a share of their own wealth, or vice versa.

2. Treat reinvestment capacity as the binding constraint, and test it explicitly

1. Take last year's free cash flow and ask, concretely, where the next unit of it goes: new stores, new acquisitions, new capacity, or out of the door as dividends and buybacks.
2. Estimate the return on the reinvested portion separately from the return on the existing base. A blended ROIC hides a deteriorating marginal return.
3. Estimate the runway in years, not in adjectives — how many more stores, acquisitions or plants exist before the opportunity is exhausted?
4. Rank candidates by reinvestment capacity when their business quality is comparable; that is where the difference in compounded outcomes comes from.
5. Reclassify a business that has run out of runway. It may still be worth owning, but it is now a cash-return story and should be valued as one.

• Reinvestment into worse opportunities as the good ones run out — growth for its own sake destroys value fastest at high-ROIC companies.
• Acquisition-driven reinvestment where returns depend on continued access to cheap targets.

3. Decide your concentration deliberately, and check what the professionals you admire actually do

1. Write down what proportion of your portfolio your top five positions represent, today.
2. Compare it with managers whose method you are borrowing. If your process is theirs but your concentration is far lower, one of the two is inconsistent.
3. Ask what your concentration implies about your confidence: a 4% position says you think the idea is roughly as good as your twenty-fifth best.
4. If you want lower concentration, be explicit that you are trading expected return for lower variance and behavioural comfort — a legitimate choice, but a choice.
5. Re-check after every purchase; concentration drifts silently as winners run.

• Copying the concentration without the underlying research depth — five positions requires knowing five businesses extremely well.
• Concentration in five names that share one mechanism, which is really one position.

4. Record purchase vintages, so the holding period is a fact rather than an aspiration

1. For every position, record the year of first purchase and never overwrite it.
2. Report returns from that date rather than from an arbitrary window; this is what makes a long-term claim checkable.
3. Review the distribution of your holding periods annually. If the median is under two years, you are not running the strategy you think you are.
4. Note the trims as well as the buys — a truthful long-term record includes the exits.

• Survivorship in the vintage list — the positions that did not work are rarely the ones profiled.
• Treating "never sell" as the rule when even the 36-year holding was eventually trimmed.

5. Support a preference with outside evidence rather than an assertion

1. When you rely on a general belief — owner-operators outperform, founders allocate better — find the study or dataset that supports it.
2. Cite it in the write-up, so the belief can be challenged on its evidence rather than on your authority.
3. Check what the study actually measured: profitability is not the same as shareholder return, and family control is not the same as founder management.
4. Note the exceptions you have personally seen; the archive's own KPG.AX governance problem is a founder-control failure.

• A study measuring the average when your decision concerns a specific company.
• Founder control that becomes entrenchment — the pledged-shares problem this archive documented elsewhere.

6. Use another manager's disclosed book as a cross-check on your own, and read the non-overlap

1. Pick managers whose stated method genuinely matches yours, not merely whose returns you envy.
2. List the intersection with your own portfolio explicitly. Shared names are corroboration — independent research reaching the same conclusion.
3. Then study the non-overlap in both directions. What do they own that you rejected, and why? What do you own that no comparable manager holds?
4. Treat the second question as the more valuable one: a position nobody like you owns is either your edge or your blind spot.
5. Never buy a name because a manager holds it. You do not know their entry price, their sizing rationale or their exit condition.

• Filings lagging by a quarter — a disclosed book is a photograph of the past.
• Comfort-seeking: overlap feels like validation, and validation is exactly what a contrarian process should not optimise for.