Actionable insights — Buying A Forever Winner

1. Test durability with three clauses in three tenses

1. Ask the question in the past tense: was this company the clear leader twenty years ago? A verifiable historical fact, not an opinion.
2. Ask it in the present: is it still the leader? If the answer changed between the two, the moat has already been breached once.
3. Ask it in the future, and require the word "probably" — the honest answer is a probability, and forcing yourself to say it out loud exposes weak cases.
4. If any clause fails, you are underwriting a change of position rather than a durable one, and the required analysis is different.
5. Use the test to filter before valuation work, so you never spend a deep dive on a business that fails clause two.

• Industries where twenty years of leadership is a warning rather than a comfort — the last incumbent before a technology shift also passed clause two.
• Leadership measured by share of a market that is itself shrinking.

2. Count the protected positions, and check whether they reinforce each other

1. Break the company into its reported segments and assess each one separately: is it a leader, and how many real competitors does it face?
2. Count the segments that are genuinely oligopolistic. One is normal for a quality business; several is unusual and worth pricing.
3. Then test the linkages: does being strong in segment A make segment B harder to attack? Independent moats add; interlocking moats multiply.
4. Name the source of protection for each — network effects, regulation, proprietary data — because different protections fail to different threats.
5. Ask what a competitor would have to do to break in, and note whether they would need to succeed in several places simultaneously.

• Segments that share a single regulatory dependency — five moats with one common failure point is really one moat.
• Conglomerate discounts: a market that values five linked businesses as five unrelated ones may keep doing so.

3. Require all three valuation methods to agree before you buy, not just before you look

1. Run the same three tests on every candidate: forward multiple versus its own ten-year average, an earnings growth model, and a reverse DCF at a 10% required return.
2. Record each as pass or fail independently, before forming a view.
3. Buy only when all three pass. Two of three is a follow-list entry, not a purchase.
4. Compare the reverse DCF's implied growth rate directly with the company's own historical rate — the sentence you want to be able to write is "it should be able to grow more than that."
5. Keep the required return fixed across all names so the tests remain comparable over time.

• A historical average multiple inflated by a period of higher growth — a cheap-versus-history reading can be a re-rating that already happened correctly.
• Unanimity produced by shared inputs; if all three lean on the same growth estimate, they are one test wearing three hats.

4. When there is no founder stake, read the compensation structure instead

1. Look for insider ownership first — a large personal stake is the strongest form of alignment.
2. Where none exists (as at most large-cap incumbents), go to the proxy statement and establish what fraction of the CEO's pay is at risk.
3. Check what it is at risk against: long-term earnings or returns on capital align with owners; revenue, share price over one year, or "strategic objectives" do not.
4. Treat a high performance-linked percentage tied to a long-term operating metric as a partial substitute for ownership, and say so explicitly rather than skipping the question.

• EPS targets met through buybacks rather than through the business.
• Performance conditions reset after a bad year, which converts at-risk pay back into fixed pay.

5. Buy with a limit set marginally below the last price, and accept the risk of not being filled

1. Decide the position size in currency before you look at the price, so the order is sized by conviction rather than by the quote.
2. Set a limit slightly below the prevailing price rather than crossing the spread. You are declining to pay for immediacy.
3. Accept that a small proportion of orders will not fill. For a long-term holding, a missed fill costs far less than a habit of paying up.
4. Publish or record the order — size, limit, quantity — so execution can be audited later against the thesis.
5. Do not chase. If the price runs away, the name returns to the shortlist at the new price and is re-tested.

• Limits so tight they never fill in a rising market, turning discipline into paralysis.
• Position sizes set by what is left in the account rather than by conviction.

6. Close the loop between the shortlist and the order, and measure the lag

1. Timestamp the moment a name enters the ranked shortlist and the moment it is bought.
2. Keep the lag short enough that the analysis is still current, and long enough that the decision is not reactive.
3. Re-check the price against the case at the moment of purchase — the shortlist was written at a different price.
4. Track how many shortlist names are eventually bought. A pipeline with near-zero conversion is a content exercise; one with near-total conversion was never a filter.

• A published shortlist moving the price you then have to pay — a real cost for a widely-read newsletter.
• Reporting lag in the portfolio record: this purchase does not appear in the holdings tables of the 2 August update.