Actionable insights — Best Buys: August 2026

1. Before buying a fallen compounder, decompose the decline into named, dateable causes

1. Establish first that the fundamentals fell, not just the price — falling revenue and profit is a different problem from a falling multiple.
2. List every distinct cause you can name, and label each one temporary or structural.
3. For each temporary cause, say what specifically reverses it and roughly when the comparison base stops being distorted.
4. Check whether the causes are independent or one event in three costumes — several unrelated headwinds landing together is the case for "temporary"; one cause described three ways is not.
5. Ask what would have to be true for the decline to be structural instead, and look for that evidence explicitly.

• Three causes that are really one — a pandemic demand spike, dealer overstocking and post-pandemic price rises are all downstream of the same event.
• No date. "Temporary" is only useful with a horizon; without one it is indistinguishable from hope.

2. Test a monopoly by asking what it would cost to rebuild it today

1. Ask literally: if a rival had unlimited capital, could they reproduce this asset? Price the land, the permits, the approvals and the time.
2. Separate the physical barrier from the regulatory one; regulatory barriers can be repealed, physical ones usually cannot.
3. Check the unit-cost advantage independently of the barrier — a protected asset that is also the cheapest way to do the job is protected twice.
4. Confirm the pricing power that should follow: has it actually raised prices at or above inflation across cycles?
5. Then separate the asset from the cycle it serves, and decide which one you are underwriting.

• A monopoly on a shrinking activity. Owning all the toll booths on a road nobody drives is worth nothing.
• Regulatory moats being regulatory risks in the other direction — rating agencies and sole-source suppliers with pricing power are recurring political targets.

3. Read a target price backwards and count how many of its inputs you supplied yourself

1. Write the target as an explicit chain: revenue × margin × multiple.
2. Label each link — company guidance, consensus, or your own assumption.
3. Flex the assumed links one at a time and see what survives. A conclusion that needs every optimistic input is not a valuation.
4. Attach a date. An unspecified "by then" turns any upside into an unfalsifiable claim, because compounding at an unknown rate over an unknown period is not a return.
5. Convert to an annualised return before comparing with anything else.

• A 25x exit multiple used as though it were neutral. It is the same "fair exit PE" that anchors the earnings-growth model across the whole Buy-Hold-Sell sheet.
• Management's own long-term targets imported as facts. They are marketing until delivered, and "long-term" is doing a lot of work.

4. Take candidates from the losers list, and check the pipeline actually runs

1. Rank your pre-vetted universe by recent performance and read the bottom of the list first.
2. Choose the spotlight from that bottom section, so the research effort is aimed where the price has already moved against consensus.
3. Track which spotlights later become purchases and how long the lag is — the conversion rate is the honest measure of whether the shortlist is a pipeline or content.
4. Watch the repeat offenders: a name appearing on the worst list two months running is either a compounding opportunity or a thesis quietly failing.

• The names that keep being screened and never bought. That residue is where the real filter lives.
• A universe narrow enough that the "worst performers" are simply the same handful of holdings each month.

5. Distinguish demand that is chosen from demand that is forced, and price the difference

1. For any distributor, supplier or service business, estimate what share of revenue is emergency or mandated rather than discretionary.
2. Ask what the buyer optimises for in that moment. If it is speed or availability rather than price, the supplier holds the pricing power.
3. Confirm the switching friction: certification, local inventory, workflow software, regulation.
4. Treat that share of revenue as quasi-recurring and underwrite it separately from the cyclical remainder.

• The mirror risk: forced demand is a political target precisely because it is forced.
• Emergency share estimated by the company itself and quoted as a range ("70%-80%") — worth checking against segment disclosure.

6. Notice when a shortlist gives you quality with no price, and refuse to complete it yourself

1. For every recommendation, ask what the entry price is. If none is given, the recommendation is a description of a business, not a decision.
2. Check whether the same source publishes prices elsewhere. If it does, the omission is a choice.
3. Go and get the number from the source's own other work rather than inventing one.
4. Only then decide, and record the price you used, so the call can be judged later.

• Business descriptions that read as buy cases. Every one of the five paragraphs would be equally true at twice the price.
• Your own tendency to supply the missing valuation charitably. If the writer did not do it, do it adversarially or not at all.

7. Separate a currency-driven slowdown from a demand-driven one before treating a de-rating as opportunity

1. Find the constant-currency growth rate alongside the reported one; the gap is translation, not trading.
2. Check whether the currency move is also hitting costs — a manufacturer producing in the same currency it reports in is affected very differently from an importer.
3. Ask whether the market has de-rated the multiple or only marked down the earnings; only the first is an opportunity.
4. Confirm the demand signal independently — for a scarcity brand, the resale market is the cleanest read.

• "Constant currency" as a permanent excuse. Currencies trend for years, and a euro reporter with global sales lives with it.
• A de-rating from very expensive to less expensive being mistaken for cheap — the previous week's sheet still had the reverse DCF demanding 18.4% growth against 11.9% expected.