Actionable insights — The Power Of Incentives

1. Translate every incentive into the behaviour it actually pays for

1. Take each metric someone is paid on and restate it from the earner's point of view — not what management intended, but what the fastest route to the payment is.
2. List the shortcuts that route creates. Paying per completed ride pays for speed and for skipping the clean-up.
3. Check the observable symptoms — complaints, sales mix, returns, churn — against that list before blaming culture or execution.
4. The rule to hold: "You will always get more of what you reward."

• A metric that looks like output but is really an input — completed rides, units shipped, calls closed — where quality is unpriced.

2. Score management pay on three specific tests before trusting the numbers

1. Open-market purchases: are insiders paid in cash and buying stock themselves, or being granted it? Granted shares dilute you; bought shares cost them.
2. Skin in the game: is the value of shares owned "significantly higher than the manager's annual salary"? Use the ratio to salary, not the percentage of the company, so it works at any size.
3. Long-term targets: are bonuses tied to multi-year measures like ROIC rather than quarterly earnings or the share price?
4. Read the proxy statement for all three before accepting a growth story — the pay scheme forecasts behaviour better than the strategy deck does.

• Insider ownership that is entirely granted equity — the percentage looks identical on a screen and means the opposite.

3. Treat short-term targets as a forecast of three specific distortions

1. When pay is tied to quarterly earnings, look for underinvestment first: R&D and maintenance cut to meet a number.
2. Then for pulled-forward revenue — channel stuffing, discounting into quarter-end — which shows up as next year's shortfall.
3. Then for buybacks executed at high prices to hit an EPS trigger rather than because the stock was cheap.
4. Each has a financial fingerprint: falling capex-to-depreciation, rising receivables, and buyback prices well above the year's average.

• Buybacks praised without a price reference — the same action is value creation or bonus management depending on the multiple it is executed at.

4. Use a public employee-review score as a culture proxy, with a stated threshold

1. Accept that culture cannot be quantified, then find an indicator rather than a measurement.
2. Use anonymous employee reviews and set a threshold in advance — "a good Glassdoor rating is usually above 3.5 out of 5" — so the test is not retrofitted.
3. Read the CEO-approval percentage alongside the overall score; they can diverge and the divergence is informative.
4. Treat it as a screen for the obviously broken rather than as evidence of excellence.

• Self-selected samples and small review counts — a 3.7 from forty reviews at a multinational tells you very little.

5. Run the qualitative work first, and let it still lose to the price

1. Do the ownership, tenure, culture and capital-allocation work before opening the valuation.
2. Score it honestly — insider ownership percentage, CEO tenure, ROIC through that tenure, employee ratings, EPS growth over a decade.
3. Then apply the valuation gate, and be willing to decline a company that passed every earlier test.
4. Keep the write-up. It is reusable the moment the price moves, which is the whole point of doing it before you need it.

• Qualitative excellence quietly becoming a reason to overpay — a great owner-operator at the wrong multiple is still the wrong purchase.

6. Name your own structural biases and check the position against them

1. Identify the biases in your own book — home country, sector familiarity, currency — and state them plainly.
2. Check whether a candidate is attractive on its merits or because it is familiar.
3. Keep the familiar name on a watchlist rather than banning it; the bias is in the reasoning, not in the company.

• Over-correction — refusing a good business because it is local is the same error with the sign flipped.