Actionable insights — 8 Stocks Quality Investors Are Buying

1. Correct every 13F list for the manager's size before you read it

1. Before treating an aggregate "most bought" list as a signal, ask what the buyer was allowed to buy: a manager of billions cannot take a meaningful position in a small company without moving its price.
2. Add the second constraint — benchmark risk. A fund measured against the S&P 500 must own the mega-caps that make up a third of it, or accept large tracking error.
3. Conclude that the mega-cap names on the list carry the least information: they are partly a constraint, not a conviction.
4. Invert to find your own edge: the positions you can take that they cannot are small caps, and the positions that do carry conviction are the small ones several funds hold — bought despite the size penalty.

• A mega-cap that appears on the aggregate buy list and the quality-fund trim list in the same quarter — that is two different constraints talking, not two opinions.

2. Read the sells before the buys — they carry the reasoning

1. Start the filing review with exits and full liquidations, not additions: a manager will explain a sale far more precisely than a purchase.
2. Pull the stated reason from the letter or interview rather than inferring one from the position change.
3. Classify the reason into one of three buckets: valuation, a broken moat, or a lost competitive position.
4. Apply the same test to your own holdings before you look for anything new — the sell reasons generalise across the sector faster than the buy ideas do.

• Sales with no published reason. Absent an explanation they are noise — redemptions, rebalancing, tax — and should not be copied.

3. Separate a temporary problem from a company falling behind

1. When a quality name derates, name the specific mechanism: a cyclical downturn, a one-off cost, a regulatory event — or a competitor that is now simply better.
2. Test the "temporary" claim against relative capability: is the company's technology, underwriting, product or cost position still at or above the peer set?
3. Buy the first category at a discount; sell the second regardless of the discount, because cheapness does not restore a lost capability.
4. Set a falsifier for the temporary reading — a metric that, if it keeps deteriorating relative to peers, converts the diagnosis to the second category.

• Your own reclassification drift: a thesis that keeps being re-described as "temporary" for successive quarters is usually the second category.

4. Treat a management guidance reset as primary evidence about the moat

1. When a new chief executive arrives, watch for the first long-term target they publish — it is the least-managed number you will get.
2. Compare it with the outgoing target. A large downward reset is management conceding a competitive fact before the financials show it.
3. Cross-check the reset against the competitive story you already suspect; agreement between the two is a stronger signal than either alone.
4. Act on the reset rather than waiting for it to appear in reported growth, which lags by a year or more.

• A reset framed as conservatism by a new manager clearing the decks — check whether the competitive evidence exists independently before believing it.

5. Price customer concentration as a certainty problem, not a size problem

1. For any business with one dominant customer, measure the share of revenue it represents and ask what could change that customer's behaviour.
2. Distinguish a customer that could shrink from a customer that could become politically volatile — the second destroys forecastability even when revenue holds.
3. Downgrade the name when the volatility is in the customer's decision process rather than in demand, because a long-term owner cannot underwrite it.

• The reverse case: a concentrated customer whose spending is legally mandated (defence floors, regulated utilities) is a different, more durable animal.

6. Weight a purchase most when the buyer is stepping outside their own style

1. Identify each tracked manager's declared style — deep value, quality, growth — before reading their filings.
2. Flag purchases that sit outside that style. A deep-value investor buying a compounder is claiming the price has fallen far enough to satisfy a value test.
3. Treat that as evidence about valuation specifically, then do your own valuation work to see whether you agree.
4. Ignore in-style purchases as confirmation — they tell you only that the manager did their usual job.

• Style drift by the manager rather than opportunity in the stock — check whether the outside-style purchases are one name or a pattern across the book.

7. Against an AI-disruption derating, test whether the asset is a network rather than a workflow

1. Ask what the AI would actually replace: a piece of software work, or the coordination of two sides of a market.
2. Software work is replaceable. A network is not — the model still has to transact against whoever holds the supply.
3. Confirm by asking where an AI agent would source the inventory or the counterparty; if the answer is the company being derated, the fear is misdirected.
4. Buy when the multiple has compressed to a multi-year low on a fear that fails this test.

• Supply-side defection: hotel groups or airlines that succeed in pulling bookings direct would attack the network itself, which the AI narrative would then merely accelerate.