The investment thesis checklist: from evidence to falsifiers

01Start with a falsifiable claim

An investable idea begins as a claim about the future. If the claim cannot be wrong, it cannot teach you anything. A useful claim names the business change you expect, the mechanism that should produce it, and the period in which evidence should appear.

Write the claim before collecting supporting material. That ordering makes it harder to quietly reshape the thesis around whichever facts look flattering later.

The one-sentence testWhat must become true, by when, for this thesis to work?

• One business outcome
• One causal mechanism
• One review window

02Separate evidence from expectations

Evidence describes what is observable now: reported results, filings, management statements, industry data, or a current market price. Expectations describe what investors appear to require next. Mixing the two turns a fact sheet into a conviction machine.

Give every material fact a source and an as-of date. Then label every forecast, interpretation, and scenario as your own. The separation makes disagreements productive because another reviewer can challenge either the source or the inference.

03State what the price appears to demand

A good company can still be a poor decision at the wrong expectations. Record the operating assumptions that appear embedded in the current price and compare them with a conservative range of outcomes.

Do not hide the bridge inside a single target price. Make the important variables visible: growth, margins, reinvestment, dilution, capital intensity, and the valuation multiple or discount rate used at the end.

04Choose falsifiers before the story gets personal

A falsifier is evidence that would weaken or break the thesis. It should be observable and specific enough to trigger review. “The story changed” is not a falsifier; a sustained loss of a named customer cohort, a missed regulatory milestone, or a balance-sheet threshold can be.

Pair each falsifier with a monitoring source and a decision response. Some signals require more research; others justify reducing risk or declining to act. The response should be written while you are still calm.

05Treat expression as a separate decision

The research conclusion and the instrument are different objects. Shares, defined-risk options, a watchlist, and no position expose you to different paths even when the underlying thesis is the same.

Compare structures only after setting a risk budget. In Thesis, any order path remains paper-only and requires an explicit broker review; the public research itself never submits an order.

06Record the decision and schedule the review

A thesis is most useful when it preserves what you knew and expected at decision time. Save the source set, scenarios, key disagreement, falsifiers, chosen action—including no position—and the next review trigger.

When the outcome arrives, judge the process separately from the return. A profitable decision can still have relied on weak evidence, and a well-structured decision can still lose. The journal exists to make that distinction inspectable.