What it corrects
Some business claims can explain every possible outcome. Growth proves the strategy works; flat results mean the market needs more time; decline means the company did not invest enough. Nothing can count against the belief, so the team cannot learn from it.
Falsifiability asks for an observable result that would make the central claim less credible. It does not demand laboratory certainty. It makes the recommendation expose its load-bearing assumption, a useful horizon, and the action that follows if the assumption fails.
That single move raises the quality of evidence and uncertainty at once: the memo becomes specific enough to be wrong and practical enough to update.
How it works
- Write the central causal claim in one sentence: “If we do X, Y changes because Z.”
- Identify the observation that would contradict Z, not merely disappoint the team.
- Give the observation a threshold and a date or event horizon.
- Decide in advance whether the result triggers a stop, revision, or deeper test.
In a Business Case Weekly case
In the Nike case, “our own brand gives us control” is too elastic. A falsifiable version names what control should produce — for example, repeat demand at acceptable unit economics within a bounded pilot — and which result would favour renewed distribution terms instead.
In your answer
- “My recommendation depends on the claim that …”
- “I would be wrong if … by …”
- “If that signal appears, I would stop/revise …”
Common misuse
Do not choose a test that can fail only after the irreversible commitment is complete. Do not hide behind a metric the team can redefine later. A useful falsifier arrives while options remain and distinguishes the central claim from a plausible rival explanation. It should make changing course easier, not merely make the memo sound scientific.