Product strategy and bets

Use this when the user names a direction, a bet, a vision, or a strategy doc, or asks whether to build something at all. The output is a strategy doc that drives one decision, not a document that ends in observations.

What you need first

The bet or direction, who the user is, what is known so far, and the decision the doc has to drive. No numbers means you size with bands and mark every assumption. If the user cannot name the user, say so and do not write the doc yet.

Write it

1. Say back the bet in one line: who wins, what changes, and why now.
2. Write the narrative in three whys — why change, why now, why us. Each why gets one paragraph grounded in something the user gave you or a source you can name. No sourced line, no paragraph.
3. Grade the evidence both ways. For is required, against is required. Mark each item A through E: A is payment or repeated usage, B is first-party behavior or real pricing, C is reviews or case studies, D is demand signals or reports, E is assertion. Never call a D or an E validation.
4. Name the substitutes, including spreadsheets, manual work, and doing nothing. A strategy that cannot beat doing nothing is not a strategy.
5. Build the business case: the cost of the bet, the return if it works, and the kill line where you stop. Bands over point estimates, and every number labelled FACT, INFERENCE, or UNKNOWN.
6. Close with the ask: the investment, the decision, and the date it is needed. A strategy doc that asks for nothing gets read by nobody.

Output

The dated strategy doc: narrative, graded evidence for and against, substitutes, business case with the kill line, and the ask. Take one round of edits, then save it. Offer to turn the first bet into roadmap items — only on the owner's yes.

When the evidence is thin

Thin evidence means a smaller bet and a louder UNKNOWN, not a confident doc. Name the one piece of evidence that would move a claim up a grade, and who would get it. Never invent a market size, a customer, or a win, and never frame an E-grade belief as evidence.