Skip to content
Build75 minutesLesson 19 of 24

The one page a busy founder finishes

Your lead has ninety seconds. Everything you have learned in six chapters has to survive that.

Lead with the decision you want

Not background. Not methodology. The first sentence is what you want the reader to do, and everything after it is evidence for that ask.

One page holds roughly one problem, one recommendation, one metric and three pieces of evidence. If your thinking does not fit, that usually means the thinking is not finished, not that you need more pages.

And name what you are not doing. The scope cuts are the most credible part of any document, because they prove you made choices rather than listed wishes.

ProblemWho hurts, how much, and the evidenceKEEPSuccess metricOne number, plus its counter-metricKEEPScopeWhat ships, and three explicit cutsKEEPOpen questionsWhat you honestly do not know yetKEEPBackgroundThree paragraphs of history nobody readsCUTCompetitive landscapeA table copied from a deckCUTAppendixEverything you could not bear to deleteCUTFour sections fit on one page. The other three are why nobody read your last one.
Four sections fit on a page. The other three are why nobody read your last one.
Work through it

A one-pager is a decision document. A reader should understand the problem, the evidence, the proposed bet, and the request without needing you to narrate the missing connections.

Use your chosen course project throughout. The additional examples below are fictional practice cases; transfer the method to your own evidence.

Start with the decision

Put the recommendation and the decision needed near the top. Name the affected user and situation, then show the evidence that makes the problem worth addressing. Separate observed facts from assumptions. A long introduction about the market can distract from the specific choice your team needs to make this week.

Show the alternative

Explain the smallest credible alternative, including doing nothing for now. Compare the trade-off using the same outcome and constraints. This makes the proposal reviewable and prevents it from becoming a sales pitch for the first idea. A good reviewer can disagree with your recommendation while still understanding your reasoning.

Make the request executable

End with owner, scope, success measure, guardrail, and review date. State what remains unknown and how the next step reduces that uncertainty. A request for a one-day investigation is different from a request for a six-week build. The document should make that commitment unmistakable.

The visual field guide

A page that supports a choice

  1. Problem: User and situation
  2. Evidence: Facts plus uncertainty
  3. Proposal: Smallest useful bet
  4. Ask: Owner and review date
A page that supports a choice. Apply this sequence to your own project; it is a conceptual guide, not measured data.

How Amazon handled it

Amazon

Public commentary. No affiliation.
The situation

Amazon requires teams to write the press release for a product before building it, plus the FAQ customers and press would ask.

What they did

Working backwards from the announcement kills a large fraction of ideas at zero engineering cost, because a feature that produces a boring press release is usually a boring feature. Kindle, AWS and Prime all started as PR FAQs before a line of code existed.

Why it matters to you

Writing is not how you report the work. It is how you find out whether the work is any good.

Doing it with AI, and where it breaks

The move

Draft with a model, delete 60 percent, then ask which sentences here could be deleted without changing what a reader would do, and delete those too.

The trap

AI written product docs have a signature: they hedge, they restate the question before answering it, and they end sections with a summary nobody needs. That padding reads as thorough and is the opposite. Learn your model's tells.

Your AI workbench

Start with your own notes or clearly labelled practice data. Remove private details before sharing. Replace the placeholders, run the prompt in your chosen AI tool, and keep the output beside its source.

Prompt worth stealing

Review this one-pager as a busy decision owner [paste]. Identify the decision requested, evidence, assumptions, alternative, scope, owner, success measure, and review date. Mark absent fields. Suggest cuts and unresolved questions without inventing facts or silently rewriting the recommendation.

Before you use the output

  • The opening names an actual decision
  • Evidence and assumptions differ visibly
  • The final request includes owner and timing
Stuck? Try this next

Ask AI to summarise the decision in one sentence before editing. If that sentence is wrong, repair the document's structure rather than polishing its wording.

Keep a brief AI log: input used, useful output, what you checked, and what you rejected. The decision remains yours.

Build it

The artefact

The one pager: problem, evidence, recommendation, metric, scope cuts.

0/6

The worked solution

Stuck, or want to compare?

There is a worked solution. Try the build first. Reading it before you have attempted anything is the fastest way to learn nothing.

Checkpoint

If you did the build, these take two minutes. If you cannot answer one of them, that is the part to go back to.

Moving on marks this lesson complete. Finish the build first, it is the part that counts.

Previous lessonWhich one is worth a quarter?