The Prose Test: Write a Real Decision Memo
Objective: Take a real, currently-undecided marketing decision and write it up as a full-sentence prose memo using the lesson's Prose Test, so any weak reasoning surfaces on the page instead of surviving into the actual decision.
You're the marketing lead at Klaviyo weighing whether to sponsor a major ecommerce conference at a six-figure price tag. Leadership wants a recommendation by Friday. You could walk in with a bullet-point slide, or you could write the reasoning out first and see if it survives the sentences.
Write a one-to-two page decision memo in full prose, no bullets, no fragment lists, about the conference sponsorship decision above, or a real decision from your own work if you'd rather use live stakes. Follow the lesson's Prose Test: state the plan, write the reasoning as complete sentences, and if it collapses into 'and then a miracle happens,' rewrite until it doesn't, or let the memo change your recommendation.
Before you start
What you'll need
Free path (everything below is enough to finish)
Free tier is a blank page with headers, which is all this deliverable needs, the format constraint (prose, not bullets) matters far more than the tool.
No access? Any plain text editor, a Google Doc, or a notebook
Final deliverable
A one-to-two page prose memo (full sentences and paragraphs only) covering: the decision on the table, the reasoning for and against written out sentence to sentence, at least one place where writing it out changed or complicated your initial view, and a final one-paragraph recommendation.
See a reference example
Illustrative structure, using Mailchimp's decision on whether to raise prices for legacy free-tier users as the topic instead of the Klaviyo one above: the memo opens by stating the plan in one sentence, then spends a paragraph on the strongest case for it (revenue per free user is negative, the cohort has grown faster than paid conversion), then a paragraph testing that case against churn risk data, and explicitly reverses from 'raise prices across the board' to 'raise prices only for accounts inactive 90+ days' once the churn-risk paragraph exposed a gap the bullet version never would have shown. The closing paragraph states the narrower recommendation and names the one remaining unknown (whether inactive accounts are dormant or simply low-usage) as still untested.
Success criteria
You're done when you can:
- Written entirely in full sentences and paragraphs, zero bullet points or fragment lists anywhere in the reasoning section
- Names at least 3 explicit, distinct uncertainties or unknowns in direct language ('I don't know,' 'unclear,' 'untested'), not buried or implied
- At least one stated position is visibly reversed or narrowed partway through the memo, with the sentence that triggered the change identifiable
- The closing recommendation follows traceably from the preceding paragraphs, not a restatement of the opening plan with different words