Two Memos, Same Decision: Which One Actually Thought It Through?
Objective: Read two memos about the identical decision, one written as prose that exposes its own reasoning, one written as confident bullets that hide the gaps, and correctly identify which specific moves in each one demonstrate (or fake) the thinking the lesson describes.
Two members of Chewy's retention team independently wrote up the same proposal, a paid loyalty program, ahead of the same budget meeting. Only one memo actually shows its work. You're sitting in for the marketing director and have to decide which one to trust, and be able to say exactly why.
Read Memo A and Memo B below, both about the identical loyalty-program decision. Identify which one demonstrates writing-to-think and which one only performs certainty. For Memo B specifically, name which bullets are hiding an untested jump in reasoning, not just 'this whole memo feels thin.'
Before you start
What you'll need
Free path (everything below is enough to finish)
This is a reading-and-judging exercise, any note-taking tool works, the free tier is the whole requirement.
No access? Pen and paper, or annotate directly on a printed copy of the two memos
The process
Specimens to review
Which memo actually demonstrates the Prose Test, and where specifically does Memo A expose or reverse reasoning that Memo B never tests at all? Point to the specific bullets in Memo B that are hiding an unexamined jump, not the memo as a vague whole.
MEMO A (prose): "I think we should launch a paid loyalty program in Q3, but I'm not fully sure the math works. Our repeat-purchase rate is 31%, and I assumed a loyalty program would lift that to 40%+ the way it did for a comparable retailer, but when I actually laid the comparison out in writing, it fell apart: their average order value was $180, ours is $42, and their program had a $99 paid entry fee that filtered for already-loyal buyers, ours would be free enrollment. Rewriting the question, it isn't 'does a loyalty program work,' it's 'does a free-entry program work at our order size,' which I don't have an answer to. Given that, I'm reversing my initial recommendation: instead of a full launch, we should run a 90-day pilot with 5,000 customers and measure repeat-rate lift before committing the year's CRM budget to a full build." MEMO B (bullets): "- Launch loyalty program in Q3 - Comparable retailers see 40%+ repeat-rate lift - Program cost: $85K build + $12K/month - Target: all 210,000 active customers - Timeline: 6 weeks to launch - Recommendation: approve full build"
Specimen: synthetic, realistic
Final deliverable
A short written verdict naming which memo demonstrates writing-to-think, plus a bullet-by-bullet call on the three flagged claims in Memo B (what jump each one hides) and the three unflagged ones (why they're fine as bullets).
See a reference example
Illustrative verdict on a different pair of memos (not the Chewy loyalty-program pair above), using Duolingo's streak-retention messaging as the comparison case: a passing answer states plainly that the prose memo wins because it names a specific place ('rewriting the question exposed...') where the act of writing changed the conclusion, while the bullet memo's most confident-sounding claim ('comparable apps see 2x day-30 retention from streak mechanics') is exactly the kind of claim that only survives as a bullet, restate it as a sentence and it demands a citation the memo never provides. The point isn't finding more flaws in the bullet memo, it's finding the right three and leaving the concrete logistics bullets alone.Success criteria
You're done when you can:
- Correctly identifies Memo A as demonstrating writing-to-think and Memo B as performing certainty without testing it
- Correctly flags all three reasoning-hiding bullets in Memo B (the comparable-retailer lift claim, the full-rollout target, the recommendation with no visible chain), not just a vague 'Memo B is worse' verdict
- Correctly leaves the three factual/logistics bullets in Memo B unflagged (cost, timeline, topic statement), doesn't over-flag the whole memo indiscriminately
- Points to the specific sentence in Memo A where the written reasoning caused a reversal, not just 'Memo A seems more thoughtful'