Borrowed Reasoning or Derived Truth? A Growth Memo Teardown
Objective: Read a realistic internal growth memo and separate the claims that are genuinely derived from first principles from the claims that are just analogy wearing a business-casual outfit, using the lesson's own name-it/test-it/rebuild-it method as the answer key.
You've been brought in to advise Allbirds' growth team for a quarter. A memo lands in your inbox from the growth lead proposing a large budget shift, six numbered claims, one recommendation, decided by end of week. Nobody has time to relitigate it from scratch, but somebody needs to check the reasoning before the budget moves.
Read the memo below. For each of the six numbered claims, decide whether it's a base truth someone actually tested (keep it) or an assumption borrowed from somewhere else and dressed up as a reason (flag it). Three of the six are borrowed reasoning. The other three look similarly confident but are genuinely derived from the team's own first-party data, don't flag those just because they sit next to the bad ones.
Before you start
What you'll need
Free path (everything below is enough to finish)
Free tier is enough for a single page of annotated reasoning, any plain-text editor works identically, the tool doesn't matter here, the thinking does.
No access? Any plain text editor or even a printed copy with a pen
The process
Specimens to review
Which of the six numbered claims are analogy-based assumptions (copied from a competitor, a trend, or someone else's result) rather than base truths this team actually tested on its own numbers? Name each one and say what it's really resting on.
INTERNAL STRATEGY MEMO — Q3 Growth Plan To: Marketing Leadership | From: Growth Lead We should shift 40% of our paid budget from Google Search to Instagram Reels this quarter. Claim 1: Our closest competitor moved budget to Reels in Q1 and their follower count doubled, so we should do the same to keep pace. Claim 2: Our own funnel data shows our best landing page converts video-referred traffic at 4.1%, versus 1.8% for search-referred traffic over the last two quarters, so shifting spend toward video-native channels should raise blended conversion rate. Claim 3: Reels is where "the algorithm rewards creators right now," and every growth newsletter is saying the same thing this month, so the timing is right. Claim 4: Our support-ticket data shows 61% of new customers this quarter cite "saw a product demo video" as their discovery source, up from 23% a year ago, so our buyers are already telling us where purchase intent is forming. Claim 5: Search CPCs in our category rose 34% year-over-year per our own account data, while last month's small Reels pilot produced a 22% lower cost-per-acquisition on the identical offer, so the unit-economics case is already showing up in our own numbers, not just in theory. Claim 6: A well-known DTC skincare brand built its entire growth engine on Reels, and if it worked for them, it will work for us. Recommendation: approve the 40% shift immediately.
Specimen: synthetic, realistic
Final deliverable
A claim-by-claim verdict (keep / flag) for all six numbered claims, each with one sentence saying what it's really resting on, plus a one-sentence overall call on whether you'd approve the budget shift as written.
See a reference example
Worked example using a different memo (not the Allbirds one above), on Warby Parker's well-known internal debate over glasses pricing: the industry-standard reasoning was 'frames cost $500 because that's what LensCrafters charges,' which is Claim-1-style analogy, a competitor's price copied wholesale with no look at what frames actually cost to source. The team instead priced from raw materials, licensing, and manufacturing costs, the same three-step move this teardown is grading: name the assumption ('$500 is the going rate'), test it against actual input costs, rebuild the number from what survived. A passing verdict on this teardown reads the same way, claim by claim, not 'these all sound confident so I'll flag half of them.'Success criteria
You're done when you can:
- Correctly flags all three analogy-based claims (1, 3, 6) as borrowed reasoning, not base truths
- Correctly leaves all three first-party-data claims (2, 4, 5) unflagged, doesn't over-flag just because they sit next to bad reasoning
- Each flagged claim's stated reason names what it's actually resting on (a competitor's result, a trend, someone else's case study), not just 'this seems off'
- Final one-sentence call is a genuine verdict on the recommendation, not a restatement of the memo's own conclusion