Rebuild One Assumption-Driven Decision From Scratch
Objective: Take one real, currently-assumption-driven marketing decision and run it through the lesson's three-step method for real, producing a short written reasoning chain that could actually be handed to a decision-maker.
You're advising Casper's growth team, and someone has already decided: 'we should run a heavy Black Friday discount because every mattress brand does.' The decision hasn't shipped yet. You have one shot to either confirm it with real reasoning or replace it with something better before the budget is committed.
Pick one assumption-driven marketing decision, either the Casper Black Friday example above, or a real decision from your own work or a company you know well. Run it through the three-step method from the lesson and write up the result as a short reasoning-chain document, not a slide, not bullets pretending to be reasoning, an actual chain someone could follow and disagree with on the merits.
Before you start
What you'll need
Free path (everything below is enough to finish)
Free tier handles a single structured document with headers for each step, no paid feature is needed for this deliverable.
No access? Any plain text editor, a Google Doc, or a notebook page photographed at the end
Paid upgrades (optional, faster/deeper)
The free path (a blank document and the three-step method) is the complete path. The AI devil's-advocate pass is a speed upgrade, not a requirement, the lesson was written and tested for decades before either tool existed.
Pasting your if-then chain and asking it to argue against each link is a fast way to catch a link you talked yourself into without really testing, useful but not required, the lesson's method works with a blank page and no AI at all.
Final deliverable
A written reasoning chain with three labeled sections: (1) the standard answer as an explicit if-then chain, (2) each link tested with a stated true/false/unclear verdict and why, (3) the rebuilt conclusion from whatever survived, plus one line on whether it matches or contradicts the standard answer.
See a reference example
Illustrative walkthrough, not the Casper decision, applied instead to a hypothetical Chewy assumption ('we should run paid social because every DTC pet brand does'): Step 1, the if-then chain is 'DTC pet brands succeed on paid social, therefore paid social will work for us, therefore we should budget for it.' Step 2, testing each link: 'DTC pet brands succeed on paid social' is true for some but conflates brand awareness with retention-driven brands like this one; 'therefore it will work for us' is unclear, no data cited on this specific audience's paid-social response; the chain does not survive intact. Step 3, rebuilt conclusion: given the retention-first model, the base truth actually in evidence is repeat-purchase behavior driven by service quality, not discovery channel, so the higher-leverage first-principles move is testing lifecycle email deepening before adding a new paid channel. Different answer than the standard playbook, which is exactly the point of running the exercise instead of skipping it.Success criteria
You're done when you can:
- The if-then chain in Step 1 has at least 3 distinct linked claims, not one vague assumption restated three times
- At least one link in Step 2 is marked false or unclear with a stated reason, not every link rubber-stamped true
- The Step 3 conclusion is derived from what specifically survived Step 2, traceable claim by claim, not just a restatement of the original decision with new words
- The closing line explicitly states whether the rebuilt conclusion matches or contradicts the standard answer, per the lesson's own framing of what counts as a genuine result either way