The exec readout
The one-pager and the ten-minute meeting: pre-empting the hard questions, handling pushback with the numbers log, and landing the ask.
The deliverable that decides your investigation's fate is small: a one-pager (or 3–5 slides) and ten minutes of a busy person's attention. The narrative structure from last lesson gives you the skeleton. What separates readouts that land from readouts that stall is preparation for the part you don't control: the questions.
Pre-empt the meeting
Here is my one-pager [paste]. Simulate the readout Q&A three times: once as the COO (cares: cost, customer trust, speed), once as the CFO (cares: is the number real, what's the ROI of the fix), once as the procurement lead who CHOSE the eco mailers (cares: not being blamed; will look for holes). For each: the 3 hardest questions they'd ask, and whether my one-pager or appendix already answers them. Flag any question I currently cannot answer.
The procurement simulation is the one analysts skip and shouldn't: someone in most readouts owns the thing your finding implicates. Anticipating their (legitimate!) scrutiny is how you keep the meeting about the fix instead of the blame.
- Answer pushback from the log, not from memory. 'Where's 19% from?' → exact source, filter, date, in one breath (your Module 2 numbers log). The speed of the answer builds more trust than the answer.
- Concede what you haven't proven — immediately. 'Correct: the fragile-flag audit is the confirming link, it's pending, and the recommendation is sequenced so step one is that audit.' Conceding fast reads as command of the material; defending reads as attachment to a theory.
- Land the ask before the time is gone. Decisions default to 'good discussion, let's follow up.' Prevent it: 'What I need today is a yes on the audit and the packaging exception — the analysis can be wrong about the mailers and both steps are still cheap and reversible.' Note what that sentence does: it makes the decision robust to your own uncertainty.
- Frame around what they optimize. The same finding is 'reversible $X/month leak' to a CFO and 'customers receiving broken kettles' to a COO who said returns are a trust promise. Same numbers, same honesty — sequenced for the listener. AI drafts these re-framings well; you pick.
A readout argues for one decision once; a dashboard monitors a metric forever. Don't ship a dashboard as your answer ('here's a refunds dashboard!') when the sponsor asked a question — and when you do spec the follow-up dashboard (your requirements doc from Module 1 already sketched it), its metrics should come from governed definitions, not hand-rolled formulas per tile. That's the semantic-layer discipline from Data Foundations; it's what keeps next quarter's dashboard from restarting this quarter's definition fight.