Skip to main content
Back to Data Analysis

Data Analysis · Doing It With AI

Capstone Audit a Full Analysis

The one idea: A confident deck can be wrong in every number that flatters its story. Audit the pipeline, not the prose — start from what one row means and whether the totals reconcile, then end with a verdict that separates what survives, what fails, and what the decision still hinges on.
⚠️ The trap: "Twelve clean slides, charts all pointing the same way, a confident recommendation — this checks out." That feels right and it's wrong. A coherent story is cheap and the errors tend to stack in its favor; audit the pipeline, not the prose — start from what one row means and whether known totals reconcile across every join and dedup step.

A deck lands in your inbox. Twelve slides, clean charts, a confident arc: "Q2 revenue grew 18% — growth driven by the new SMB segment — recommend doubling SMB acquisition spend." (SMB means small and medium business.) Leadership meets Thursday. Your job isn't to redo the analysis; it's to audit it — to run the checks this track has taught, in order, and decide whether the recommendation survives.

The specific numbers below are fictional. The failure modes are the ones that actually appear in real decks, in roughly this order of frequency. By the end, the checklist should run in your head automatically every time someone shows you a number.

The claimed analysis

The deck's key figures:

MetricQ1Q2Change
Revenue$2.10M$2.48M+18.1%
— Enterprise segment$1.60M$1.55M−3.1%

| — SMB segment (new) | $0.50M | $0.93M…

Keep reading: Capstone Audit a Full Analysis

Unlock the full Data Analysis track — every lesson, the AI tutor, and full mock assessments.

  • Full lesson content
  • AI tutor for this section
  • Practice questions