Structured method for estimating the causal effect of a known intervention — "did X work," "what was the impact of Y," "how much lift did Z drive," "prove the campaign/feature/rebrand/price change moved the metric." Use this skill whenever a cause is named and its effect must be measured or defended, including A/B test analysis, pre/post comparisons, campaign measurement, incrementality questions, and any request containing causal verbs (drove, caused, resulted in, boosted, reduced) applied to an intervention. Do NOT use when the effect is known and the cause is unknown (use root-cause-analysis) or for open-ended pattern hunting (use exploratory-data-analysis).
