Prove conversion ads value via incrementality

An e‑commerce advertiser runs conversion‑optimized ads and wants proof the ads cause incremental conversions (lift). Propose a rigorous plan: (1) choose among user‑level RCT holdouts, geo‑experiments with matched markets, or PSA/ghost‑bid approaches; justify assumptions and spillover risks; (2) define the primary metric (absolute/relative lift in conversion rate) and attribution windows; address view‑through bias, cross‑device, and post‑view latency; (3) perform sample‑size/power calculations to detect a 2% relative lift from a 3% baseline CR at α=0.05, stating formulas and assumptions; (4) outline analysis (difference‑in‑differences, CUPED, pre‑trend diagnostics), interference handling, budget/pacing constraints, and how to guard against p‑hacking; (5) compute incremental ROAS with confidence intervals and present results/actionability to the advertiser, including sensitivity analyses and repeatability criteria.