Design a rigorous test plan to compare the proposed scoring model against the current baseline for [CompanyName]. Use historical backtesting (rolling windows) and a forward-looking A/B split across inbound channels in [Region]. Define sample size using [BaselineConversion] and desired detectable uplift of [MinUplift] during [Timeframe]. Assign treatment routing to [RoleSDR] with identical SLAs to isolate the scoring effect.
Output format:
Test Charter (hypotheses, units of randomisation, guardrails);
Sample Size & Power Table with calculations;
Success Metrics (SQL rate, opportunity creation, speed-to-first-touch);
Data Collection Spec and instrumentation checklist;
Executive Readout template with visualised lift and decision criteria.
Include pre-registration of analysis, contamination mitigation, and a rollback plan. Provide guidance on post-test calibration if lift is heterogeneous by segment or channel.