Whenever you're ready, walk me through how you'd approach this.
What do you do if your experiment results are inconclusive in a data-driven project for a company?
Explain how you would distinguish between no meaningful effect, insufficient power, poor instrumentation, and an invalid experiment. Address the primary metric, guardrails, MDE, sample-size assumptions, randomization unit, statistical analysis, and decision criteria. State how you would handle peeking, sample ratio mismatch, and inconclusive subgroup results. Use illustrative assumptions only where numerical calculations are required.