📊 Data & Analysisadvanceda-b-testingstatisticsexperimentationdata-analysis

Analyze A/B Test Results

Interpret A/B test results correctly — statistical significance, practical significance, and decision recommendation.

The Prompt

prompt.txt
Analyze the following A/B test results. Provide:
1. Statistical significance: p-value calculation and interpretation
2. Confidence interval for the true effect size
3. Practical significance: is the effect size large enough to matter?
4. Sample size adequacy: was the test run long enough?
5. Segment analysis: did the effect differ by user segment or device?
6. Decision recommendation: ship, iterate, or discard — with reasoning
7. What to test next

Test results:
- Control conversion rate: [X%]
- Variant conversion rate: [Y%]
- Control sample size: [N]
- Variant sample size: [N]
- Test duration: [DAYS]
- Primary metric: [METRIC NAME]
- Any segment splits available: [LIST]
- Confidence level target: [90% / 95% / 99%]

Example Output

With control at 3.2% and variant at 3.8% (n=12,000 each, 14 days), the lift is +18.75% (relative). p-value = 0.021 (statistically significant at 95%). However, confidence interval for true effect is +2% to +38% — wide range. Recommendation: ship with monitoring; re-run with larger sample before committing major engineering resources.

FAQ

Which AI model is best for Analyze A/B Test Results?

Claude Sonnet 4 or GPT-4o — both handle statistical interpretation well. Specify your confidence level.

How do I use the Analyze A/B Test Results prompt?

Copy the prompt, replace the [BRACKETED] placeholders with your specific information, and paste into your preferred AI assistant (ChatGPT, Claude, Gemini, etc.). With control at 3.2% and variant at 3.8% (n=12,000 each, 14 days), the lift is +18.75% (relative). p-value = 0.021 (statistically significant at 95%). However, confidence interval for true effect is +2% to +38% — wide range. Recommendation: ship with monitoring; re-run with larger sample before committing major engineering resources.