A/B Test Results Interpreter

Reads raw A/B test numbers and returns a statistical significance verdict with a clear ship or kill call.

Best For

•Solo marketers running their own landing page tests who don't have a statistician to tell them if results are real.
•Growth-focused founders who want a fast significance check before committing engineering time to ship a variant.
•Freelance CRO consultants who need a repeatable framework to present test results to clients clearly.

Prompt Template

You are a data analyst interpreting A/B test results. I'm going to paste the raw results. Tell me whether the results are statistically significant and what to do next. Output: 1. Test Summary: what was tested, control vs. variant 2. Results Table: side-by-side metrics for control and variant 3. Statistical Significance: is the difference statistically significant at 95% confidence? Include confidence level, p-value estimate, sample size assessment 4. Practical Significance: even if statistically significant, is the effect size large enough to matter? 5. Verdict: SHIP IT / KILL IT / KEEP RUNNING / ITERATE 6. Caveats: anything that could invalidate these results 7. Next Test Suggestion: what to test next based on what you learned Test results: [PASTE RESULTS, include: sample sizes, conversion rates, confidence intervals if available, test duration]

Pro Tip

Always include the test duration and total sample size, not just the conversion rates, since a result can look dramatic but still be statistically meaningless with too few visitors or too short a run.

Example Output

A sample of what this prompt produces once you fill in the placeholders.

An analysis of a homepage CTA test finding a 33% relative CTR lift with 97% confidence, a SHIP IT verdict, and a caveat recommending one more week of data to rule out weekly traffic cycles before fully committing.

Recommended Tools

ChatGPT

Claude

Unbounce

How to use this prompt

Copy the prompt template using the button above.
Paste it into your preferred AI assistant (ChatGPT, Claude, Gemini, etc.).
Replace all bracketed placeholders like [TOPIC] with your specific details.
Send the prompt and refine the output as needed.