Overview
What it is
A ChatGPT prompt that takes raw experiment results and shapes them into Cabana's standard readout: hypothesis, metric movement, confidence, segments, and a recommended decision.
When to use it
After any A/B test or quasi-experiment, before sharing results with the team.
Who it’s for
Data Science, PMs, and growth-minded engineers.
Expected output
A consistent readout that makes the decision easy to make.
Setup instructions
- Open ChatGPT.
- Paste the prompt.
- Drop in results table and hypothesis.
Usage
Good inputs
- • Hypothesis
- • Primary + guardrail metrics
- • Sample sizes and durations
Example prompts
- Readout: shorter session reminder copy, primary metric is open rate, ran 14 days.
Expected outputs
- • A 1-page readout ending in ship / iterate / kill.
Tips & limitations
- • Always report guardrails even when the primary metric wins.
- • Note any peeking or unplanned variants explicitly.
Tags
#data
#experimentation
#readout