Reference table
The one-page analysis plan for product research
Seven fields to fill in before fielding, so every claim was made before the data spoke
A seven-field analysis plan for product and insights teams, each field paired with the analytic choice it pins down before the data arrives. It turns the forks that produce p-hacked findings, such as which metric counts, when to stop and who to exclude, into commitments made in advance. Copy it into the front of any study brief. In Koji, the brief is written and versioned before recruiting starts, so it can hold this plan.
Where this appears
What's in this reference table
| Field | What it fixes |
|---|---|
| The decision | What action changes based on the result, and who owns it |
| Primary outcome | One metric, defined exactly, including the top-box rule and scale handling |
| Comparison family | Every comparison that can produce a reported finding, and how many there are |
| Sample size and stopping rule | Target n and the rule for ending the study, decided in advance |
| Exclusion rules | Speeder cut-off, attention-check policy, and what happens if exclusions change the result |
| Covariates | Which controls you apply, plus a promise to report the uncontrolled result too |
| Decision thresholds | What size of effect triggers which action, written before you know the answer |
Use this image
Free to use in your docs, decks and posts under CC BY 4.0, as long as you credit Koji with a link to this page. To embed it, paste this into your page:
html
<a href="https://www.koji.so/visuals/one-page-analysis-plan-for-product-research"><img src="https://www.koji.so/visuals/one-page-analysis-plan-for-product-research/one-page-analysis-plan-for-product-research.png" alt="Table of a one-page analysis plan with seven fields: decision, primary outcome, comparison family, stopping rule, exclusions, covariates and thresholds." width="1200" height="956" /></a>
<p>Source: <a href="https://www.koji.so/visuals/one-page-analysis-plan-for-product-research">The one-page analysis plan for product research</a>, Koji</p>
