agentleFS
product-on-purposeproduct-on-purpose/pm-skills/skills/measure-experiment-results/SKILL.md

measure-experiment-results

Skill777 starsChanged 10 months agoApache-2.0View on GitHub

Documents the results of a completed experiment or A/B test with statistical analysis, learnings, and recommendations. Use after experiments conclude to communicate findings, inform decisions, and build organizational knowledge.

The file

We split this skill into the rules it gives an agent

5 rules. About 0 are common, 4 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈4

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    After an A/B test or experiment reaches statistical significance

    line 20
  2. Distinctive:

    When an experiment is ended early (for any reason)

    line 21
  3. Distinctive:

    When asked to document experiment results, follow these steps:

    line 35
  4. Distinctive:

    Look for differential effects across user segments (platform, tenure, plan type, etc.).

    line 50
  5. Very specific:

    Use the template in references/TEMPLATE.md to structure the output.

    line 63

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md
<!-- PM-Skills | https://github.com/product-on-purpose/pm-skills | Apache 2.0 -->

Experiment Results

An experiment results document captures what happened when you tested a hypothesis, including statistical outcomes, segment analysis, learnings, and clear recommendations. Good results documentation turns individual experiments into organizational knowledge that improves future decision-making.

When to Use

  • After an A/B test or experiment reaches statistical significance
  • When an experiment is ended early (for any reason)
  • To communicate findings to stakeholders who weren't involved
  • During decision-making about whether to ship, iterate, or kill a feature
  • To build a repository of learnings that inform future experiments

When NOT to Use

  • The experiment is not designed or run yet -> use measure-experiment-design
  • The results demand a direction decision -> use iterate-pivot-decision; this skill reports the evidence, that one decides
  • You want the transferable learning banked for the organization -> follow up with iterate-lessons-log
  • Your data is survey responses, not a controlled experiment -> use measure-survey-analysis

Instructions

When asked to document experiment results, follow these steps:

  1. Summarize the Experiment Provide context: what was tested, when it ran, how much traffic it received. Link to the original experiment design document if one exists.
  2. Restate the Hypothesis Remind readers what you believed would happen and why. This frames the results interpretation.
  3. Present Primary Results Show the primary metric outcome clearly: what were the values for control and treatment? Include statistical significance (p-value), confidence intervals, and sample sizes. Be honest about whether results are conclusive.
  4. Analyze Secondary Metrics Present guardrail metrics that ensure you didn't cause unintended harm. Note any secondary metrics that moved unexpectedly.both positive and negative.
  5. Segment the Data Look for differential effects across user segments (platform, tenure, plan type, etc.). Sometimes overall results mask important segment-level insights.
  6. Extract Learnings What did you learn beyond the numbers? Include surprising findings, questions raised, and implications for the product hypothesis. Negative results are valuable learnings.
  7. Make a Recommendation Be clear: should we ship, iterate, or kill? Support the recommendation with the evidence. If the decision is nuanced, explain the trade-offs.
  8. Define Next Steps Specify what happens now.engineering work to ship, follow-up experiments, metrics to continue monitoring, or documentation to update.

Output Format

Use the template in references/TEMPLATE.md to structure the output. A complete readout fills every template section: Summary; Hypothesis Recap; Results; Segment Analysis; Visualization; Learnings; Recommendation; Next Steps; and Appendix.

Quality Checklist

Before finalizing, verify:

  • Statistical methods and significance are clearly stated
  • Confidence intervals are included (not just p-values)
  • Segment analysis checked for differential effects
  • Secondary/guardrail metrics are reported
  • Learnings go beyond just the numbers
  • Recommendation is clear and actionable
  • Negative or inconclusive results are reported honestly

Examples

See references/EXAMPLE.md for a completed example.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.