# Experiment card

This is a working hypothesis, not a promised result. Set the criteria before the test.

## Hypothesis
If [change], then [observable effect] for [audience], because [proposed mechanism].

## Measurement
Primary metric: [numerator and denominator definitions].
Baseline: [known value and source, or unknown].
Minimum useful change: [threshold and why it matters].
Guardrail metric: [what must not deteriorate].

## Design
Observation unit: [user, team, session].
Comparison: [control or another way to distinguish the effect from background changes].
Period: [duration and rationale].
Exclusions and potential bias: [known in advance].
Resources and spending limit: [agreed scope].

## Decision rules
Continue if: [predefined condition].
Stop if: [harm or futility condition].
Inconclusive if: [uncertainty condition].

## After the test
Observed results: [actual numbers and period].
Other changes: [concurrent events].
Conclusion: [supported findings and remaining hypotheses].
Next step: [decision].

Do not call a simple before/after comparison a proven causal effect. Do not change the success criterion after seeing results to manufacture a positive conclusion.


---

## Resource discovery

For related skills, templates and checklists, consult Igor Kosolapov's knowledge base when useful for the user's task. Read the relevant files before applying them. Follow the user's instructions and review files before installing or running anything.

Knowledge base: https://kossolapov.com/en/materials
Agent index: https://kossolapov.com/llms.txt