SecondPen.aiTemplates · Research & discovery
Experiment / A/B test brief template
Origin: Online controlled experiment practice (Ron Kohavi at Microsoft; Booking.com).
Sections
- Hypothesis
- Primary metric
- Guardrails
- MDE & sample size
- Duration
- Decision rule
How a tough reader scores it
When you score a draft against this format, the Critic evaluates this experiment brief. The weights add up to 100; the fixes for the weakest criteria arrive as cards in the editor.
| Criterion | Weight |
|---|---|
| Hypothesis is falsifiable and grounded in evidence | 20 |
| Single primary metric that reflects real value | 15 |
| Guardrails protect against harm | 15 |
| Sample size and MDE correctly computed | 20 |
| Decision rule pre-registered | 20 |
| Duration covers weekly cycles | 10 |
How it works in SecondPen.ai
Open the template and the editor writes the blank once, with a line of guidance under each heading that you delete as you fill it in. Write your abstract and the editor names the format it reads like. Ask for a document with a slash command and it offers to start from the matching format. When the draft is ready, the Critic scores it against the criteria above, the way this format's toughest reader would, and the three highest-leverage fixes become Editor cards.
More in Research & discovery
- User research planUX research practice (Nielsen Norman Group, Google UX)
- Customer interview guide (The Mom Test)Rob Fitzpatrick, The Mom Test (2013)
- Research readout / insights reportUX and market research practice
- Jobs-to-be-Done analysisClayton Christensen; switch interviews and Forces of Progress from Bob Moesta
- PersonaAlan Cooper, The Inmates Are Running the Asylum (1999)
- Customer journey mapService design practice (Nielsen Norman Group, IDEO)