BRAC IT

Product Discovery Framework

Use when

  • A live product has enough traffic and instrumentation to compare alternatives.
  • The team needs evidence about which option performs better on a measurable outcome.

Do not use when

  • Traffic is too low or instrumentation is not trustworthy.
  • The team still does not understand the user problem or task well enough.

Best for

  • Optimization with live traffic
  • Comparing alternatives
  • Rollout validation

How to design it

  • Define one decision and one success metric clearly.
  • Control audience assignment and duration intentionally.
  • Avoid running experiments when the underlying experience is still fundamentally unclear.
Experiment Template

Use this when you need a practical starting template.

View template
  • What is the hypothesis?
  • Which audience sees each variant?
  • What metric decides the winner and what would count as inconclusive?

Output

  • Experiment plan
  • Variant performance summary
  • Decision and follow-up action