Testing velocity

Velocity is tests per week or per quarter. Higher velocity = more learning, faster improvement.

Velocity by maturity

• Early: 1-2 tests/month
• Growing: 3-5 tests/month
• Mature: 2-3 tests/week

Building velocity

• Documented test queue
• Templated test design
• Dedicated CRO infrastructure
• Regular planning cadence

Quality still matters

Bad tests at high velocity is waste. Balance speed and rigor.

What to do with this

• Target 3-5 tests per week at meaningful traffic, fewer means slow learning, more means poorly-designed tests
• Balance velocity with rigor, 10 noisy tests per week produces no actionable learning
• Use feature flags + testing tools that reduce friction, velocity is capped by how painful it is to ship a test
• Track winners per month, if win rate drops below 10%, the hypothesis quality is the problem, not the test count
• Document everything in a rolling doc, the velocity advantage is lost if learnings don't compound