When to do CRO

CRO experiments require statistical significance, meaning enough visitors to detect differences. Below a certain scale, it's too slow to be worth doing.

Minimum traffic

10K unique visitors/month per tested page, roughly. Lower scale = longer tests = less learning.

When to prioritize

• High traffic, low conversion (CRO wins big)
• High-spend paid ads (every conversion matters)
• Stable product (testing isn't chasing a moving target)

When to delay

• Pre-product-market-fit (fix the product first)
• Very low traffic
• Radical site redesign planned anyway