Measure

A/B testing

Randomly split people into a control and a variant, change one thing, and compare outcomes to see which version performs better.

How it's used

Half of visitors see the current checkout; half see one with the shipping cost shown earlier. After enough orders, the variant wins on completed purchases, not just clicks.

How I help

I help design tests that can actually reach a conclusion: the right metric, the right sample size and a plan for what happens after.

Common toolsSubject lines, landing pages, offers, checkout steps