Test what works. Keep what wins.

It depends on traffic and how big the difference between variants turns out to be, not a fixed number of days. We set a minimum sample size and conversion count per variant up front and don't call a winner — or make ranking or revenue promises — until that threshold and a reliable confidence level are actually reached, which usually lands somewhere between two and four weeks for most of the pages we test.

Whichever pages already carry the most traffic and intent — usually the primary landing page, a lead or checkout form, or a pricing page. We start from funnel and heatmap analysis to find where visitors are actually dropping off, then test that specific step, rather than running tests everywhere at once.

CRO doesn't touch your ad spend or targeting — it makes the traffic you're already paying for convert at a higher rate, which effectively lowers your cost per lead without spending more. It pairs naturally with our Google Ads or Meta Ads work, but doesn't require it — the same testing applies to organic or referral traffic too.