Persuasion, measured.

CRO is the practice of systematically raising the share of visitors who buy — testing changes against a control instead of redesigning on instinct.

Test, don't redesign

CRO is not a redesign with better taste. It is the habit of changing one thing, splitting traffic between the old page and the new one, and keeping the change only when the numbers say it earned its place. Most tests lose — even at mature programs, roughly two out of three variants fail to beat the control. That is not a flaw in the method; it is the method. Every losing variant is a redesign you were spared from shipping.

Sample size is where the honesty lives. Detecting a small lift takes thousands of orders per variant, which most stores simply do not have. If you take a few hundred orders a month, skip the button colors and test big swings: a new offer, a rewritten product page, a shorter checkout. Big changes produce effects large enough to measure quickly — and when they lose, you have learned something worth knowing.

Where the wins hide

The winning tests repeat from store to store. Show shipping costs early — surprise fees at checkout are the single biggest driver of cart abandonment. Make pages faster, especially on phones, where most sessions now live. Cut form fields — every box a buyer must fill is a place to quit. None of this is glamorous, which is exactly why the gains are still sitting there. Start with your conversion rate as the baseline, pick the ugliest friction you can find, and test the fix against a control.