Customer Portal
New
Smart Upsell A/B Testing lets you run two versions of your upsell carousel against each other and see which one earns more from the same shoppers.
Most upsell tweaks get judged by squinting at next month’s revenue, where a new badge, a reordered rule, and a holiday weekend all land in one number and nobody can say which did the work. Now you start a test from your Smart Upsell settings, duplicate your current setup into Variant B, and change whatever you want to prove, from the products a rule prioritizes and rule order to badge copy, tag colors, and targeting. Shoppers split 50/50, and each one stays on the same variant every time they come back, so a returning customer never watches the carousel reshuffle mid-test. You pick the winning metric (revenue per visitor by default, or attach rate, conversion rate, or AOV), and revenue per visitor counts only upsell spend divided by everyone who saw that variant, so a carousel selling one pricey add-on to a handful of people can’t fake a win. Set it to end on a date or at 90, 95, or 99 percent confidence, and Skio names the variant that’s ahead and how confident it is. It won’t promote the winner for you, that call stays yours.
What you get is a real answer before you roll out a change, instead of learning a quarter later whether your new carousel added revenue or quietly gave some back.
