Sticky assignment. Each visitor gets a random, anonymous ID. Their variant is derived from a hash of that ID and stored in a cookie, so a visitor never switches variants.
Exposure. A visitor counts toward a test only when they actually view a page the test targets.
Conversions. A conversion is the primary-metric event (add to cart, checkout start or purchase) happening after the visitor's first exposure.
Conflicts. We block launching two tests on the same page type at the same time (or a sitewide test alongside any other), because overlapping tests contaminate each other.
Changes are CSS-selector operations: replace text or HTML, insert HTML before or after an element, add CSS, hide an element, or set an attribute. Use Preview to see the treatment on your storefront. Changed areas are outlined there, and preview visits are never tracked.
