What Radar can tell you

Radar compares public storefront captures made under consistent conditions. Repeated theme or price variations across independent sessions can indicate an experiment. Content and tool changes describe differences between captured pages.

Observations, with limits

A Certain label means the observation meets our corroboration rules. It does not prove random assignment, reveal conversion rates, or establish a winning variant. Geography, personalization, market settings and ordinary deployments can also change a storefront. When one observed variant persists, we describe that persistence without claiming it improved results.

Public statistics

Public history excludes the newest three days and captures older than thirty days. Statistics use that same window and current public membership. Test-opening counts count unique observed lifecycle openings. Tool popularity counts stores with the tool on their latest eligible pages. Tested-page-type figures count stores with an opening on that type of page. Category breakdowns use curated categories.

The sample is a curated index, not a representative census of Shopify. Stores with missing supporting evidence are omitted. A quiet capture does not prove that a store is not testing. Old or moderated observations never become evidence of an active test.

Respect for store owners

Radar reads public pages, respects crawl restrictions and excludes Convertify customer stores. We are not affiliated with scanned brands. Store owners and rights holders can request exclusion or screenshot removal through our takedown process.