Data source. Data comes from the HTTP Archive monthly crawl (2026-09-01, mobile client, home pages only). Shopify stores are those detected as Shopify via Wappalyzer technologies in httparchive.crawl.pages (729,553 stores in this crawl).
Metrics. For each store we take every script request in httparchive.crawl.requests, match its URL to an app, and sum the response size (summary.respSize) and request count per app per store. We report the median (p50) and p75 across stores where the app appears. Apps seen on fewer than 20 stores are not ranked. Where a Lighthouse lab figure is shown, it comes from Lighthouse bootup-time and third-party-summary in HTTP Archive's 10k-page sample, and only when at least 20 Shopify stores in that sample load the app.
Attribution. Scripts are matched to apps by domain and by Shopify theme app extension paths (cdn.shopify.com/extensions/<id>/<handle>-<version>/). Domain lists partly come from third-party-web (MIT).
Caveats. Attribution is imperfect: scripts can be self-hosted, bundled, loaded via tag managers, or shared between apps; lab runs are single loads on emulated mobile and vary run to run; correlation with store speed is not causation. Stores using more apps differ in other ways too.
Field context. Real-user context comes from the Chrome UX Report (CrUX) by Google, used under CC BY 4.0.