AppBloat’s Shopify rankings measure scripts, not slowdown

AppBloat’s new leaderboard offers a starting point for a Shopify storefront script audit. It ranks 100 apps and services using a September HTTP Archive crawl covering 729,553 detected Shopify stores. Its main measure is the median script response size attributed to each app on stores where that app appears.

That makes it useful for deciding what to investigate. It does not tell you how much an app slows your particular store, what revenue it generates, or whether removing it improves the buying experience.

What the figures actually cover

The published methodology specifies the 1 September 2026 mobile crawl and home pages only. It identifies Shopify stores through Wappalyzer detections, matches script request URLs to apps, then adds response size and request count for each app on each store. Apps need to appear on at least 20 stores to enter the ranking.

At the time of review, Privy appears first at 849 KB median script weight, across 10,959 detected stores. That is AppBloat’s reported observation of this sample, not a fresh measurement of every Privy installation or a performance test conducted by The Founder’s Inn. The full 729,553 store figure is the crawl’s Shopify population, not the sample size for each app.

The leaderboard also contains services such as Google Tag Manager and Facebook. Merchants should therefore read it as a view of detected storefront technologies rather than a definitive inventory of installed Shopify App Store apps.

Where attribution can miss things

AppBloat matches domains and Shopify theme app extension paths. It acknowledges that self hosting, bundling, tag managers and shared scripts complicate attribution. Some entries carry a marker explaining that only the app’s own script domain is counted and its Shopify extension scripts are not yet included.

A low number can therefore reflect incomplete capture. A high number still needs context: when the code loads, what it does and how the store is configured all matter. Home page observations also leave product pages, carts and later interactions outside this particular comparison.

The site’s corrections page provides a route for reporting attribution errors. It showed no listed correction entries when reviewed. We checked the published method and displayed figures but did not reproduce the underlying dataset query.

Turn the ranking into a store specific check

Start with the services your store actually uses. Record the relevant page, device conditions and features before changing anything. Compare repeat measurements under the same conditions, then test one change at a time in an appropriate test environment.

Check function as well as speed: a lighter page that loses a required consent control, breaks search or removes useful support is not automatically an improvement. For a suspected heavy script, ask the provider about conditional loading and configuration before deciding whether to replace the app.

Use the leaderboard to prioritise that investigation. Keep its September snapshot separate from your current store measurements, and avoid treating kilobytes as a direct estimate of sales lost.

AppBloat’s leaderboard announcement, 19 October 2026

Read AppBloat’s original announcement on X if the embed does not load.