We audited 1,302 Shopify stores.

82.0% fail Google's mobile speed bar.

The median store scores 33/100 on Lighthouse mobile. Just 0.4% land in Google's "good" band (90+). Speed is a confirmed ranking and conversion factor — most stores are leaving both on the table.

Median 15.0s to paint the main content.

97.1% of stores have a "poor" Largest Contentful Paint (over 4s) in the lab; only 1.4% hit Google's 2.5s target. But layout stability is mostly fine — 84.3% pass Cumulative Layout Shift. The bottleneck isn't shifting content; it's heavy themes, render-blocking apps, and unoptimized hero imagery delaying the first real paint.

These are Lighthouse mobile lab figures (a throttled mid-tier device) — they run harsher than real-world field data, so read them as relative, not as what every shopper experiences.

89.5% are invisible to AI shopping assistants.

When we asked Claude to name the top brands in each store's category, it recommended the store by name only 10.5% of the time. As shoppers increasingly ask AI "what should I buy," a brand that never comes up simply isn't in the consideration set.

This is a strict, name-level bar: a store counts as "visible" only if the model recommends it unprompted among category leaders. Smaller and newer stores expectedly fail — which is exactly the gap structured data and content can close.

Only 2.2% tell machines what they sell.

Just 2.2% of stores expose Product structured data (schema.org JSON-LD) in their server-rendered HTML, and 21.6% ship no JSON-LD at all. That markup is how Google builds rich product results and how AI assistants reliably parse price, availability, and reviews. Without it, a store is guessing-fodder.

Measured from the initial HTML response; schema injected later by client-side JavaScript isn't counted (and is itself less reliable for crawlers).

Every store surfaced an issue — 99.6% critical or high.

Our diagnostic surfaces the single highest-impact conversion problem per store, and it found one on all 1,302. Of those, 54.3% were critical and 45.3% high severity; only 0.4% topped out at medium. Broken or missing meta descriptions, sprawling unfiltered catalogs, absent trust signals, and policy/checkout gaps recurred most.