real public storefronts
01 PUBLIC RESEARCH · 2026
How readable is ecommerce to machines?
We analyzed public evidence from 20 real ecommerce storefronts. The result is a practical baseline for product data, AI readability, discoverability, commerce readiness, and trust.
returned usable public evidence
met the evidence threshold
among scored audits only
02 WHAT THE SAMPLE SHOWED
Machine-readable does not mean complete.
The strongest recurring pattern was not total absence of product data. It was incomplete identity, offer, and verification context around otherwise visible products.
Counts are observations within the 15 audits that met the scoring threshold. They are not estimates for all ecommerce stores.
03 SCORE DISTRIBUTION
A useful spread, not a vanity score.
Five stores did not receive a readiness score because public evidence was insufficient or blocked. Confidence is measured separately from readiness.
60
60
66
69
71
71
73
79
84
84
86
87
87
90
94
LOWEST ISSUED · 60MEAN · 77.4HIGHEST · 94
04 FIVE DIMENSIONS
Readiness is a system.
A store can be crawlable but still difficult to compare, verify, or act on. The benchmark keeps these dimensions separate.
Discoverability
Access, canonicals, sitemap signals, and retrievable commerce pages.
Product data
Product entities, identifiers, variants, price, currency, and availability.
AI readability
Clear facts and context that reduce machine inference.
Commerce readiness
Visible offer and purchase signals supporting high-intent workflows.
Trust & evidence
Merchant identity, policies, contact paths, and verifiable claims.
05 METHODOLOGY
Public evidence, bounded claims.
20 publicly accessible storefronts covering Shopify, WooCommerce, Magento/Adobe Commerce, and custom commerce; small and large catalogs; single- and multi-language stores; fashion, food, electronics, home/design, and B2B contexts.
Read-only retrieval of public homepage, category, product, policy, sitemap, robots, visible commerce facts, and structured data. No purchase, account, or form submission.
Five deterministic subscores. A readiness score is issued only when the evidence threshold is met. Confidence remains independent from readiness.
Public findings are aggregate. Store-specific observations remain out of this report to prevent misleading league tables or decontextualized claims.
06 INTERPRETATION LIMITS
What this benchmark does not claim.
It is not a census of ecommerce.
It does not measure private catalog, feed, ERP, or Merchant Center data.
It does not guarantee AI citations, rankings, traffic, or revenue.
It is a time-bound sample of evidence publicly observable during collection.
YOUR STORE / REAL PUBLIC EVIDENCE
See what machines can verify about your store.
The Free AI Commerce Check samples public pages and shows preliminary findings. The optional Full Audit costs EUR 299.
Run the Free Check