Shopping is moving into assistants, and assistants read product content differently than people do. The Feedmind Score measures one thing: how well a catalog can be read, understood and answered from by an AI agent. We publish the method in full, because a score nobody can audit is a score nobody should trust.
Every dimension is a set of named checks. A product is scored on each dimension it is eligible for, and the dimensions are combined by the weights below. A store that publishes one language is scored over three dimensions rather than being penalised for the fourth.
| Dimension | Weight | What it asks |
|---|---|---|
| Agent accessibility | 25 percent | Can an assistant reach and read the page at all? We check crawler permissions, whether content survives without JavaScript, and whether the page states what it is before a human ever scrolls. |
| Structured data | 25 percent | Machine readable facts beat prose. Product and Offer markup with real identifiers is what lets an assistant match your item to the one a shopper is asking about, and quote a price it trusts. |
| Content answerability | 30 percent | The heaviest weight, because it decides recommendations. We measure whether your copy contains the specifics buyers ask about, such as fit, compatibility, materials and what is in the box, and whether those facts survive summarization. |
| Multi-language parity | 20 percent | European catalogs answer in several languages. A thin German page is a German shopper you lose silently, so we compare depth, attributes and markup across every locale you publish. |
Checks marked v1.1 were added with the scoring version of the same name. Every check that existed in v1.0 produces byte-identical results under the current version.
robots.txt: GPTBot, PerplexityBot, Google-Extended
Server rendered content without JS execution
A single, clear H1 and a declared document language
Image alt text coverage, how images read without eyes (v1.1)
schema.org Product and Offer JSON-LD validity
GTIN, MPN and brand coverage across SKUs
Price, currency and availability freshness
Standard taxonomy category assignment (v1.1)
Named spec coverage: material, capacity, warranty, compatibility, colour (v1.1)
Description depth and attribute density
FAQ content and FAQPage markup
Buyer question answer rate in simulation
Tag quality: use case and discovery phrasing (v1.1)
Word count and attribute parity per locale
hreflang correctness and locale routing
Localized structured data, not just copy
The store level bands below are what a catalog score and a public scorecard are graded against. Per product rows in the app use a more forgiving catalog scale, because a single thin product is not the same failure as a whole catalog that reads badly. Mixing them would flatter one of the two.
| Grade | Band | What it reads as |
|---|---|---|
| A | 85 to 100 | Agent-ready. Assistants can answer almost anything about the catalog. |
| B | 65 to 84 | Solid. Gaps appear on comparison and compatibility questions. |
| C | 50 to 64 | Readable but thin. Frequently skipped for better-documented rivals. |
| D | 35 to 49 | Structural problems. Most buyer questions go unanswered. |
| F | below 35 | Effectively invisible to assistants. |
The current version is FM-SCORE v1.1. When a check is added, removed or reweighted, the version increments and the change is listed on the changelog before it takes effect. Every score we store carries the version it was measured under, so your own history stays comparable even after the method moves.