How We Measure AI Shopping Visibility

We design realistic shopping scenarios that mirror how actual buyers ask AI assistants for product recommendations. For example: "What's the best harness for a small dog?" or "Recommend a durable laptop backpack under $100."

Queries are selected based on your product category, target market, and competitive landscape. We only evaluate your products against genuinely relevant scenarios — a small-dog harness brand won't be penalized for missing from large-dog equipment queries.

Each query is run across supported AI shopping interfaces under controlled conditions. We record the full response including product recommendations, ordering, and any relevant context.

We use repeated runs where necessary to assess result variability. AI recommendations are not static — the same query may produce different results across sessions.

We test platforms independently and never combine results from different interfaces without documented normalization.

Sponsored placements are identified and classified separately. Only organic recommendations count toward your AI Shelf Score. This ensures the benchmark reflects genuine product visibility, not advertising spend.

Each observed recommendation receives a weighted placement credit based on its position:

These weights are provisional and may be recalibrated based on benchmark distribution analysis.

The AI Shelf Score is calculated as:

The score ranges from 0 to 100. It represents your brand's observed organic recommendation visibility within the tested benchmark — not a universal measure of product quality or market share.

Thresholds are provisional design settings and may be adjusted based on actual benchmark distributions.

Your score is compared against directly comparable competitors tested using the same methodology, queries, and conditions. We call this your Category Rank.

Competitors are selected based on product category overlap and market relevance. We do not compare unrelated brands or inflate competitive sets.

AI Shelf Share represents your brand's proportion of normalized recommendation placement credits within the benchmark competitor set.

It does not represent actual market share or the percentage of AI shoppers exposed to your brand. It is a relative benchmark metric only.

Catalog Readiness is scored independently from AI Shelf Score. It evaluates the technical quality of your product data: identifiers (GTIN/UPC), descriptions, variant consistency, structured data, and merchant feed health.

Improving catalog readiness may support better AI product matching, but it does not guarantee a higher recommendation rank.

Every finding in your report is classified into one of three categories:

Every benchmark stores its complete methodology version, test parameters, prompts, timestamps, and observed results. This ensures reproducibility and allows meaningful comparison across measurement periods.

Historical trends are only shown when prior measurements exist and the methodology remains comparable.

AI Shelf Score measures performance within a defined, standardized benchmark. It does not capture every possible AI shopping interaction, all buyer intents, or all geographic variations.

AI recommendation systems are proprietary and dynamic. A missing recommendation does not necessarily indicate a technical problem, poor indexing, or a defective product. Scores may vary over time as AI systems evolve.

We do not claim to have access to all real user queries or actual platform-wide shopping exposure data.