Shopping Visibility
Track how your brand appears in AI-powered shopping recommendations, product comparisons, and purchasing advice.
Overview
Shopping Visibility tracks how your brand appears when users ask AI chatbots for shopping advice, product recommendations, and purchase comparisons. As AI increasingly influences buying decisions, understanding your brand's position in AI shopping responses is critical for e-commerce and product businesses.
What Is a Shopping Query?
The system identifies shopping queries using a weighted shopping-intent score. A prompt is classified as shopping-related when its score reaches 3 or more:
- Transactional keywords count 2 points each —
buy,purchase,order,shop,shopping,deal,discount,coupon,checkout,add to cart,subscribe - Informational keywords count 1 point each —
price,cost,best,top,recommend,compare,vs,versus,review,rating,cheap,affordable,premium,worth,alternative,product,brand
Keywords are matched on word boundaries, so top does not match laptop and price still matches pricing. (A single transactional word like buy scores 2 — below the threshold on its own — so genuinely shopping-oriented prompts usually combine intent and product terms.)
You can also explicitly mark prompts with type shopping for guaranteed inclusion.
Example Shopping Prompts
- "What is the best laptop to buy in 2025?"
- "Compare iPhone 16 vs Samsung Galaxy S25"
- "Best affordable wireless headphones under $100"
- "Which running shoes are best for marathon training?"
Key Metrics
Every figure on the page divides by runs — one question asked once of one engine — and says so on the card: "34 of 120 runs on shopping questions", "12 of 34 runs that named you". The four cards:
Named in the answer
Of every run on a shopping question in the window, how many named your brand. Carries a 90% credible interval on the card once there are at least 10 runs; below that the interval is withheld rather than printed on too little.
In the first three named
Of the runs that named you, how many put you among the first three companies the answer named, by the extracted rank. A named run with no extractable rank counts as not in the first three — the honest reading of "we cannot see a rank", never an inferred placement. This is a run-level reading; the question-level Top-3 presence on the Performance review counts questions, not runs.
Price shown & correct
Of the runs that named you, how many stated a price for you within 10% of the price recorded in Brand Facts. Both halves have to hold — a run with no price, or a wrong one, does not count. Price accuracy vs Brand Facts, further down the page, divides the same numerator by only the runs that stated a price near a brand mention; when no run did, that card says so instead of showing 0%.
Link present
Of the runs that named you, how many carried at least one link or citation in the answer. Any cited source counts — this is not the Performance review's Citation rate, which asks whether your own domain was cited. (Until 9 September 2026 this card was labelled "Citation Rate", the same words with a different meaning two pages apart.)
Tone
One line under the cards: the leading verdict — Positive, Neutral, Negative, or Mixed — over the runs that named your brand and were judged, with the count beside it, then how many named runs were not judged and how many runs did not name you. Under four judged answers only the count is shown. The same word appears per question in the table.
Shopping questions table
One row per shopping question, every column a count over its runs in the window: runs · named (n of runs) · in the first three (n of named) · price shown & correct (n of named) · link present (n of named) · tone.
Competitor Analysis
Which competitors appear alongside your brand in shopping queries, as one ranked bar — the share of shopping runs each brand was named in, your row emphasised, the count on every row — and a table:
- Named in — How many shopping runs named the competitor, as "n of N"
- With a price — How often their pricing was shown beside them
A competitor's position is not in the data (the rank extractor tracks your brand only), and position is a retired measure regardless.
How to Set Up
- Create shopping prompts — Write prompts that mimic how shoppers ask AI for advice.
- Include comparison prompts — "Compare [your brand] vs [competitor] for [use case]"
- Add pricing to Brand Facts — Go to Settings → Brand Facts and add
priceorpricingentries so the system can verify price accuracy. - Capture snapshots — The system automatically identifies shopping queries.
- Visit Shopping Visibility — Navigate from the sidebar to see results.
Optimization Tips
- Publish accurate pricing — Ensure current pricing is publicly available on your website with Product schema markup.
- Maintain product pages — Up-to-date specs, availability, and pricing in structured data.
- Encourage reviews — AI models weight third-party endorsements heavily in shopping contexts.
- Create comparison content — Publish honest comparison pages on your own site to influence AI training data.
- Monitor competitors — See which rivals are named on more of your shopping runs than you are, and how often a price travels with their name, then study the content behind it.
- Optimize for citations — Clear, linkable product pages with unique URLs improve citation rates.
Retired on 9 September 2026
- Average Position (a KPI, a competitor column, and the "Position Distribution" card) — a position within one answer moves between runs of the same question, so an average of it is not a measure a reader can act on. The dictionary keeps Top-3 presence and excludes rank and position; the definition of Avg. Rank is marked retired in the Metric definitions appendix. In the first three named is what replaces it here.
- The sentiment donut — three slices beside a legend that already said the same three numbers. The tone line and the per-question tone word carry it now.
- "Citation Rate" as the name of the link card — renamed Link present, because on the Performance review the same words mean "our domain was cited".
Plan Requirements
Shopping Visibility requires the Advanced AI Insights feature, available on Pro-Individual plans and above.