Example report
Explore an entirely fictional product report, created to illustrate the intended workflow. All product details, buyer comments, figures and sample answers are invented.
Share of visibility
Of 148 answers naming the category
Counts over the same 148 answers, so they add to a real total rather than to 100.
Where you appear
Appearances by measured surface
A zero is a finding: that surface answered and did not name you. Failed calls are reported apart.
Illustrative sources behind the answers
31 cited in total
The pages the answers actually cited — and where a source opportunity sits.
Answer “is hybrid non-stick safe?” on your own site. It shapes 19 of the 50 questions measured; three competitors are cited on it, you are cited on none.
Explore example reportWhat it looks like inside
The existing report screens with an example product and illustrative figures. Product identity, source names and excerpts are fictional; nothing is scanned.
Tavren Pace 3
road racing shoes · example-product.html
Your visibility and this week's priorities
Weekly summary
2026-08-31 to 2026-09-06 · 1 completed scan(s)
62% discovery visibility · scan #2 · 2026-09-01
Brand or product named in discovery answers, averaged across 32 question–assistant pairs (9 questions, 4 assistants). Repeated answers share one pair's weight. 32 successful answers; 0 failed answers excluded.
No earlier completed scan is available for this period's comparison.
0 task(s) signed off; 0 change(s) marked published by your team.
What to do this week
- Nothing to recommend from this scan meets the evidence checks. This is a finding about the evidence. Review any held recommendations.
Product and setup review
Tavren Pace 3
284.99 GBP · road racing shoes
A fictional road racing shoe used to illustrate the report. Product details, price, measurements and scores are examples, rather than claims about a real product.
An engine has to guess at a lot of this page. Guessed facts are the ones that come back wrong in an answer. · 9 of 9 facts declared, 1 blocking issue
Scan steps and supporting detail
Steps 4 and 5
A first read, while the scan is out
Answers already collected about tavren.example, taken 2026-08-31T20:39:26+00:00. This is not your measurement. It asked none of your questions, covers the whole domain rather than this product, and has a different denominator — the scan below is the figure your plan is for.
of the time an assistant names you — averaged over 32 question–assistant pairs (9 questions put to 4 assistants)
named youleft you outnot asked
Share of every brand named across the answers. Rank order, never reordered to put you first.
bigger → comes up on more of your questions
What the product page says
Tavren Pace 3
284.99 GBP · road racing shoes
A fictional road racing shoe used to illustrate the report. Product details, price, measurements and scores are examples, rather than claims about a real product.
An engine has to guess at a lot of this page. Guessed facts are the ones that come back wrong in an answer. · 9 of 9 facts declared, 1 blocking issue
Also proposed: racing shoes, shoes.
What the page declares
9 facts9 of 9 are declared in the page's own machine-readable data; 0 had to be inferred from the text. Fictional specifications for this example product.
| Product name | Tavren Pace 3 Men's Road Racing Shoes - Sage/Stone - Size 6 | json-ld |
| Price | 284.99 | json-ld |
| Currency | GBP | json-ld |
| Description | A fictional road racing shoe used to illustrate the report. Product details, price, measurements and scores are examples, rather than claims about a real product. | json-ld |
| Image | /static/example-product.svg | json-ld |
| GTIN / UPC | 00000000000000 | json-ld |
| MPN | EXAMPLE-003 | json-ld |
| Colour | Sage/Stone | json-ld |
| Condition | New | json-ld |
What an assistant has to guess at
2 gapsGuessed facts are the ones that come back wrong in an answer.
6 scored fields absent
No canonical URL
- stored as
- example-product.html
- how
- fetched from the site, once
- page title
- Tavren Pace 3 Men's Road Racing Shoes. Tavren UK
- canonical
- none — the page names no single address for itself
- language
- en-GB
- data blocks
- 1 machine-readable block(s): brand, offer, peopleaudience, product, productgroup
96 sample comments across 12 illustrative sources
Sample collection date 2026-08-31 · searched Tavren Pace Men's site:source-46.example, Tavren Pace Men's example-forum worth it, Tavren Pace Men's forum, Tavren Pace Men's discussion, Tavren Pace Men's review, best racing shoes for 5k example-forum, best racing shoes for marathon example-forum, best racing shoes running example-forum
source-46.example
80 posts“I would like to compare the fit and sizing before choosing a road racing shoe.”
“Clear information about the intended running distance would help me decide.”
“I want to understand the listed materials and care instructions before buying.”
“A comparison should explain the differences without promising a personal result.”
“I would check the size guide and return conditions before placing an order.”
“I am looking for clear specifications rather than a claim that one shoe suits everyone.”
“I would like to compare the fit and sizing before choosing a road racing shoe.”
“Clear information about the intended running distance would help me decide.”
“I want to understand the listed materials and care instructions before buying.”
“A comparison should explain the differences without promising a personal result.”
“I would check the size guide and return conditions before placing an order.”
“I am looking for clear specifications rather than a claim that one shoe suits everyone.”
“I would like to compare the fit and sizing before choosing a road racing shoe.”
“Clear information about the intended running distance would help me decide.”
“I want to understand the listed materials and care instructions before buying.”
“A comparison should explain the differences without promising a personal result.”
“I would check the size guide and return conditions before placing an order.”
“I am looking for clear specifications rather than a claim that one shoe suits everyone.”
“I would like to compare the fit and sizing before choosing a road racing shoe.”
“Clear information about the intended running distance would help me decide.”
“I want to understand the listed materials and care instructions before buying.”
“A comparison should explain the differences without promising a personal result.”
“I would check the size guide and return conditions before placing an order.”
“I am looking for clear specifications rather than a claim that one shoe suits everyone.”
“I would like to compare the fit and sizing before choosing a road racing shoe.”
“Clear information about the intended running distance would help me decide.”
“I want to understand the listed materials and care instructions before buying.”
“A comparison should explain the differences without promising a personal result.”
“I would check the size guide and return conditions before placing an order.”
“I am looking for clear specifications rather than a claim that one shoe suits everyone.”
“I would like to compare the fit and sizing before choosing a road racing shoe.”
“Clear information about the intended running distance would help me decide.”
“I want to understand the listed materials and care instructions before buying.”
“A comparison should explain the differences without promising a personal result.”
“I would check the size guide and return conditions before placing an order.”
“I am looking for clear specifications rather than a claim that one shoe suits everyone.”
“I would like to compare the fit and sizing before choosing a road racing shoe.”
“Clear information about the intended running distance would help me decide.”
“I want to understand the listed materials and care instructions before buying.”
“A comparison should explain the differences without promising a personal result.”
“I would check the size guide and return conditions before placing an order.”
“I am looking for clear specifications rather than a claim that one shoe suits everyone.”
“I would like to compare the fit and sizing before choosing a road racing shoe.”
“Clear information about the intended running distance would help me decide.”
“I want to understand the listed materials and care instructions before buying.”
“A comparison should explain the differences without promising a personal result.”
“I would check the size guide and return conditions before placing an order.”
“I am looking for clear specifications rather than a claim that one shoe suits everyone.”
“I would like to compare the fit and sizing before choosing a road racing shoe.”
“Clear information about the intended running distance would help me decide.”
“I want to understand the listed materials and care instructions before buying.”
“A comparison should explain the differences without promising a personal result.”
“I would check the size guide and return conditions before placing an order.”
“I am looking for clear specifications rather than a claim that one shoe suits everyone.”
“I would like to compare the fit and sizing before choosing a road racing shoe.”
“Clear information about the intended running distance would help me decide.”
“I want to understand the listed materials and care instructions before buying.”
“A comparison should explain the differences without promising a personal result.”
“I would check the size guide and return conditions before placing an order.”
“I am looking for clear specifications rather than a claim that one shoe suits everyone.”
“I would like to compare the fit and sizing before choosing a road racing shoe.”
“Clear information about the intended running distance would help me decide.”
“I want to understand the listed materials and care instructions before buying.”
“A comparison should explain the differences without promising a personal result.”
“I would check the size guide and return conditions before placing an order.”
“I am looking for clear specifications rather than a claim that one shoe suits everyone.”
“I would like to compare the fit and sizing before choosing a road racing shoe.”
“Clear information about the intended running distance would help me decide.”
“I want to understand the listed materials and care instructions before buying.”
“A comparison should explain the differences without promising a personal result.”
“I would check the size guide and return conditions before placing an order.”
“I am looking for clear specifications rather than a claim that one shoe suits everyone.”
“I would like to compare the fit and sizing before choosing a road racing shoe.”
“Clear information about the intended running distance would help me decide.”
“I want to understand the listed materials and care instructions before buying.”
“A comparison should explain the differences without promising a personal result.”
“I would check the size guide and return conditions before placing an order.”
“I am looking for clear specifications rather than a claim that one shoe suits everyone.”
“I would like to compare the fit and sizing before choosing a road racing shoe.”
“Clear information about the intended running distance would help me decide.”
source-43.example
4 posts“I want to understand the listed materials and care instructions before buying.”
“A comparison should explain the differences without promising a personal result.”
“I would check the size guide and return conditions before placing an order.”
“I am looking for clear specifications rather than a claim that one shoe suits everyone.”
source-07.example
2 posts“I would like to compare the fit and sizing before choosing a road racing shoe.”
“Clear information about the intended running distance would help me decide.”
source-22.example
2 posts“I want to understand the listed materials and care instructions before buying.”
“A comparison should explain the differences without promising a personal result.”
source-38.example
1 post“I would check the size guide and return conditions before placing an order.”
source-35.example
1 post“I am looking for clear specifications rather than a claim that one shoe suits everyone.”
source-39.example
1 post“I would like to compare the fit and sizing before choosing a road racing shoe.”
source-16.example
1 post“Clear information about the intended running distance would help me decide.”
source-20.example
1 post“I want to understand the listed materials and care instructions before buying.”
source-21.example
1 post“A comparison should explain the differences without promising a personal result.”
source-36.example
1 post“I would check the size guide and return conditions before placing an order.”
source-18.example
1 post“I am looking for clear specifications rather than a claim that one shoe suits everyone.”
The questions to target for road racing shoes
Review product setup
Use the model name buyers recognise. Leave colourways, retailer suffixes and sizes out unless they define the product being tracked.
Approve the questions
Approved questions are eligible for future scans. Excluded questions stay in the evidence record.
best road racing shoes for marathon — approved
Provenance: observed-demand. Editing creates a customer-added question; saved observed wording stays intact.
best road running racing shoes — approved
Provenance: observed-demand. Editing creates a customer-added question; saved observed wording stays intact.
top road racing shoes — approved
Provenance: observed-demand. Editing creates a customer-added question; saved observed wording stays intact.
road racing running shoes — approved
Provenance: observed-demand. Editing creates a customer-added question; saved observed wording stays intact.
carbon plate running shoes for road racing — approved
Provenance: observed-demand. Editing creates a customer-added question; saved observed wording stays intact.
carbon road racing shoes — approved
Provenance: observed-demand. Editing creates a customer-added question; saved observed wording stays intact.
mens road racing shoes — approved
Provenance: observed-demand. Editing creates a customer-added question; saved observed wording stays intact.
What are the best road racing shoes alternatives to Tavren Pace 3? — approved
Provenance: semantic-coverage. Editing creates a customer-added question; saved observed wording stays intact.
How does Tavren Pace 3 compare with Aveltric for road racing shoes? — approved
Provenance: semantic-coverage. Editing creates a customer-added question; saved observed wording stays intact.
How does Tavren Pace 3 compare with Corvane for road racing shoes? — approved
Provenance: semantic-coverage. Editing creates a customer-added question; saved observed wording stays intact.
Is Tavren Pace 3 worth it? — approved
Provenance: semantic-coverage. Editing creates a customer-added question; saved observed wording stays intact.
Who should consider buying Tavren Pace 3? — approved
Provenance: semantic-coverage. Editing creates a customer-added question; saved observed wording stays intact.
Tavren Pace 3
284.99 GBP · road racing shoes
A fictional road racing shoe used to illustrate the report. Product details, price, measurements and scores are examples, rather than claims about a real product.
An engine has to guess at a lot of this page. Guessed facts are the ones that come back wrong in an answer. · 9 of 9 facts declared, 1 blocking issue
no DATAFORSEO_LOGIN or DATAFORSEO_PASSWORD — search volumes need both. Add them to config/keys.env from the API Access tab of the DataForSEO dashboard. Until then the live purchase is not offered. The mock press above sends nothing to DataForSEO and bills nothing — it shows what this step does without needing a key at all.
12 questions, grouped by what they measure
✓ Discovery questions do not name Tavren; competitive and branded questions intentionally do · ·
Discovery visibility
7 questions46 postsNon-branded shortlist questions where this product is a credible answer. This is the primary AI visibility metric.
“best road racing shoes for marathon”
25 supporting passages — open and read them
“I would like to compare the fit and sizing before choosing a road racing shoe.”
“Clear information about the intended running distance would help me decide.”
“I want to understand the listed materials and care instructions before buying.”
“A comparison should explain the differences without promising a personal result.”
“I would check the size guide and return conditions before placing an order.”
“I am looking for clear specifications rather than a claim that one shoe suits everyone.”
“I would like to compare the fit and sizing before choosing a road racing shoe.”
“Clear information about the intended running distance would help me decide.”
“best road running racing shoes”
32 supporting passages — open and read them
“I want to understand the listed materials and care instructions before buying.”
“A comparison should explain the differences without promising a personal result.”
“I would check the size guide and return conditions before placing an order.”
“I am looking for clear specifications rather than a claim that one shoe suits everyone.”
“I would like to compare the fit and sizing before choosing a road racing shoe.”
“Clear information about the intended running distance would help me decide.”
“I want to understand the listed materials and care instructions before buying.”
“A comparison should explain the differences without promising a personal result.”
“top road racing shoes”
Where this phrase came from · 2 other wordings
“road racing running shoes”
32 supporting passages — open and read them · 1 other wording
“I would check the size guide and return conditions before placing an order.”
“I am looking for clear specifications rather than a claim that one shoe suits everyone.”
“I would like to compare the fit and sizing before choosing a road racing shoe.”
“Clear information about the intended running distance would help me decide.”
“I want to understand the listed materials and care instructions before buying.”
“A comparison should explain the differences without promising a personal result.”
“I would check the size guide and return conditions before placing an order.”
“I am looking for clear specifications rather than a claim that one shoe suits everyone.”
“carbon plate running shoes for road racing”
33 supporting passages — open and read them
“I would like to compare the fit and sizing before choosing a road racing shoe.”
“Clear information about the intended running distance would help me decide.”
“I want to understand the listed materials and care instructions before buying.”
“A comparison should explain the differences without promising a personal result.”
“I would check the size guide and return conditions before placing an order.”
“I am looking for clear specifications rather than a claim that one shoe suits everyone.”
“I would like to compare the fit and sizing before choosing a road racing shoe.”
“Clear information about the intended running distance would help me decide.”
“carbon road racing shoes”
4 supporting passages — open and read them
“I want to understand the listed materials and care instructions before buying.”
“A comparison should explain the differences without promising a personal result.”
“I would check the size guide and return conditions before placing an order.”
“I am looking for clear specifications rather than a claim that one shoe suits everyone.”
“mens road racing shoes”
Where this phrase came from · 2 other wordings
Competitive visibility
3 questions0 postsExplicit comparisons or switching questions that name this product. A rival-only search never counts against it.
“What are the best road racing shoes alternatives to Tavren Pace 3?”
Where this phrase came from
“How does Tavren Pace 3 compare with Aveltric for road racing shoes?”
Where this phrase came from
“How does Tavren Pace 3 compare with Corvane for road racing shoes?”
Where this phrase came from
Branded visibility
2 questions0 postsQuestions from buyers already considering this product: suitability, value and drawbacks.
“Is Tavren Pace 3 worth it?”
Where this phrase came from
“Who should consider buying Tavren Pace 3?”
Where this phrase came from
What the assistants said when asked these questions
Three questions, three jobs
The scores are not averaged together. 11 historic off-target question(s) were excluded from every score.
Non-branded shortlist questions where this product is a credible answer. This is the primary AI visibility metric.
Explicit comparisons or switching questions that name this product. A rival-only search never counts against it.
Questions from buyers already considering this product: suitability, value and drawbacks.
Change withheld against the scan on 31 Aug. The question sets differ (20 added, 41 dropped), so the two runs are not measuring the same thing. Re-run the baseline's questions if you want a like-for-like number. The answered questions, surfaces or measurement settings differ; change is withheld. chatgpt-web-search answered from gpt-5-6 in the baseline and chatgpt-web-search later. A model change moves these numbers on its own — whatever moved here is not evidence that the work worked. The baseline scan is not complete; change is withheld.
The three targets — where the product appears
One row per eligible query type; one column per assistant. The number is how often this product was named in the answers that came back. Discovery is the primary metric; competitive and branded visibility are kept separate. A hatched cell is a type that assistant never answered — not an absence.
| Query type | ChatGPT Search | Gemini Web | Google AI Mode | Google AI Overview | All |
|---|---|---|---|---|---|
| Discovery visibilitydiscovery · 9 eligible question(s) · 32 answer(s) | 78% | 67% | 44% | 60% | 62% |
Read the Discovery visibility row first — it carries 32 of the 32 eligible answers in this scan. Only the discovery row moves the headline.
Ground held
Defend these. A scan that keeps them did not go backwards.
Ground lost — and who took it
The weakest questions in the scan, each with the brand recommended in answers this product was missing from.
What to do this week
- Nothing to recommend from this scan meets the evidence checks. This is a finding about the evidence. Review any held recommendations.
Held back until you say who this is
Real findings, with no action written for them, because the name they turn on has not been identified. Work aimed at out-publishing a shop or a feature is work aimed at nothing, so nothing is written until a person has said which it is. Identify them — it takes a minute and these become actions.
The whole run in one picture
One row per assistant, one square per question — 32 reading(s) in all. Point at any square to read the question that was put to the assistant. The 0 outlined square(s) are calls that never came back: they leave the denominator rather than counting as answers you were absent from.
ChatGPT Search
selvoro road racing shoesanswered, did not name you
ChatGPT Search
Veylora road racing shoesnamed you
ChatGPT Search
carbon road racing shoesnamed you
ChatGPT Search
best road running racing shoesnamed you
ChatGPT Search
road racing running shoesnamed you
ChatGPT Search
top road racing shoesnamed you
ChatGPT Search
best road racing shoes for marathonnamed you
ChatGPT Search
carbon plate running shoes for road racingnamed you
ChatGPT Search
mens road racing shoes7 of 9Gemini Webanswered, did not name you
Gemini Web
selvoro road racing shoesanswered, did not name you
Gemini Web
Veylora road racing shoesanswered, did not name you
Gemini Web
carbon road racing shoesnamed you
Gemini Web
best road running racing shoesnamed you
Gemini Web
road racing running shoesnamed you
Gemini Web
top road racing shoesnamed you
Gemini Web
best road racing shoes for marathonnamed you
Gemini Web
carbon plate running shoes for road racingnamed you
Gemini Web
mens road racing shoes6 of 9Google AI Modeanswered, did not name you
Google AI Mode
selvoro road racing shoesanswered, did not name you
Google AI Mode
Veylora road racing shoesanswered, did not name you
Google AI Mode
carbon road racing shoesnamed you
Google AI Mode
best road running racing shoesanswered, did not name you
Google AI Mode
road racing running shoesanswered, did not name you
Google AI Mode
top road racing shoesnamed you
Google AI Mode
best road racing shoes for marathonnamed you
Google AI Mode
carbon plate running shoes for road racingnamed you
Google AI Mode
mens road racing shoes4 of 9Google AI Overviewanswered, did not name you
Google AI Overview
selvoro road racing shoesnot asked of this assistant
Google AI Overview
Veylora road racing shoesnot asked of this assistant
Google AI Overview
carbon road racing shoesanswered, did not name you
Google AI Overview
best road running racing shoesnamed you
Google AI Overview
road racing running shoesnamed you
Google AI Overview
top road racing shoesnot asked of this assistant
Google AI Overview
best road racing shoes for marathonnamed you
Google AI Overview
carbon plate running shoes for road racingnot asked of this assistant
Google AI Overview
mens road racing shoes3 of 5
No assistant can be reached
Add the credentials for at least one configured surface to config/keys.env, then restart the app. janus keys names what is missing without ever printing a secret.
Just want to see what step 4 produces? The mock walks the whole stage, asks nobody anything, and sends no request of any kind — no assistant, no search, no key touched, nothing billed and nothing counted against your plan. Every row it writes is stamped a mock in four places. It measures nothing.
Look at part of this run
Narrow the whole report to one assistant, one kind of question, or the answers naming one rival. Nothing here asks an assistant anything — it re-reads answers already stored.
The evidence underneath
Every figure above is derived from stored answers, and the rows behind each one are here rather than on another page.
Every question against every assistant
Where it wins and where it loses
One square is one question put to one assistant. The mark says what happened, so the reading does not depend on telling the colours apart; a hatched square is a call that failed and is not an absence. Point at any square for the sentence it was read from.
chatgpt-web-search
1 repeat(s) · asked once, so nothing here says it would hold
selvoro road racing shoes
gemini-web
1 repeat(s) · asked once, so nothing here says it would hold
selvoro road racing shoes
google-ai-mode
1 repeat(s) · asked once, so nothing here says it would hold
selvoro road racing shoes
google-ai-overview
1 repeat(s) · asked once, so nothing here says it would hold
selvoro road racing shoes
chatgpt-web-search
1 repeat(s) · asked once, so nothing here says it would hold
Veylora road racing shoes
gemini-web
1 repeat(s) · asked once, so nothing here says it would hold
Veylora road racing shoes
google-ai-mode
1 repeat(s) · asked once, so nothing here says it would hold
Veylora road racing shoes
chatgpt-web-search
1 repeat(s) · asked once, so nothing here says it would hold
carbon road racing shoes
gemini-web
1 repeat(s) · asked once, so nothing here says it would hold
carbon road racing shoes
google-ai-mode
1 repeat(s) · asked once, so nothing here says it would hold
carbon road racing shoes
chatgpt-web-search
1 repeat(s) · asked once, so nothing here says it would hold
best road running racing shoes
gemini-web
1 repeat(s) · asked once, so nothing here says it would hold
best road running racing shoes
google-ai-mode
1 repeat(s) · asked once, so nothing here says it would hold
best road running racing shoes
google-ai-overview
1 repeat(s) · asked once, so nothing here says it would hold
best road running racing shoes
chatgpt-web-search
1 repeat(s) · asked once, so nothing here says it would hold
road racing running shoes
gemini-web
1 repeat(s) · asked once, so nothing here says it would hold
road racing running shoes
google-ai-mode
1 repeat(s) · asked once, so nothing here says it would hold
road racing running shoes
google-ai-overview
1 repeat(s) · asked once, so nothing here says it would hold
road racing running shoes
chatgpt-web-search
1 repeat(s) · asked once, so nothing here says it would hold
top road racing shoes
gemini-web
1 repeat(s) · asked once, so nothing here says it would hold
top road racing shoes
google-ai-mode
1 repeat(s) · asked once, so nothing here says it would hold
top road racing shoes
google-ai-overview
1 repeat(s) · asked once, so nothing here says it would hold
top road racing shoes
chatgpt-web-search
1 repeat(s) · asked once, so nothing here says it would hold
best road racing shoes for marathon
gemini-web
1 repeat(s) · asked once, so nothing here says it would hold
best road racing shoes for marathon
google-ai-mode
1 repeat(s) · asked once, so nothing here says it would hold
best road racing shoes for marathon
chatgpt-web-search
1 repeat(s) · asked once, so nothing here says it would hold
carbon plate running shoes for road racing
gemini-web
1 repeat(s) · asked once, so nothing here says it would hold
carbon plate running shoes for road racing
google-ai-mode
1 repeat(s) · asked once, so nothing here says it would hold
carbon plate running shoes for road racing
google-ai-overview
1 repeat(s) · asked once, so nothing here says it would hold
carbon plate running shoes for road racing
chatgpt-web-search
1 repeat(s) · asked once, so nothing here says it would hold
mens road racing shoes
gemini-web
1 repeat(s) · asked once, so nothing here says it would hold
mens road racing shoes
google-ai-mode
1 repeat(s) · asked once, so nothing here says it would hold
mens road racing shoes
The same thing as rows, question by question
Question by question
The questions come from stages 1–4 and are stored as they were asked, so re-wording a question later never rewrites what a past run measured.
Visibility by query type
Discovery, competitive and branded questions are scored separately because each measures a different buyer moment. A rival-only search is not eligible for any row.
What the answers actually said, rate by rate
What the answers actually said
Raw discovery answer counts: the right side names the brand or product; the left side does not. These proportions count all 32 successful answers, including repeats. The headline instead gives equal weight to each question–assistant pair, so its percentage can differ when repeat counts vary.
12 of 32 answer(s) · 38%
20 of 32 answer(s) · 62%
Every rate, with the arithmetic behind it
How visible it is
Out of 32 answer(s) that came back. 0 call(s) failed and are excluded from the denominator, not counted as absences.
How named at all was calculated
Averaged over 32 question×assistant cell(s) — 9 question(s) put to 4 assistant(s) — from 32 answer(s) that came back = 62%.
The unit is the cell, not the answer. Each question is scored on its own repeats and the cells are then averaged, so a question that was asked a third time does not outweigh one that settled after two. Counted flat across every answer instead, this reads 62% (20 of 32).
0 call(s) failed and are excluded rather than counted as the client being absent — a run where the assistants fell over must not read as a run where nobody named the client. A cell whose every repeat failed is left out entirely.
Each answer was read by the lexical-v2 classifier, which decides from the sentence containing the brand and stores that sentence so the call can be checked. How this was arrived at
How recommended was calculated
Averaged over 32 question×assistant cell(s) — 9 question(s) put to 4 assistant(s) — from 32 answer(s) that came back = 0%.
The unit is the cell, not the answer. Each question is scored on its own repeats and the cells are then averaged, so a question that was asked a third time does not outweigh one that settled after two. Counted flat across every answer instead, this reads 0% (0 of 32).
0 call(s) failed and are excluded rather than counted as the client being absent — a run where the assistants fell over must not read as a run where nobody named the client. A cell whose every repeat failed is left out entirely.
Each answer was read by the lexical-v2 classifier, which decides from the sentence containing the brand and stores that sentence so the call can be checked. How this was arrived at
How compared against a rival was calculated
Averaged over 32 question×assistant cell(s) — 9 question(s) put to 4 assistant(s) — from 32 answer(s) that came back = 0%.
The unit is the cell, not the answer. Each question is scored on its own repeats and the cells are then averaged, so a question that was asked a third time does not outweigh one that settled after two. Counted flat across every answer instead, this reads 0% (0 of 32).
0 call(s) failed and are excluded rather than counted as the client being absent — a run where the assistants fell over must not read as a run where nobody named the client. A cell whose every repeat failed is left out entirely.
Each answer was read by the lexical-v2 classifier, which decides from the sentence containing the brand and stores that sentence so the call can be checked. How this was arrived at
How warned about was calculated
Averaged over 32 question×assistant cell(s) — 9 question(s) put to 4 assistant(s) — from 32 answer(s) that came back = 0%.
The unit is the cell, not the answer. Each question is scored on its own repeats and the cells are then averaged, so a question that was asked a third time does not outweigh one that settled after two. Counted flat across every answer instead, this reads 0% (0 of 32).
0 call(s) failed and are excluded rather than counted as the client being absent — a run where the assistants fell over must not read as a run where nobody named the client. A cell whose every repeat failed is left out entirely.
Each answer was read by the lexical-v2 classifier, which decides from the sentence containing the brand and stores that sentence so the call can be checked. How this was arrived at
How absent was calculated
Averaged over 32 question×assistant cell(s) — 9 question(s) put to 4 assistant(s) — from 32 answer(s) that came back = 38%.
The unit is the cell, not the answer. Each question is scored on its own repeats and the cells are then averaged, so a question that was asked a third time does not outweigh one that settled after two. Counted flat across every answer instead, this reads 38% (12 of 32).
0 call(s) failed and are excluded rather than counted as the client being absent — a run where the assistants fell over must not read as a run where nobody named the client. A cell whose every repeat failed is left out entirely.
Each answer was read by the lexical-v2 classifier, which decides from the sentence containing the brand and stores that sentence so the call can be checked. How this was arrived at
Average position when named: 1.4 (best 1)
How stable the answers are
Every question was asked once, so nothing here can say whether an answer would hold. Assistants vary between calls: a rate from single asks is a snapshot, not a measurement. Re-run with --real (3 repeats) to find out which of these answers are stable.
How it is perceived
Quoted, not summarised — these are the assistants' own sentences, with the run they came from.
The whole field, and who is named instead of you
Who is getting the attention
This product is in blue, every rival in grey. A recommendation counts for more than a shortlisting, which counts for more than a name in passing. Which names count as brands is a heuristic, so the tail is folded rather than drawn — pin the set with competitors: in this product's engagement.yaml.
The full field, and who is named instead of this product
Who is doing better
Share of voice weights a recommendation above a shortlisting above a passing mention. Competitor names are extracted from the answers themselves and are a heuristic — pin them with a competitors: list in this product's engagement.yaml.
How share of voice was calculated
Each brand's appearances are weighted by how strong each one was — a recommendation counts for more than a shortlisting, which counts for more than a name in passing — and each brand's total is divided by every brand's total. It is a share of the naming that happened, not a share of a market.
Which names count as brands is decided by a capitalisation heuristic stamped heuristic-v1, so this table should be read as a proposal. Pinning the set replaces the heuristic rather than adding to it. How this was arrived at
Who is named instead
Counted only over the 12 answer(s) this product was absent from, so a rival that appears beside it scores nothing here and one that appears instead of it scores every time.
What rivals are credited with
Quoted whole from the answer, and attributed only to the brand the clause follows — a sentence naming two brands does not credit both. Lexical, low confidence, and worth reading rather than counting.
asked: selvoro road racing shoes
asked: How should buyers compare road racing shoes for super?
asked: best road racing shoes for marathon
asked: How should buyers compare road racing shoes for super?
asked: Aveltric road racing shoes
asked: mens road racing shoes
asked: mens road racing shoes
asked: How should buyers compare road racing shoes for super?
asked: mens road racing shoes
What was named — the product, or only the company
What was named
Four figures, never one. Each is the share of question×assistant cells that named you at that level or more specifically — naming the exact product also names the company, so the figures widen as they go down. Told they appear in 62% of answers, a customer who then finds most of those were the company name in a list of manufacturers has been misled.
Read against “named at all” above: the gap between them is visibility you have that is not visibility for this product.
Every opening, and what scored it
What is worth working on
Up the page is how much of the answering it drives; across is how reachable it is. A filled dot is an answer the assistants built from pages they looked up, so publishing can still change it. A hollow dot is an answer the model produced from what it already remembered — nothing published this quarter will move it, and the two are identical in any list of scores.
worth 100% · reachable 57%
publishing can move thissource-29.example
worth 100% · reachable 57%
publishing can move this(publisher not named)
worth 100% · reachable 44%
publishing can move thissource-37.example
worth 100% · reachable 38%
publishing can move thissource-36.example
worth 100% · reachable 38%
publishing can move thissource-13.example
worth 100% · reachable 38%
publishing can move thissource-14.example
worth 100% · reachable 38%
publishing can move thissource-11.example
worth 100% · reachable 38%
publishing can move thissource-17.example
worth 95% · reachable 38%
publishing can move thissource-05.example
worth 92% · reachable 38%
publishing can move thissource-19.example
worth 89% · reachable 38%
publishing can move thissource-03.example
worth 84% · reachable 38%
publishing can move thissource-44.example
worth 84% · reachable 32%
publishing can move thissource-02.example
worth 70% · reachable 38%
publishing can move thissource-33.example
worth 66% · reachable 38%
publishing can move thissource-47.example
worth 66% · reachable 38%
publishing can move thissource-12.example
worth 60% · reachable 38%
publishing can move thissource-24.example
worth 59% · reachable 57%
publishing can move thiswww.capeunionmart.co.za
worth 51% · reachable 72%
publishing can move thissource-09.example
worth 51% · reachable 72%
publishing can move thissource-01.example
worth 51% · reachable 38%
publishing can move thism.joybuy.de
worth 51% · reachable 38%
publishing can move thissource-08.example.br
worth 49% · reachable 72%
publishing can move thissource-30.example
worth 49% · reachable 72%
publishing can move this
Every opening, with the factors that scored it
What can realistically be influenced
Feasibility is how movable the source is; influence is how much of the answering it appears to drive. They are kept apart so you can see which of the two a priority came from, and every factor that fired is listed with the sentence behind it so the score can be argued with rather than believed.
a community platform — open to participate in, not to control. Cited 16 time(s) across 7 question(s) by 3 assistant(s). No page of the client's is cited here. Confidence held down by: more than one assistant cited it.
a community platform — open to participate in, not to control. Cited 26 time(s) across 12 question(s) by 4 assistant(s). No page of the client's is cited here. Confidence held down by: more than one assistant cited it.
no platform signal in the domain — worth a human look. Cited 52 time(s) across 13 question(s) by 2 assistant(s). No page of the client's is cited here. Confidence held down by: page type unknown, more than one assistant cited it.
the URL reads as a published article (inferred from the URL, not from the domain). Cited 28 time(s) across 7 question(s) by 2 assistant(s). No page of the client's is cited here. Confidence held down by: more than one assistant cited it.
the URL reads as a published article (inferred from the URL, not from the domain). Cited 30 time(s) across 12 question(s) by 4 assistant(s). No page of the client's is cited here. Confidence held down by: more than one assistant cited it.
the URL reads as a published article (inferred from the URL, not from the domain). Cited 14 time(s) across 8 question(s) by 2 assistant(s). No page of the client's is cited here. Confidence held down by: more than one assistant cited it.
the URL reads as a published article (inferred from the URL, not from the domain). Cited 22 time(s) across 9 question(s) by 4 assistant(s). No page of the client's is cited here. Confidence held down by: more than one assistant cited it.
the URL reads as a published article (inferred from the URL, not from the domain). Cited 16 time(s) across 12 question(s) by 2 assistant(s). No page of the client's is cited here. Confidence held down by: more than one assistant cited it.
the client's own domain. Cited 2 time(s) across 2 question(s) by 1 assistant(s). The client is already cited here, so the work is deepening an existing page rather than winning a new one.
the URL is a product listing, so its copy is supplied (inferred from the URL, not from the domain). Cited 2 time(s) across 1 question(s) by 1 assistant(s). No page of the client's is cited here.
Where the questions came from
What the questions are based on
Only direct observed demand is evidence that buyers searched for a question. Every weaker layer is coverage: useful to measure, but never presented as observed demand.
Every question, and who wins it
One row per underlying question, weakest first. Several wordings of the same question are one row and one finding — the wordings are under each row, so a fold you disagree with is visible rather than buried. The rows under “Where it wins and where it loses” are the same answers listed one wording at a time, to match the picture above them.
Which assistant and which model answered
How AI assistants answer for road racing shoes
Observed interface visibility
Non-personalised public-interface results collected for the stated location, language, device and time. Not a record of what every user saw — signed-in state, personalisation, tests and the moment can differ.
Recommended in 0%, first-named in 60%, warned about in 0%. Repeats agreed on 0% of the 0 question(s) asked more than once.
Recommended in 0%, first-named in 44%, warned about in 0%. Repeats agreed on 0% of the 0 question(s) asked more than once.
Recommended in 0%, first-named in 33%, warned about in 0%. Repeats agreed on 0% of the 0 question(s) asked more than once.
Recommended in 0%, first-named in 67%, warned about in 0%. Repeats agreed on 0% of the 0 question(s) asked more than once.
There is no overall score across observed interfaces and modelled assistants. They are different kinds of evidence, and averaging them would produce a number that means nothing.
Which model answered
The same run split by the model behind each assistant, not just by its name. A provider whose model changed between runs is the commonest reason a number moves without anybody touching the site.
What the assistants cite
What is shaping the answers
Retrieval estimate: memory 3 · mixed 1 · retrieval 28 (average confidence 77%). Retrieval is inferred from citations, adapter capability and phrasing. It is an estimate with a confidence, never a statement about what the assistant did.
The pages behind the answers
One row per page, deduplicated by normalised URL so the same article arriving with two tracking tails is one source. Nothing here has opened any of these pages: the type is read off the URL and the client/rival column counts what the citing answer named, which is co-occurrence and not what the page says.
26 citation(s) across 6 question(s), 2 assistant(s) · appears alongside both · publication 361Sport
“Manufacturer Trials: Brands like [Veylora Running](”
invented sample answer excerpt for this illustrative source (about Veylora), not from the page itself — nothing here has opened it. Page type: the path reads as a headline slug.
11 citation(s) across 3 question(s), 3 assistant(s) · appears alongside neither · publication Runnersneed
page type article inferred: the path reads as a headline slug
10 citation(s) across 10 question(s), 1 assistant(s) · appears alongside both · publication Runnersworld
“Tavren Pace 3 — Fast marathon PBs / elite racing — Extremely bouncy, efficient, and protective over long distances; widely regarded as one of the strongest marathon racers.”
invented sample answer excerpt for this illustrative source (about the client), not from the page itself — nothing here has opened it. Page type: the path reads as a headline slug.
8 citation(s) across 3 question(s), 2 assistant(s) · appears alongside both · publication Example outdoor journal
“Brands like Tavren fit narrow, while others offer wider platforms or better lockdown.”
invented sample answer excerpt for this illustrative source (about the client), not from the page itself — nothing here has opened it. Page type: the path reads as a headline slug.
8 citation(s) across 2 question(s), 1 assistant(s) · appears alongside both · publication Runrepeat
“Tavren Pace 3”
invented sample answer excerpt for this illustrative source (about the client), not from the page itself — nothing here has opened it. Page type: the path reads as a headline slug.
7 citation(s) across 7 question(s), 1 assistant(s) · appears alongside both · publication Example running journal
“Tavren Pace 3 — Fast marathon PBs / elite racing — Extremely bouncy, efficient, and protective over long distances; widely regarded as one of the strongest marathon racers.”
invented sample answer excerpt for this illustrative source (about the client), not from the page itself — nothing here has opened it. Page type: the path reads as a headline slug.
7 citation(s) across 5 question(s), 3 assistant(s) · appears alongside both · publication Runnersneed
“Manufacturer Trials: Brands like [Veylora Running](”
invented sample answer excerpt for this illustrative source (about Veylora), not from the page itself — nothing here has opened it. Page type: the path reads as a headline slug.
7 citation(s) across 1 question(s), 1 assistant(s) · appears alongside competitors · publication Example running journal
“Selvoro’s road racing and speed-oriented lineup features dedicated carbon-plated marathon tier racers, plated speed trainers, and lightweight everyday trainers.”
invented sample answer excerpt for this illustrative source (about Selvoro), not from the page itself — nothing here has opened it. Page type: the path reads as a headline slug.
6 citation(s) across 1 question(s), 1 assistant(s) · appears alongside competitors · publication Fleetfeet
“Aveltric has one of the strongest lineups in road racing, heavily driven by carbon-plated "super shoes" tailored to different running styles, alongside lightweight options for tempo efforts and training.”
invented sample answer excerpt for this illustrative source (about Aveltric), not from the page itself — nothing here has opened it. Page type: the path reads as a headline slug.
6 citation(s) across 1 question(s), 1 assistant(s) · appears alongside neither · publication Rouvy
page type article inferred: the path reads as a headline slug
5 citation(s) across 1 question(s), 1 assistant(s) · appears alongside both · publication Rundna
“Soft and Bouncy: Brands like Tavren (responsive foam) and Aveltric (responsive foam) offer extremely plush, highly energetic foams that isolate you from the road.”
invented sample answer excerpt for this illustrative source (about the client), not from the page itself — nothing here has opened it. Page type: the path reads as a headline slug.
4 citation(s) across 4 question(s), 1 assistant(s) · appears alongside both · publication Example running guide
“Tavren Sprint 2 - £104.”
invented sample answer excerpt for this illustrative source (about the client), not from the page itself — nothing here has opened it. Page type: the path reads as a headline slug.
4 citation(s) across 3 question(s), 2 assistant(s) · appears alongside competitors · publication Rei
“Manufacturer Trials: Brands like [Veylora Running](”
invented sample answer excerpt for this illustrative source (about Veylora), not from the page itself — nothing here has opened it. Page type: no shape in the URL says what kind of page it is.
4 citation(s) across 1 question(s), 3 assistant(s) · appears alongside client · publication Fleetfeet
“Racing shoes often feature ultra-responsive superfoams and carbon-fibre plates (such as the Tavren [Tavren Tempo 6 Go to product viewer dialogue for this item.”
invented sample answer excerpt for this illustrative source (about the client), not from the page itself — nothing here has opened it. Page type: the path reads as a headline slug.
Readings worth arguing with
Waiting for a second opinion
20 weak classification(s) flagged and 0 corrected by hand. Only calls that change what to do next are flagged — the client's own row, and anything read as a recommendation or a warning. Override one with janus measure review set; the classifier's original reading is kept beside it.
mention read at 40% confidence, below the 50% threshold, and it is a call that changes what to do next · google-ai-overview · ref
response:216|brand:tavrenmention read at 40% confidence, below the 50% threshold, and it is a call that changes what to do next · google-ai-overview · ref
response:217|brand:tavrenmention read at 40% confidence, below the 50% threshold, and it is a call that changes what to do next · google-ai-overview · ref
response:220|brand:tavrenmention read at 40% confidence, below the 50% threshold, and it is a call that changes what to do next · google-ai-mode · ref
response:233|brand:tavrenmention read at 40% confidence, below the 50% threshold, and it is a call that changes what to do next · google-ai-mode · ref
response:234|brand:tavrenmention read at 40% confidence, below the 50% threshold, and it is a call that changes what to do next · google-ai-mode · ref
response:237|brand:tavrenmention read at 40% confidence, below the 50% threshold, and it is a call that changes what to do next · chatgpt-web-search · ref
response:252|brand:tavrenmention read at 40% confidence, below the 50% threshold, and it is a call that changes what to do next · chatgpt-web-search · ref
response:253|brand:tavrenmention read at 40% confidence, below the 50% threshold, and it is a call that changes what to do next · chatgpt-web-search · ref
response:254|brand:tavrenmention read at 40% confidence, below the 50% threshold, and it is a call that changes what to do next · chatgpt-web-search · ref
response:255|brand:tavrenmention read at 40% confidence, below the 50% threshold, and it is a call that changes what to do next · chatgpt-web-search · ref
response:258|brand:tavrenmention read at 40% confidence, below the 50% threshold, and it is a call that changes what to do next · gemini-web · ref
response:267|brand:tavrenmention read at 40% confidence, below the 50% threshold, and it is a call that changes what to do next · gemini-web · ref
response:268|brand:tavrenmention read at 40% confidence, below the 50% threshold, and it is a call that changes what to do next · gemini-web · ref
response:269|brand:tavrenmention read at 40% confidence, below the 50% threshold, and it is a call that changes what to do next · gemini-web · ref
response:270|brand:tavrenmention read at 40% confidence, below the 50% threshold, and it is a call that changes what to do next · gemini-web · ref
response:271|brand:tavrenmention read at 40% confidence, below the 50% threshold, and it is a call that changes what to do next · chatgpt-web-search · ref
response:285|brand:tavrenmention read at 40% confidence, below the 50% threshold, and it is a call that changes what to do next · chatgpt-web-search · ref
response:286|brand:tavrenmention read at 40% confidence, below the 50% threshold, and it is a call that changes what to do next · gemini-web · ref
response:289|brand:tavrenmention read at 40% confidence, below the 50% threshold, and it is a call that changes what to do next · google-ai-mode · ref
response:290|brand:tavrenChoose this week's work
What to do this week
- Nothing to recommend from this scan meets the evidence checks. This is a finding about the evidence. Review any held recommendations.
Full plan and task ownership
Your to-do list
nothing to rankYou are being held on all 2 factor(s) this scan measured, so nothing is ranked. The 8 factor(s) below were raised by buyers and this scan asked nothing about them, so they cannot be ranked either.
Nothing is invented to fill this. A plan built from factors with no measured gap would be ordering work by buyer post count and presenting it as likely effect, which is the one thing this page must never do.
Raised by buyers, not measured by this scan
8These are not in the plan and must not be worked on yet. Buyers raise concerns continuously and a scan asks a frozen set of questions, so the two drift. We can see the demand and we have no reading of how assistants treat it — so any ranking of these would be ordering work by post count and calling it likely effect.
That is a gap in the instrument, not a finding about your product. The thing to change is the question set, not a page.
| distance | 11 buyer post(s) raise it | this scan asked no question about it |
| long | 8 buyer post(s) raise it | this scan asked no question about it |
| runner | 8 buyer post(s) raise it | this scan asked no question about it |
| speed | 8 buyer post(s) raise it | this scan asked no question about it |
| half | 7 buyer post(s) raise it | this scan asked no question about it |
| plated | 7 buyer post(s) raise it | this scan asked no question about it |
| ran | 7 buyer post(s) raise it | this scan asked no question about it |
| carbon | 4 buyer post(s) raise it | this scan asked no question about it |
Add questions on these factors and the next scan can measure them.
10 recommendation(s) held for review
Resolve these evidence gaps before assigning work.
- running: The extracted topic is not a specific, checkable product attribute.
- feel: The extracted topic is not a specific, checkable product attribute.
- distance: Too few measured answers to prioritise this change.; The extracted topic is not a specific, checkable product attribute.
- long: Too few measured answers to prioritise this change.; The extracted topic is not a specific, checkable product attribute.
- runner: Too few measured answers to prioritise this change.; The extracted topic is not a specific, checkable product attribute.
- speed: Too few measured answers to prioritise this change.; The extracted topic is not a specific, checkable product attribute.
- half: Too few measured answers to prioritise this change.; The extracted topic is not a specific, checkable product attribute.
- plated: Too few measured answers to prioritise this change.; The extracted topic is not a specific, checkable product attribute.
- ran: Too few measured answers to prioritise this change.; The extracted topic is not a specific, checkable product attribute.; No saved page passage or linked buyer quote supports the proposed edit.; A new claim needs linked supporting buyer evidence for review.
- carbon: Too few measured answers to prioritise this change.; The extracted topic is not a specific, checkable product attribute.
Visibility charts and supporting evidence
of the time an assistant names you — averaged over 32 question–assistant pairs (9 questions put to 4 assistants)
named youleft you outnot asked
Share of every brand named across the answers. Rank order, never reordered to put you first.
bigger → comes up on more of your questions
Open the detailed action plan and scoring
Why each job is ranked where it is
No decision factor in this scan has both a gap and a route to closing it. That is a finding about the evidence, not about the product.
Ground held
Named at or above your overall rate on these. Nothing to do — shown so a fall is visible next scan.
What this diagnosis does not cover
4 of 12 decision factor(s) had a question in this scan. 16 answer(s) sit outside every factor.
Raised by buyers and never asked about — this is a gap in the question set, not a finding about the product:
Questions were asked on men, Corvane, Aveltric, run, size, compared and no discovered factor matched them, so the diagnosis above does not speak for those answers.
Frozen demonstration screens from a saved scan. Use the report sections above or the menu inside the frame to explore them. 20 buyer questions put to 4 AI surfaces (Google AI Overviews, Google AI Mode, ChatGPT Search, Gemini); 80 answers stored and read back. Tavren is a fictional brand — this sample illustrates what a product scan could produce. Nothing here can be pressed into doing anything.