Read the Executive Diagnosis, then assign the work in the Action Brief.
Check the red and amber scorecard layers.
Execute the top First Fix Queue actions through the shared packages.
Use the Appendix for methodology, glossary and evidence.
1Management Decision
Overall
The report evaluates 24 fetched pages as a website system, not as isolated URLs.
Main visibility break
AI Journey Coverage is not sufficiently evidenced: AI Journey Coverage 39/100: the weakest measured surfaces are answer surface for AI questions 17/100 and category selection context 33/100; strongest measured surface: product decision surface 67/100.
Strongest signal
21 fetched page(s) expose structured data.
First priority
Citation Status (5/100) is the weakest outcome layer. It is not fixed directly: the first actionable levers are AI Journey Coverage, Product Page Readiness and Category Page Readiness. The First Fix Queue starts there, and improvement is validated against Citation Status movement in the next measurement.
Business meaning
8 high-priority finding(s) currently decide whether AI systems can use this website as an answer source. The work plan is sorted by that risk.
Decision summary · derived from this scan
Priority work orders
Architecture Clarity is 100/100 and should be preserved. The main actionable break is Category Page Readiness at 36/100: product and category surfaces do not yet explain enough selection, comparison, proof and purchase context. Citation Status 5/100 is the outcome to monitor after repairs, not a task the team can fix directly.
100Architecture Clarity
39AI Journey Coverage
5Citation Status
1
Add buyer-question blocks for choice, comparison, fit and purchase terms
A buying-decision block is a short, visible section in the product or category template that answers a real buyer question directly. Why: AI answers need quotable decision evidence on the shop page itself; otherwise they pull selection, comparison, delivery or warranty facts from marketplaces, review sites or competitors. Product example: "Will this model fit X?", "Which size/variant should I choose?", "Delivery, returns and warranty". Category example: "Which variant for use case A vs. B?", "Important selection criteria", "Comparison of subgroups". Use a blog only when the question cannot be answered well inside the template.
Owner: SEO + content + developer
Done when: Rerun Full AI Visibility Audit and confirm AI Journey Coverage becomes usable or strong.
2
Mark up only visible, real reviews
Only if real review data is visible: mark Review/AggregateRating in the product template. Do not create artificial reviews.
Owner: SEO + developer
Done when: Rerun the crawl and confirm Review Schema no longer fails or warns on the affected product pages URLs.
3
Add breadcrumb schema to the template
Add BreadcrumbList JSON-LD to the affected template and align it with visible navigation, canonical and category path.
Owner: SEO + developer
Done when: Rerun the crawl and confirm Breadcrumb Schema no longer fails or warns on the affected product pages URLs.
Remeasure the outcome afterwards: citation and prompt signals are not direct repair switches. Implement the actionable levers first, then rerun the same query set.
1BAction Brief
ACTION BRIEF · ASSIGN THE WORK
What the numbers mean — and what the team should do
This operating layer assigns 9 queue action(s) to shared work packages instead of turning every failed metric into a separate project.
Plain-language verdict
Architecture Clarity is 100/100 and should be preserved. The main actionable break is Category Page Readiness at 36/100: product and category surfaces do not yet explain enough selection, comparison, proof and purchase context. Citation Status 5/100 is the outcome to monitor after repairs, not a task the team can fix directly.
Build one shared answer component. Start with questions evidenced in the audit: selection, meaningful differences, proof, suitability boundaries, delivery and returns. Claims must match visible, approved product information.
2 · Category selection guide · 2 action(s)
Owner: Merchandising/content + developer
Explain purpose, variants, decision criteria and important subgroups before the grid. Add missing category metadata from the same approved positioning; do not write filler merely to reach a word count.
3 · Catalog schema · 3 action(s)
Owner: Developer + SEO validation
Implement BreadcrumbList once across product and category templates; add CollectionPage/ItemList to categories. Visible navigation, canonicals, product lists and JSON-LD must describe the same facts.
4 · Reviews and proof · 1 action(s)
Owner: Product/customer team + developer
Mark only real, visible reviews. If none exist, omit Review/AggregateRating. Keep relevant documents, seller identity and policies visible; never create hidden or invented evidence.
Measured topics/questions: metilen mavisi USP grade · metilen mavisi satın al · nattokinaz kapsül
Assignment and completion checks
Package
Deliverable
Done when
Buyer-answer blocks
Reusable, visible answers on affected product and category templates.
Answer Structure passes on sampled URLs; answers are direct, evidence-bound and approved by the responsible product team.
Category selection guide
Selection context before the product grid plus an accurate page summary.
A buyer can choose without opening a separate blog; category context, Answer Structure and missing-meta checks pass.
Catalog schema
Shared BreadcrumbList and category CollectionPage/ItemList implementation.
Schema validation passes and visible paths, canonicals, product lists and JSON-LD agree.
Reviews and proof
Visible evidence with matching markup, or a documented decision not to add review schema.
Markup matches the exact visible count and rating; no invented, hidden or site-wide product rating is used.
Implementation sequence and decision gate
1Specify and stage
Start with one product and one category template; save current URLs and scores as the baseline.
2Validate and roll out
Check visible content, compliance, canonicals and schema before broad rollout.
3Rerun and decide
If template criteria pass but citations do not move, plan stronger source-worthy evidence. If criteria still fail, repair implementation first.
Avoid generic FAQ filler, fabricated reviews or evidence, vague GEO tickets, and new pages justified only by a missing citation. Every opportunity needs a documented editorial disposition; sensitive topics need qualified specialist review.
Metric definitions are in the Appendix so this section remains an action brief.
SEO Foundation scores 62/100; the weakest measured components are Heading structure and Single H1. Evidence band: usable with visible gaps.
Weakest weighted components:
Heading structureweight 10% · contribution 3.5
Single H1weight 8% · contribution 2.8
Meta descriptionweight 8% · contribution 2.8
Full breakdown in appendix
GEO Signal Layer
62/100
GEO Signal Layer scores 62/100; the weakest measured components are Conversational tone and Citation worthiness. Evidence band: usable with visible gaps.
Weakest weighted components:
Conversational toneweight 7% · contribution 0
Citation worthinessweight 9% · contribution 3.2
Direct answersweight 7% · contribution 2.4
Full breakdown in appendix
AI Quality
70/100
AI Quality scores 70/100; the weakest measured components are Tone and Structure. Evidence band: usable with visible gaps.
Architecture clarity is derived from the share of semantic vs opaque URL patterns.
Product Page Readiness
67/100
Product Page Readiness scores 67/100; the weakest measured components are Review Schema and Breadcrumb Schema. Evidence band: usable with visible gaps.
Weakest weighted components:
Review Schemaweight 1%
Breadcrumb Schemaweight 1%
Answer Structureweight 1%
Category Page Readiness
36/100
Category Page Readiness scores 36/100; the weakest measured components are Collection JSON-LD and Breadcrumb Schema. Evidence band: structurally weak.
Weakest weighted components:
Collection JSON-LDweight 1%
Breadcrumb Schemaweight 1%
Intro Text ≥80 wordsweight 1%
AI Journey Coverage
39/100
AI Journey Coverage 39/100: the weakest measured surfaces are answer surface for AI questions 17/100 and category selection context 33/100; strongest measured surface: product decision surface 67/100.
Citation Status
5/100
AI Share of Voice 7% (share of tested AI answers where the domain appears as a source), mention rate 0% across 3 platform(s).
AEO / Agent Readiness
38/100
Can AI agents read contact, services, actions and freshness? Score 38.
Full breakdown in appendix
Prompt Discovery
50/100
10 sector AI questions tested. Coverage 50%, mention 20%.
Technical GEO
Technical GEO was not measured as a separate score in this report.
AI-Agent Readiness
agents.md, llms.txt, llms-full.txt, Universal Commerce Protocol endpoint detected. The site exposes machine-readable guidance for AI agents and crawlers. agents.md: platform-provided (Shopify). AI crawlers allowed: GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, anthropic-ai, Claude-Web +6
Sector Benchmark
Sector baseline, not a site score: 6 top results measured for "metilen mavisi USP grade". Avg word count 532.2, schema coverage 66.7%.
3Priorities: What Must Be Fixed First
Compact triage: sorted by severity, with only the short measurement signal. Full evidence and visuals stay in the matching report chapter.
No.
Priority
Topic
Finding
Short signal
Visual
1
High
AI Journey
AI Journey Coverage is not sufficiently evidenced
Product Pages: 3
-
2
High
Product Schema
Product Pages: Review Schema
Review Schema: 0/3 / 0%
3
High
Product Schema
Product Pages: Breadcrumb Schema
Breadcrumb Schema: 0/3 / 0%
4
High
Product Schema
Product Pages: Answer Structure
Answer Structure: 1/3 / 33%
5
High
Category Context
Category Pages: Collection JSON-LD
Collection JSON-LD: 0/2 / 0%
6
High
Category Context
Category Pages: Breadcrumb Schema
Breadcrumb Schema: 0/2 / 0%
7
High
Category Context
Category Pages: Intro Text ≥80 words
Intro Text ≥80 words: 0/2 / 0%
8
High
Category Context
Category Pages: Answer Structure
Answer Structure: 0/2 / 0%
9
Medium
Category Context
Category Pages: Meta Description
Meta present: 1/2 / 50%
10
Low
Crawl Inventory
Crawl inventory defines the evidence boundary
Discovered: 26
-
11
Low
Finding
AI-Agent Readiness
agents.md: https://www.ts-wellness.com/agents.md
-
12
Low
URL Semantics
URL semantics affect page-role clarity
Semantic: 26/26 - 100%
13
Low
Slug Variants
Slug variant clusters can inflate catalog size
Product URLs: 3
-
14
Low
Schema Coverage
Structured data coverage frames machine-readable context
Pages with schema: 21/24 / 88%
4Detailed Analysis and Evidence
4Detailed analysis
Crawl Inventory
Crawl Inventory
Defines the evidence boundary of the report.
Summary
26 URLs discovered, 24 pages fetched, 2 URLs recorded only.
Evidence boundary
Diagnosis
No crawl cap was hit. 2 URLs were deliberately recorded without fetching: the crawl loads a representative selection per template family (variant URLs are skipped) and keeps the remaining URLs as inventory. Findings apply to the fetched pages.
Evidence
URLs discovered: 26
Pages fetched: 24
Recorded only: 2
4Detailed analysis
Architecture
Architecture Findings
Explains how site structure affects machine readability.
Summary
3 Content Page(s) are classified as CMS/content pages. 0% opaque URL ratio. 0% slug variant inflation.
AI-Agent Readiness
01AEO Signal
What this signal is: Checks whether the site publishes machine-readable instruction files for AI agents — agents.md, llms.txt, llms-full.txt, a Universal Commerce Protocol endpoint and an AI-crawler policy in robots.txt. These files tell AI assistants and shopping agents what the site offers, which actions are safe and how to interact with it; without them, an agent has to guess everything from raw HTML.
Diagnosis
agents.md is a positive signal: Machine-readable instructions for AI agents and agentic commerce flows. agents.md: platform-provided (Shopify). AI crawlers allowed: GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, anthropic-ai, Claude-Web +6
Helps AI agents understand site capabilities, safe actions and preferred interaction paths before using the site.
Action
Recommended change
Keep it and update it when offers, checkout paths or policies change.
Semantic URL Check
02SEO Signal
What this signal is: Measures how many crawled URLs are human-readable and descriptive (semantic) versus built from IDs, hashes or parameters (opaque). Search engines and AI systems use the URL slug as an early clue to what a page is about — opaque URLs shift the entire burden of page-role identification onto schema and headings.
Diagnosis
0% opaque URL ratio in the provided evidence.
Interpretation
Opaque URLs weaken page-role clarity and increase dependence on schema and headings.
Evidence
Semantic ratio: 100%
Mixed ratio: 0%
Opaque ratio: 0%
Action
Recommended change
Clarify URL patterns or strengthen template-specific schema and headings.
Slug Variant Inflation
03SEO Signal
What this signal is: Detects clusters of near-identical URLs growing out of the same base slug — color/size variants, tracking parameters, session paths. Inflated variant sets waste crawl budget, split ranking and citation signals across duplicates, and make the catalog look larger but less trustworthy to machine readers than it really is.
Diagnosis
0 inflated URL(s) across 3 stem(s).
Interpretation
Variant inflation can make a site look larger while weakening page-role clarity.
Evidence
Inflated count: 0
Variant inflation ratio: 0%
Action
Recommended change
Separate parent URLs and variant URLs clearly, and expose the relationship in schema/canonical logic.
4Detailed analysis
SEO Foundation
SEO Foundation
Site-wide SEO foundation criteria measured across the fetched pages — each with measurement, interpretation and first fix.
Summary
3 weighted criteria did not fully pass; fully passed criteria are listed in the Score Methodology appendix.
Keyword positioning
04SEO Signal
What this signal is: Where the target term sits across five strategic positions: title, H1, meta description, first 100 words, URL slug. Density itself only guards against stuffing.
Diagnosis
critical
Interpretation
Ranking and extraction in 2026 work on position signals, not repetition counts. A term missing from title and H1 makes every other signal work harder.
Evidence
Measurement: “metilen”: 1/5 in the keyword-positioning/over-optimization check; status: critical.
Action
Recommended change
Place the term in the missing core positions (title, H1, meta description); do not increase repetition in the body text.
Keyword in first 100 words
05SEO Signal
What this signal is: Whether the target terms appear within the first ~100 words, where crawlers and answer systems weight topical signals most.
Diagnosis
partial
Interpretation
The opening defines what the page is about. If the core term arrives late, machines anchor the page to the wrong topic or to none.
Evidence
Measurement: 5/20 keywords appear in the first 100 words; status: partial.
Action
Recommended change
Rewrite the opening paragraph around the target term and the page promise — naturally, one clear mention is enough.
Paragraph length
06SEO Signal
What this signal is: Whether paragraphs stay short enough to be read and extracted as single units.
Diagnosis
partial
Interpretation
Walls of text hide the answer sentence. Short paragraphs map one idea to one extractable block.
Evidence
Measurement: Short-paragraph ratio was not forwarded; deterministic neutral partial credit applied.
Action
Recommended change
Split paragraphs that run past ~5 lines, one idea each, and lead with the conclusion sentence.
4Detailed analysis
GEO Signal Layer
GEO Signal Layer
Reads whether site content is extractable, attributable and answer-ready.
Summary
GEO Signal Layer scores 62/100; evidence band: usable with visible gaps.
Commercial facts: pricing disclosure and access
07GEO Signal
What this signal is: Checks whether commercial pages expose usable pricing facts or explained pricing logic directly in HTML and whether pricing sources remain reachable.
Diagnosis
3/3 detected commercial pages disclose a usable price or explained pricing logic.
Interpretation
When pricing facts are missing or only indirectly reachable, agents must infer the answer or fall back to third-party sources.
Add the two or three natural next questions — cost, compatibility, alternatives, process — as short sections.
Social signals
11GEO Signal
What this signal is: Whether the page or site exposes connected social and profile surfaces that help entity reconciliation.
Diagnosis
partial
Interpretation
Profiles do not prove authority alone. They help machines connect the website to a recognizable entity across the web.
Evidence
Measurement: Only 2 social platforms detected: instagram, twitter. Aim for 3+ for stronger signals.
Action
Recommended change
Link the active official profiles from the site and keep names, logos, addresses, and descriptions consistent.
Conversational structure
12GEO Signal
What this signal is: Whether the content flows in a question-answer and explanation rhythm rather than brochure prose.
Diagnosis
partial
Interpretation
Answer systems reuse passages that already work as dialogue turns. Brochure copy needs rewriting before it fits an answer — so it usually gets skipped.
What this signal is: Checks whether each fetched product page embeds a schema.org Product object in JSON-LD — name, description, image, brand and identifiers. This block is the primary machine-readable statement of what the product is; AI answer engines rely on it to cite the product with correct facts instead of guessing from free text.
Diagnosis
3/3 pass, 100% coverage.
Evidence
Status: stable
Measurement
3/3 / 100%
Product JSON-LD
Product JSON-LD
Action
Recommended change
Keep Product JSON-LD and recheck it against visible product data after theme or app changes.
Machine-readable Price Data
15GEO Signal
What this signal is: Verifies that price, currency and availability are exposed as structured Offer data (price, priceCurrency, availability), not only as display text. Assistants and shopping agents can only compare or recommend a product whose price they can read programmatically — text-only prices are routinely misread or dropped from answers.
Diagnosis
3/3 have machine-readable Offer data; 3/3 show a visible price.
Evidence
Status: stable
Measurement
3/3 / 100%
Visible price
3/3 / 100%
Machine-readable Offer
Visible price
Machine-readable Offer
Action
Recommended change
Keep Offer data and continue syncing price, currency, availability and URL with the visible page.
Review Schema
16GEO Signal
What this signal is: Checks for structured rating and review markup (AggregateRating / Review) on product pages. Ratings are one of the strongest trust signals AI systems attach to a product recommendation; without the markup, customer reviews that exist on the page stay invisible to the answer layer.
Diagnosis
0/3 pass, 0% coverage.
Evidence
Status: critical
Top fail URLs: 3 example URL(s) in the dropdown; full list in Evidence JSON.
Only if real review data is visible: mark Review/AggregateRating in the product template.
Detail
Do not create artificial reviews.
Breadcrumb Schema
17SEO Signal
What this signal is: Checks for BreadcrumbList JSON-LD stating where the product sits in the catalog hierarchy (shop → category → product). Breadcrumb data hands crawlers the site’s category tree in machine-readable form and anchors each product to its category context, which supports both rankings and AI category understanding.
Diagnosis
0/3 pass, 0% coverage.
Evidence
Status: critical
Top fail URLs: 3 example URL(s) in the dropdown; full list in Evidence JSON.
Add BreadcrumbList JSON-LD to the affected template and align it with visible navigation, canonical and category path.
Canonical Declared
18SEO Signal
What this signal is: Verifies that every product page declares a canonical URL. The canonical tag tells crawlers which URL is the authoritative version of this content, preventing variant, parameter and duplicate URLs from splitting indexing and citation signals across several addresses.
Diagnosis
3/3 pass, 100% coverage.
Evidence
Status: stable
Measurement
3/3 / 100%
Canonical Declared
Canonical Declared
Action
Recommended change
Keep the signal: Canonical Declared currently passes on the product template; recheck after template changes.
Meta Description
19SEO Signal
What this signal is: Checks whether the product page ships a meta description of usable length and substance. It is frequently the snippet search and answer engines reuse to summarize the page — a missing or trivial description leaves that summary entirely to the machine.
Diagnosis
3/3 have a meta description; 100% are unique (0 duplicate).
Evidence
Status: stable
Measurement
3/3 / 100%
Meta present
3/3 / 100%
Unique
Meta present
Unique
Site-wide measurement
SEO Foundation · Meta description: partial — Meta description length: 193 characters; target range 120–160. Status: partial.
First fix: Write one specific 140-160 character description that names the topic, the buyer/user intent, and the concrete outcome of the page.
Action
Recommended change
Keep unique meta descriptions; do not publish new product or category pages with duplicate copy.
Image Alt ≥ 80%
20SEO Signal
What this signal is: Measures the share of product images carrying meaningful alt text; the pass threshold is 80%. Alt text is the only way non-visual systems “see” images — it feeds image search and accessibility, and gives AI models additional product facts they cannot get from pixels.
Diagnosis
3/3 pass, 100% coverage.
Evidence
Status: stable
Measurement
3/3 / 100%
Image Alt ≥ 80%
Image Alt ≥ 80%
Site-wide measurement
SEO Foundation · Image alt text: partial — Image alt-text coverage: 85%; status: partial.
First fix: Describe what the image actually shows — product plus distinguishing attribute; avoid keyword stuffing.
Action
Recommended change
Keep the alt-text rule: continue describing main image and gallery from product name plus a distinguishing attribute.
Answer Structure
21GEO Signal
What this signal is: Checks whether the product page contains directly quotable answer units — question-style headings, short factual paragraphs, FAQ blocks, definition-first sections. AI engines lift answers verbatim; pages without extractable answer units get read but rarely cited.
Diagnosis
1/3 pass, 33% coverage.
Evidence
Status: critical
Top fail URLs: 2 example URL(s) in the dropdown; full list in Evidence JSON.
SEO Foundation · Direct-answer snippet: partial — BLUF 33%: some headings provide direct answers, but snippet formatting is inconsistent.
First fix: After key H2/H3 headings, place a 40-70 word answer paragraph before examples, caveats, or sales copy.
Site-wide measurement
GEO Signal Layer · Direct answers: partial — BLUF (answer-first) ratio: 33%; direct-answer coverage is partial. Stored guidance: 10-80 words per answer paragraph.
First fix: Convert the main user questions into H2/H3 blocks and answer each one in the first sentence of the section.
Action
Recommended change
Add short answer blocks to the template: compatibility, selection, use, delivery/warranty or real buyer questions.
What this signal is: Checks whether category pages embed CollectionPage / ItemList structured data listing the products they contain. This gives AI systems a machine-readable inventory of the assortment — without it, a category page is just a wall of links with no declared meaning.
Diagnosis
0/2 pass, 0% coverage.
Evidence
Status: critical
Top fail URLs: 2 example URL(s) in the dropdown; full list in Evidence JSON.
Add CollectionPage or ItemList to the category template and align it with the visible product list.
Breadcrumb Schema
23SEO Signal
What this signal is: Checks for BreadcrumbList JSON-LD on category pages, placing each category inside the shop’s navigation tree. This is how machines learn the catalog’s structure — which categories exist, how they nest, and which page is the entry point for each product family.
Diagnosis
0/2 pass, 0% coverage.
Evidence
Status: critical
Top fail URLs: 2 example URL(s) in the dropdown; full list in Evidence JSON.
Add BreadcrumbList JSON-LD to the affected template and align it with visible navigation, canonical and category path.
Canonical Declared
24SEO Signal
What this signal is: Verifies that category pages declare a canonical URL. Category pages generate many URL variants through filters, sorting and pagination; the canonical keeps all of those signals consolidated on one authoritative address.
Diagnosis
2/2 pass, 100% coverage.
Evidence
Status: stable
Measurement
2/2 / 100%
Canonical Declared
Canonical Declared
Action
Recommended change
Keep the signal: Canonical Declared currently passes on the category template; recheck after template changes.
Meta Description
25SEO Signal
What this signal is: Checks whether the category page carries a meta description that actually describes the assortment. For category pages this is often the only editorial summary machines get — it shapes the search snippet and the way answer engines characterize what this part of the shop sells.
Diagnosis
1/2 have a meta description; 1 page(s) are missing one. Existing descriptions are 100% unique (0 duplicate).
Evidence
Status: partial
Top fail URLs: 1 example URL(s) in the dropdown; full list in Evidence JSON.
SEO Foundation · Meta description: partial — Meta description length: 193 characters; target range 120–160. Status: partial.
Action
Recommended change
Add meta descriptions to the 1 missing page(s): use the page/category name plus a distinguishing attribute.
H1 Present
26SEO Signal
What this signal is: Verifies each category page has one clear H1 naming the category. The H1 is the strongest on-page statement of the page’s topic; a missing or duplicated H1 forces crawlers and AI models to infer the topic from weaker, more ambiguous signals.
Diagnosis
2/2 pass, 100% coverage.
Evidence
Status: stable
Measurement
2/2 / 100%
H1 Present
H1 Present
Site-wide measurement
SEO Foundation · Single H1: partial — Average 0.63 H1 headings per page; each page should contain exactly one H1. Status: partial.
First fix: Keep one H1 for the page promise, then move secondary ideas into H2/H3 headings.
Action
Recommended change
Keep the signal: H1 Present currently passes on the category template; recheck after template changes.
Intro Text ≥80 words
27GEO Signal
What this signal is: Checks that the category page opens with at least 80 words of descriptive text about the assortment — what it contains, for whom, and how to choose. This intro is usually the only extractable prose on a category page; it is what an AI engine can quote when asked what this shop offers in that category.
Diagnosis
0/2 pass, 0% coverage.
Evidence
Status: critical
Top fail URLs: 2 example URL(s) in the dropdown; full list in Evidence JSON.
Add short selection context to category pages: purpose, variants, selection criteria and important subgroups.
Answer Structure
28GEO Signal
What this signal is: Checks whether the category page offers extractable answer units — selection guidance, comparison hints, question-formatted headings or FAQ blocks. These structures let AI assistants use the category page for “which one should I pick” questions instead of only as a link list.
Diagnosis
0/2 pass, 0% coverage.
Evidence
Status: critical
Top fail URLs: 2 example URL(s) in the dropdown; full list in Evidence JSON.
SEO Foundation · Direct-answer snippet: partial — BLUF 33%: some headings provide direct answers, but snippet formatting is inconsistent.
Site-wide measurement
GEO Signal Layer · Direct answers: partial — BLUF (answer-first) ratio: 33%; direct-answer coverage is partial. Stored guidance: 10-80 words per answer paragraph.
Action
Recommended change
Add short answer blocks to the template: compatibility, selection, use, delivery/warranty or real buyer questions.
4Detailed analysis
AI Journey
AI Journey Coverage
Connects product and category evidence to the AI visibility question: is there enough answer, selection and buying-decision surface?
Summary
AI Journey Coverage 39/100: the weakest measured surfaces are answer surface for AI questions 17/100 and category selection context 33/100; strongest measured surface: product decision surface 67/100.
BOFU product decision surface
29AI Journey Signal
What this signal is: Measures whether product pages carry the decision-stage (bottom-of-funnel) facts a buyer — or an AI assistant advising one — needs to say yes: specifications, price, availability, trust and review data in extractable form. Weakness here means losing exactly the moment where a recommendation would turn into a purchase.
Diagnosis
67/100
Evidence
Product Pages: 3
Product/Offer/Review: 67/100
Action
What to add
make the product page close the buying decision on the page itself.
Why
AI shopping and answer systems need quotable facts for price, availability, trust and fit; without them, the recommendation is built from other sources.
Example
visible price plus Offer data, availability, real reviews/warranty, and a short block such as "Will this product fit X?" or "Which variant should I choose?".
Done when
Product/Offer/Review and the main buying questions pass visibly and machine-readably in the product template.
MOFU category and selection surface
30AI Journey Signal
What this signal is: Measures whether category pages support the comparison and selection stage (middle of funnel): assortment overview, selection criteria and structured product lists. This is the surface on which AI assistants decide which of the shop’s products make it onto a shortlist.
Diagnosis
33/100
Evidence
Category Pages: 2
Category context: 33/100
Action
What to add
make the category behave like a selection guide, not only a product grid.
Why
AI systems cannot easily cite a list; they need explained selection context before they can use the category as an answer source.
Example
a short intro block explains purpose, variants, important subgroups and criteria such as material, size, configuration, use case or compatibility.
Done when
H1, CollectionPage/ItemList and selection text clearly show which variant fits which need.
Answer surface for AI questions
31AI Journey Signal
What this signal is: Measures whether the site answers informational buyer questions (top of funnel) with extractable content — FAQs, guides, definition blocks tied to real query language. This surface earns the early mentions in AI answers, before a user ever asks for a specific product.
Diagnosis
17/100
Evidence
Product answer structure: 33/100
Category answer structure: 0/100
Action
What to add
give each recurring buyer question a short, directly quotable answer surface.
Why
if the answer is not in the shop, an AI system will take it from guides, marketplaces, forums or competitors.
Example questions
"What is the difference between A and B?", "How is it used?", "What are the risks or limits?", "What is it compatible with?".
Done when
product and category templates answer these questions visibly and the Answer Structure measurement improves.
4Detailed analysis
Citation Status
Citation Evidence (Add-on)
Point-in-time / snapshot: Live-snapshot layer: AI Citation and Prompt Discovery are point-in-time measurements against live external systems (Perplexity / ChatGPT / Google AIO) (2026-06-26); they use a fixed query set, but external answers can change day to day.
Do AI systems mention or cite the domain on buyer questions?
Citation means the domain is used as a source. Mention means the domain is named, but not used as a source.
Multi-ASOV 7% means here: ts-wellness.com was a source in 1 of 15 query-platform results.
"No sources returned" counts as absent. It is an evidence boundary of the AI answer, not an automatic site defect.
Interpretation: this snapshot is directional evidence. The first fix queue lists hypotheses to validate; a re-audit measures what changed without assigning causality to any single edit.
Perplexity
Diagnosis
ASOV 20%, mention 0%.
ChatGPT
Diagnosis
ASOV 0%, mention 0%.
Google AI Overview
Diagnosis
ASOV 0%, mention 0%.
4Detailed analysis
Finding
Citation Status — Query Drilldown
Point-in-time / snapshot: Live-snapshot layer: AI Citation and Prompt Discovery are point-in-time measurements against live external systems (Perplexity / ChatGPT / Google AIO) (2026-06-26); they use a fixed query set, but external answers can change day to day.
Shows which tested questions produced citation, mention or absence on each platform.
Buyer-question language: Turkish. Questions remain in the tested language as measured evidence; the report language is English.
Summary
15 query-platform result(s) from the citation add-on.
How to use this table
The "sources went to" domains are your real AI competition - not necessarily your classic competitors.
Questions where specialist sources are cited instead of marketplaces are the fastest to win; answer them in answer blocks on the matching product/category page.
Query
Perplexity
ChatGPT
Google AI Overview
metilen mavisi USP grade
absent
absent
absent
metilen mavisi satın al
absent
absent
absent
nattokinaz kapsül
absent
absent
absent
CoA sertifikalı metilen mavisi
cited
absent
absent
metilen mavisi nedir
absent
absent
absent
Legend: cited = domain is a source; mentioned = domain is named; absent = no citation/mention; - no sources counts as absent, but is not a site defect.
metilen mavisi USP grade
Diagnosis
0 cited, 0 mention-only, 3 absent.
Interpretation
Frequently cited domains for this question: hepsiburada.com, aquarius-prolife.com, trendyol.com, instagram.com (Social Media / Platforms).
Evidence
Perplexity: absent — sources went to: hepsiburada.com, aquarius-prolife.com, instagram.com, trendyol.com
ChatGPT: no sources returned in this answer
Google AI Overview: no sources returned in this answer
metilen mavisi satın al
Diagnosis
0 cited, 0 mention-only, 3 absent.
Interpretation
Frequently cited domains for this question: cimri.com, kimyaciniz.com, hepsiburada.com, trendyol.com.
Evidence
Perplexity: absent — sources went to: cimri.com, kimyaciniz.com, hepsiburada.com, trendyol.com
ChatGPT: absent — sources went to: destekkimya.com, labor.com.tr, okulgen.com, egenanotek.com
Google AI Overview: no sources returned in this answer
nattokinaz kapsül
Diagnosis
0 cited, 0 mention-only, 3 absent.
Interpretation
Frequently cited domains for this question: velavit.com, trendyol.com, hepsiburada.com, redfood24.de.
Evidence
Perplexity: absent — sources went to: velavit.com, trendyol.com, hepsiburada.com, redfood24.de
ChatGPT: no sources returned in this answer
Google AI Overview: no sources returned in this answer
CoA sertifikalı metilen mavisi
Diagnosis
1 cited, 0 mention-only, 2 absent.
Interpretation
Frequently cited domains for this question: metilen-mavisi.com, laborteknik.com, ts-wellness.com, drozdogan.com.
What this signal is: Checks whether an AI agent can programmatically extract who operates the site and how to reach them — address, e-mail, contact channels, imprint and Organization schema. Agents rank operators they can identify and verify above anonymous ones, and contact data is a precondition for any agent-initiated transaction.
Diagnosis
Contact extractability: critical (20/100). Contact path and ContactPoint signals are evaluated as the AEO contact layer. Evidence page: /pages/iletisim
Service/product readability
33AEO Signal
What this signal is: Rates how well the offer itself can be machine-read: whether what is sold, its properties, conditions and scope are available in structured or clearly-formatted form rather than only in marketing prose. This decides whether an agent can restate the offer accurately.
Diagnosis
Servis/ürünler fiyat verisiyle tamamen yapılandırılmış (name, description, url, price, availability) — AI ajanları teklifleri programatik karşılaştırabilir. Evidence page: /products/metilen-mavisi-usp-grade-100-ml
Action flow accessibility
34AEO Signal
What this signal is: Rates whether the key actions — buying, booking, contacting — are reachable through machine-followable paths: clean links, working forms, structured checkout entry points instead of script-only UI. Agentic commerce fails at exactly this step when the action path is invisible to the agent.
Diagnosis
Action flow accessibility: critical (30/100). The visible next step and machine-readable actions determine agent accessibility. Evidence page: /products/metilen-mavisi-usp-grade-100-ml
Data freshness
35AEO Signal
What this signal is: Checks for machine-readable recency signals — dateModified, price validity, stock status, sitemap lastmod. Agents discount data they cannot date; facts that look stale get excluded from answers even when they are still correct.
Diagnosis
Data freshness: critical (20/100). Freshness and verifiability signals show whether agents can judge the state of the data. Evidence page: /products/metilen-mavisi-usp-grade-100-ml
4Detailed analysis
Sector Benchmark
Sector Benchmark (Add-on)
Special benchmark layer for metilen mavisi USP grade: it shows not only averages, but how Mundus turns the competitive field into a measurable AI-search baseline.
How To Use This Benchmark
This is not an extra score. This layer turns the sector comparison into work decisions: which market signals AI answer systems can recognize, select, and reuse.
Read market signals
Use the top-result pages to see which content depth, schema coverage, answer format, freshness, and extractable facts are expected in this query field.
Choose fix direction
Compare the audited site against the sector baseline and decide which below-baseline signals should enter the fix queue first.
Re-measure the gap
After fixes, run the same benchmark again: did the gap close, and did the page become stronger answer material?
Recommended Fix Direction
AI extractability
Start with AI extractability because these benchmark cards show the gap: Proper H1 Hierarchy, Synthesizability Score.
Use the H1, schema, synthesizability, metadata and structure cards. Make the answer reusable through clear headings, structured data and explicit facts.
6 results analyzed. The benchmark remains sector context, but it should be read as a decision field: which signals are visible in the market, where the audited site sits below median, and which gap should be re-measured after the next fix?
How to read the Sector Benchmark
Mundus builds this layer from the live query field for "metilen mavisi USP grade": 6 top-result pages become the market baseline, then the audited site is compared against it.
AI search is no longer a simple ranking check. Systems split questions into subtopics, select only a small set of sources, and can cite pages that do not map one-to-one to classic search results.
This is why the section is a work compass: it shows which content depth, answer structure, schema coverage, and extraction signals are visible in the sector, and which gaps should be re-measured first.
Short summary
Diagnosis
12 of 17 comparison metrics are at sector level or better.
Interpretation
Worse than sector: Proper H1 Hierarchy, Synthesizability Score. Not measured: Readability Score, Question Ratio (%), FAQ Headings.
Action
Recommended change
Start with AI extractability.
Detail
Use these cards below: Proper H1 Hierarchy, Synthesizability Score.
Detail
Word Count is only the starting point when it is named as a gap.
Proper H1 Hierarchy
36Benchmark Signal
What this signal is: Compares the page’s heading hierarchy — exactly one H1, cleanly ordered H2/H3 levels — against the measured top results for the target query. A clean hierarchy is how machines segment content into sections they can cite individually.
Diagnosis
Current: absent; sector: 50% of top pages comply; interpretation: absent.
Evidence
Current: absent
Sector: 50% of top pages comply
Interpretation: absent
Site-wide measurement
SEO Foundation · Heading structure: partial — The measured H1→H2/H3 hierarchy is partial; H2 average 8.63 per page.
First fix: Rewrite headings as a table of contents: one H1, descriptive H2 sections, and H3 detail blocks only where needed.
Site-wide measurement
SEO Foundation · Single H1: partial — Average 0.63 H1 headings per page; each page should contain exactly one H1. Status: partial.
Action
Recommended change
AI extractability: Use the H1, schema, synthesizability, metadata and structure cards.
Detail
Make the answer reusable through clear headings, structured data and explicit facts.
Synthesizability Score
37Benchmark Signal
What this signal is: Mundus’ measure of how easily an AI model can compress the page into an answer: fact density, self-contained paragraphs, low filler. Benchmarked against competing pages for the same query — the more synthesizable side is the one that gets paraphrased into the AI answer.
GEO Signal Layer · Synthesizability: partial — Synthesizability score: 30/100; other measured values 0. Status: partial.
First fix: Turn core claims into lists, tables, definitions, and step blocks. Each block should be usable without reading the whole page.
Action
Recommended change
AI extractability: Use the H1, schema, synthesizability, metadata and structure cards.
Detail
Make the answer reusable through clear headings, structured data and explicit facts.
Readability Score
38Benchmark Signal
What this signal is: Standard readability scoring (sentence length, word complexity) compared with the sector’s top results. Readable text is also machine-friendly text — unusually complex prose correlates with lower extraction and citation rates.
Diagnosis
Current: not measured; sector: 7.7; interpretation: not measured.
Evidence
Current: not measured
Sector: 7.7
Interpretation: not measured
Action
Recommended change
Make this card measurable first if it belongs to the chosen fix direction; then rerun the same benchmark.
Question Ratio (%)
39Benchmark Signal
What this signal is: The share of headings and sections phrased as questions, compared with the top results. Question-formatted sections map almost one-to-one onto real user prompts, which makes them preferred lift-out targets for AI answers.
Diagnosis
Current: not measured; sector: 0; interpretation: not measured.
Make this card measurable first if it belongs to the chosen fix direction; then rerun the same benchmark.
FAQ Headings
40Benchmark Signal
What this signal is: Counts explicit FAQ blocks and headings against the competition. FAQ format is the single most reliably extracted content pattern across AI engines — each Q&A pair is a ready-made answer unit.
Diagnosis
Current: not measured; sector: 0.3; interpretation: not measured.
First fix: Add a real FAQ section with the questions buyers/users ask, then mark it up only if the visible content matches the schema.
Action
Recommended change
Make this card measurable first if it belongs to the chosen fix direction; then rerun the same benchmark.
Word Count
41Benchmark Signal
What this signal is: Total content length versus the measured top results for the query. It indicates the depth of coverage the sector’s winning pages deliver — the goal is to close large gaps in substance, not to copy the average number blindly.
SEO Foundation · Content length: partial — 746 words per page. Status: partial.
First fix: Extend thin templates with the selection context, answers and product detail users actually need; do not add filler.
Action
Recommended change
Do not start by adding volume while this page is at or above sector level.
Detail
Add missing subtopics instead of filler.
Internal Links
42Benchmark Signal
What this signal is: Internal link count compared with the sector’s top results. Internal links define the crawl paths and topical relationships machines use to work out which pages of a site matter and how its topics connect.
Keep this signal stable; the first fix belongs to the cards named as weak or not measured in the short summary.
External Links
43Benchmark Signal
What this signal is: Outbound reference links versus the competition. Citing verifiable sources is a trust cue in the E-E-A-T sense — pages that reference sources are treated as better-sourced claims, which supports citation-worthiness.
Keep this signal stable; the first fix belongs to the cards named as weak or not measured in the short summary.
Alt Tag Coverage (%)
44Benchmark Signal
What this signal is: The share of images with alt text, compared with the sector’s top results. Beyond accessibility, alt text is extra machine-readable context — competitors with higher coverage hand AI systems more usable facts per page.
SEO Foundation · Image alt text: partial — Image alt-text coverage: 85%; status: partial.
Action
Recommended change
Keep this signal stable; the first fix belongs to the cards named as weak or not measured in the short summary.
Passive Voice (%)
45Benchmark Signal
What this signal is: The share of passive-voice sentences versus the competition. Active, subject-first sentences state facts the way answer engines like to quote them; heavy passive use blurs who does what and lowers quotability.
Diagnosis
Current: 0; sector: 0.2; interpretation: at sector level.
Evidence
Current: 0
Sector: 0.2
Interpretation: at sector level
Delta: -0.2
Action
Recommended change
Keep this signal stable; the first fix belongs to the cards named as weak or not measured in the short summary.
Meta Description Length
46Benchmark Signal
What this signal is: Meta description length versus the sector’s top results — checking both presence and usable snippet length (~150–160 characters). The meta description is the page’s default summary in search results and some answer surfaces.
SEO Foundation · Meta description: partial — Meta description length: 193 characters; target range 120–160. Status: partial.
Action
Recommended change
Keep this signal stable; the first fix belongs to the cards named as weak or not measured in the short summary.
4Detailed analysis
Prompt Discovery
Prompt Discovery (Add-on)
Point-in-time / snapshot: Live-snapshot layer: AI Citation and Prompt Discovery are point-in-time measurements against live external systems (Perplexity / ChatGPT / Google AIO) (2026-06-26); they use a fixed query set, but external answers can change day to day.
Shows which buyer questions AI systems answer in the sector and whether the domain appears as a source.
Buyer-question language: Turkish. Questions remain in the tested language as measured evidence; the report language is English.
Summary
Prompt map with 10 questions. Top domain: trendyol.com (8).
How to Use the Prompt Map
Coverage means: share of Prompt Discovery questions with citation. 50% = 5 of 10 questions.
ts-wellness.com appears in this question space at rank #5: 6 citations, 50% coverage. Neighbors: eksisozluk.com / cimri.com; social platforms are counted separately.
A missing citation is only a candidate signal. Before any content decision, review topical fit, existing coverage, user value, original evidence, commercial bias and template uniqueness; then choose improve existing, create only if justified, consolidate, no action or specialist review.
Provenance: AI-generated research hypotheses; these are not search-volume or observed-demand data and are not automatic content briefs.
Social platforms appear as a separate channel: instagram.com, youtube.com.
Citation Evidence and Prompt Map measure different question spaces: citation Multi-ASOV 7% (1/15), Prompt Discovery coverage 50%. Both values are correct.
Query
Perplexity
ChatGPT
Google AI Overview
metilen mavisi nedir ve ne işe yarar, insan sağlığına faydaları var mı?
absent
absent
absent
USP grade metilen mavisi ile normal metilen mavisi arasındaki fark ne?
cited
absent
absent
Türkiye'de CoA sertifikalı USP grade metilen mavisi nereden satın alabilirim?
cited
cited
absent
nattokinaz kapsül günde kaç mg alınmalı, dozajı nasıl ayarlanır?
absent
absent
absent
metilen mavisi alırken nelere dikkat etmeliyim, sahtesini nasıl anlarım?
absent
absent
absent
nattokinaz mı serrapeptaz mı daha etkili, ikisi arasındaki fark nedir?
absent
absent
absent
USP grade ne demek, ilaç saflığında bu sertifikanın önemi var mı?
cited
absent
absent
en güvenilir metilen mavisi markası hangisi, CoA belgesi olan ürün önerisi?
absent
cited
absent
nattokinaz kapsül kalp sağlığı için kullanıyorum, metilen mavisi ile birlikte alınır mı?
cited
absent
absent
metilen mavisi fiyatları neden bu kadar farklı, ucuz olanı almak riskli mi?
absent
absent
absent
Legend: cited = domain is a source; mentioned = domain is named; absent = no citation/mention; - no sources counts as absent, but is not a site defect.
Google AI Overview: no sources returned in this answer
nattokinaz kapsül günde kaç mg alınmalı, dozajı nasıl ayarlanır?
Diagnosis
0 cited, 0 mention-only, 3 absent.
Interpretation
Frequently cited domains for this question: validus.com.tr, kirmizibeyazz.com, recete.com, youtube.com (Social Media / Platforms).
Evidence
Perplexity: absent — sources went to: validus.com.tr, kirmizibeyazz.com, recete.com, youtube.com
Google AI Overview: no sources returned in this answer
ChatGPT: no sources returned in this answer
metilen mavisi alırken nelere dikkat etmeliyim, sahtesini nasıl anlarım?
Diagnosis
0 cited, 0 mention-only, 3 absent.
Interpretation
Frequently cited domains for this question: cimri.com, aktifinternational.com, eksisozluk.com, tetkik.com.tr.
Evidence
Perplexity: absent — sources went to: cimri.com, aktifinternational.com, eksisozluk.com, tetkik.com.tr
Google AI Overview: no sources returned in this answer
ChatGPT: no sources returned in this answer
nattokinaz mı serrapeptaz mı daha etkili, ikisi arasındaki fark nedir?
Diagnosis
0 cited, 0 mention-only, 3 absent.
Interpretation
Frequently cited domains for this question: amerikasepetim.com, redfood24.de, eksisozluk.com, youtube.com (Social Media / Platforms).
Evidence
Perplexity: absent — sources went to: amerikasepetim.com, redfood24.de, youtube.com, eksisozluk.com
ChatGPT: no sources returned in this answer
Google AI Overview: no sources returned in this answer
5Implementation Backlog
Compact work extract: action, owner and validation. The diagnosis stays in the relevant report chapter.
Why there is no "improve citation status" action here
Citation is an outcome, not a switch. Template, answer-structure and source actions are the levers later validated in the citation chapter, query drilldown and prompt map.
Buyer-answer blocks
3 action(s) · priorities 1, 4, 8
Category selection guide
2 action(s) · priorities 7, 9
Catalog schema
3 action(s) · priorities 3, 5, 6
Reviews and proof
1 action(s) · priorities 2
5Implementation Backlog
AI JourneyPriority 1
1
TO DO: Add buyer-question blocks for choice, comparison, fit and purchase terms
Technical context: AI Journey Coverage is not sufficiently evidenced · Work package: Buyer-answer blocks
owner: SEO + content + developer
Effort: mediumEffect: high
Work order
A buying-decision block is a short, visible section in the product or category template that answers a real buyer question directly. Why: AI answers need quotable decision evidence on the shop page itself; otherwise they pull selection, comparison, delivery or warranty facts from marketplaces, review sites or competitors. Product example: "Will this model fit X?", "Which size/variant should I choose?", "Delivery, returns and warranty". Category example: "Which variant for use case A vs. B?", "Important selection criteria", "Comparison of subgroups". Use a blog only when the question cannot be answered well inside the template.
Trigger: AI Journey Coverage (ai_journey_coverage)
Observed: BOFU product decision surface: 67/100; MOFU category and selection surface: 33/100; Answer surface for AI questions: 17/100
Target basis: Derived from product/category criteria; no fixed word count, FAQ count or universal content length is assumed.
Validation: Rerun Full AI Visibility Audit and confirm AI Journey Coverage becomes usable or strong.
5Implementation Backlog
FindingPriority 2
2
TO DO: Mark up only visible, real reviews
Technical context: Product Pages: Review Schema · Work package: Reviews and proof
owner: SEO + developer
Effort: mediumEffect: high
Work order
A review-schema work order marks only reviews that a buyer can actually see on the product page. Why: AI systems and search features use reviews as trust evidence only when markup matches visible content; invisible or invented reviews create risk. Example: if a product visibly shows 23 verified reviews and a 4.8 rating, mark exactly that count and rating as AggregateRating/Review in Product JSON-LD. Done when affected product templates show real visible reviews and Review/AggregateRating validates without warnings.
A breadcrumb-schema work order describes the visible catalog path as BreadcrumbList JSON-LD. Why: AI systems need the path to place a product or category inside the shop hierarchy; without it, the page is an isolated item. Example: Home > Products > Product category > Product name is represented with the same names, positions and URLs in BreadcrumbList markup. Done when visible breadcrumbs, canonical URL and BreadcrumbList names/URLs match.
An answer-block work order is a short visible section that answers a real buyer question directly. Why: if the answer is not in the template, an AI system builds the answer from other sources and the shop page loses quotable evidence. Example: "Which variant should I choose?", "Is this product compatible with X?", "How is it used?", "Delivery, returns and warranty". Done when affected product or category templates answer these questions visibly and the Answer Structure measurement improves.
A CollectionPage/ItemList work order makes the category machine-readable as a real collection. Why: AI systems understand a product list better when the category, included items and order are explicitly marked. Example: the category template includes CollectionPage or ItemList with the visible product cards and URLs. Done when markup and visible listing show the same products.
A breadcrumb-schema work order describes the visible catalog path as BreadcrumbList JSON-LD. Why: AI systems need the path to place a product or category inside the shop hierarchy; without it, the page is an isolated item. Example: Home > Products > Product category > Product name is represented with the same names, positions and URLs in BreadcrumbList markup. Done when visible breadcrumbs, canonical URL and BreadcrumbList names/URLs match.
Target basis: Target is passing the measured template criterion on affected URLs; no fixed word count or universal content volume is assumed.
Validation: Rerun the crawl and confirm Breadcrumb Schema no longer fails or warns on the affected category pages URLs.
5Implementation Backlog
FindingPriority 7
7
TO DO: Add category context above the listing
Technical context: Category Pages: Intro Text ≥80 words · Work package: Category selection guide
owner: SEO + developer
Effort: mediumEffect: high
Work order
A category-intro work order explains before the product list how the buyer should choose. Why: a pure list does not answer a selection question; AI systems need criteria, variants and use cases to use the category as an advisory source. Example: purpose, variants, material, size, configuration, use case and important subgroups are explained briefly. Done when the category gives selection context before the listing.
Trigger: intro_text
Observed: 0/2 category pages pass Intro Text ≥80 words.
An answer-block work order is a short visible section that answers a real buyer question directly. Why: if the answer is not in the template, an AI system builds the answer from other sources and the shop page loses quotable evidence. Example: "Which variant should I choose?", "Is this product compatible with X?", "How is it used?", "Delivery, returns and warranty". Done when affected product or category templates answer these questions visibly and the Answer Structure measurement improves.
Target basis: Target is passing the measured template criterion on affected URLs; no fixed word count or universal content volume is assumed.
Validation: Rerun the crawl and confirm Answer Structure no longer fails or warns on the affected category pages URLs.
5Implementation Backlog
FindingPriority 9
9
TO DO: Add missing meta descriptions
Technical context: Category Pages: Meta Description · Work package: Category selection guide
owner: SEO + developer
Effort: lowEffect: medium
Work order
What to do: Add missing meta descriptions. Why: Missing meta descriptions force search and AI systems to infer the snippet or page summary from weaker body-text signals. Example: use the affected template and align visible content, structured data and canonical/URL signals. Done when validation passes: Rerun the crawl and confirm Meta Description no longer fails or warns on the affected category pages URLs.
Trigger: unique_meta_desc
Observed: 1/2 category pages have a meta description; 1 page(s) are missing one. The finding is about missing meta descriptions.
Target basis: Target is passing the measured template criterion on affected URLs; no fixed word count or universal content volume is assumed.
Validation: Rerun the crawl and confirm Meta Description no longer fails or warns on the affected category pages URLs.
6Methodology and Raw Data
6Methodology and Raw Data
FindingEvidence type: guidance_glossary / 6
Metric Glossary — How to Read the Score Layers
Reference definitions kept outside the Action Brief.
metric
meaning
boundary
SEO Foundation
Discovery, description and indexing hygiene.
Not traffic or revenue potential.
GEO Signal Layer
Whether facts and answers are extractable and attributable.
Not guaranteed AI placement.
AI Journey Coverage
Whether product/category pages support selection and buying decisions.
Not a conversion rate.
Citation vs mention
A citation uses the domain as a source; a mention only names it.
Absence is not automatic proof of a site defect.
Sector Benchmark
Comparison with measured top results for the chosen query.
Do not copy average word count blindly.
AEO / Agent Readiness
Extractability of contact, offer/action and freshness information.
Not the presence of agent files alone.
Rows shown: 6 of 6
6Methodology and Raw Data
FindingEvidence type: methodology / 44
Score Methodology
Reconstructable components, weights and weighted contributions for deterministic SEO/GEO/E-E-A-T/AEO scores.
Product Page Readiness
8 template criteria, measured on fetched product pages.
Category Page Readiness
7 template criteria, measured on fetched category pages.
AI Journey Coverage
Blend of product decision surface, category selection surface and answer surface for buyer questions.
Citation Status
Sampled live-AI snapshot at measurement time; outcome layer, not a direct repair switch.
Prompt Discovery
10 tested questions; coverage and mention logic across the tested prompt set.
score
component
value
weight
contribution
total
Layer note
Deterministic vs live-snapshot
Deterministic layer: crawl/engine metrics (SEO, GEO, E-E-A-T, AEO, schema, template) use fixed rules; same input gives the same score and can be re-measured. Live-snapshot layer: ...
-
-
-
SEO Foundation
Keyword in first 100 words
warning
7%
2.4
-
SEO Foundation
Keyword positioning
fail
4%
0
-
SEO Foundation
LSI / semantic keywords
pass
5%
5
-
SEO Foundation
Heading structure
warning
10%
3.5
-
SEO Foundation
Single H1
warning
8%
2.8
-
SEO Foundation
Content length
warning
6%
2.1
-
SEO Foundation
Paragraph length
warning
3%
1
-
SEO Foundation
Sentence variety
pass
2%
2
-
SEO Foundation
Passive voice
pass
4%
4
-
SEO Foundation
Internal links
pass
8%
8
-
SEO Foundation
External links
pass
5%
5
-
SEO Foundation
Image alt text
warning
6%
2.1
-
SEO Foundation
Direct-answer snippet
warning
5%
1.8
-
SEO Foundation
Schema markup
pass
10%
10
-
SEO Foundation
Meta description
warning
8%
2.8
-
SEO Foundation
E-E-A-T signals
pass
9%
9
-
SEO Foundation
TOTAL
headline score
100%
62
62
GEO Signal Layer
Entity clarity
pass
9%
9
-
GEO Signal Layer
Entity attributes
pass
7%
7
-
GEO Signal Layer
Citation worthiness
warning
9%
3.2
-
GEO Signal Layer
Sourced claims
pass
7%
7
-
GEO Signal Layer
Q&A format
warning
6%
2.1
-
GEO Signal Layer
Direct answers
warning
7%
2.4
-
GEO Signal Layer
Semantic completeness
pass
7%
7
-
GEO Signal Layer
Conversational structure
warning
4%
1.4
-
GEO Signal Layer
Knowledge graph / ontology
pass
9%
9
-
GEO Signal Layer
AI cliche detection
warning
4%
1.4
-
GEO Signal Layer
Follow-up query coverage
warning
5%
1.8
-
GEO Signal Layer
Freshness signal
pass
7%
7
-
GEO Signal Layer
Synthesizability
warning
7%
2.4
-
GEO Signal Layer
Social signals
warning
5%
1.8
-
GEO Signal Layer
Conversational tone
fail
7%
0
-
GEO Signal Layer
TOTAL
headline score
100%
62
62
E-E-A-T
Experience
26
25%
6.5
-
E-E-A-T
Expertise
84
25%
21
-
E-E-A-T
Authority
90
25%
22.5
-
E-E-A-T
Trust
87
25%
21.8
-
E-E-A-T
TOTAL
headline score
100%
72
72
AEO / Agent Readiness
Contact extractability
20
25%
5
-
AEO / Agent Readiness
Service/product readability
70
30%
21
-
AEO / Agent Readiness
Action flow accessibility
30
30%
9
-
AEO / Agent Readiness
Data freshness
20
15%
3
-
AEO / Agent Readiness
TOTAL
headline score
100%
38
38
Rows shown: 44 of 44
6Methodology and Raw Data
Crawl InventoryEvidence type: crawl_summary / 4
Crawl Evidence
Crawl totals and report evidence boundary.
label
value
URLs discovered
26
Pages fetched
24
URLs recorded only
2
Crawl cap hit
no
Rows shown: 4 of 4
6Methodology and Raw Data
Schema CoverageEvidence type: metric_summary / 3
Structured Data Evidence
Schema coverage from crawler/parser evidence.
label
value
Pages with schema
21
Total fetched pages
24
Schema page coverage
88%
Rows shown: 3 of 3
6Methodology and Raw Data
FindingEvidence type: table / 3
Citation — Platform Raw Data
Per-platform ASOV / mention / score from the citation add-on.
platform
asov
mention
score
Perplexity
20%
0%
15
ChatGPT
0%
0%
0
Google AI Overview
0%
0%
0
Rows shown: 3 of 3
6Methodology and Raw Data
FindingEvidence type: table / 15
Citation — Query Platform Evidence
Normalized query-platform table behind ASOV, mention rate and citation score.
