BrandlightResearch Get a demo
Chapter 5 · Third-party

How do you make retailer product pages machine-legible?

When AI recommends a product, it cites retailer product pages, not your brand site. Six pillars decide whether your PDPs get cited or stay invisible.

Retailer product page with ratings, structured data and semantic descriptions

For commerce brands, the pages AI cites most are retailer product pages, not your own .com. You shape them through the content briefs you send retailers. Most brands do not send briefs. The ones that do get cited.

The pattern is blunt. On unbranded queries, page structure influences citations more than brand alone. Two brands in the same category, same retailer, same price range: one earns citations, the other is invisible. The difference is not prestige. It is the six pillars below.

✓ Cited
Optimised brand
868
citations · 281 pages
AggregateRating + Review schema on every retailer PDP
Semantic description: type, notes, occasion, fit as plain text
Earns citations on broad structured commerce pages

The whitespace is universal. Contextual use-cases, awards, and FAQ are missing on every page, cited brands and invisible brands alike. One route to differentiation.

The six pillars

Six pillars of PDP optimisation

Each one is a Brandlight best-practice rule, made machine-legible and pushed through retailer content briefs. Every play routes through a brief.

01

Semantic descriptions

Lead with what it is and who it is for. AI relies heavily on semantic retrieval, so meaning matters more than keyword matching. When a shopper asks "best [product] for [occasion]", AI needs product type, notes, occasion, and fit as plain text in the description.

Good · semantic-led
PRODUCTCO · Eau de Toilette 100ml
BOSS Bottled Eau de Toilette for Men 100ml
A modern-classic Eau de Toilette for men. Crisp apple and bergamot over an aromatic heart of cinnamon and clove, on a smooth sandalwood-and-cedar base. A confident, versatile day-to-evening signature.

Smells like: crisp apple, warm spice, smooth woods · Best for: office to evening · Choose it if you want a versatile everyday classic.
Bad · tagline-led
PRODUCTCO · Fragrance
BOSS Bottled EdT
Ein Duftklassiker der eleganten Ausstrahlung...

Notes appear only lower in the page. Occasion, audience and use case are absent. AI cannot extract meaning for situational queries.
  • Lead with type and occasion
    Spell the product type in full (Eau de Toilette, not EdT). State who it is for and when. AI needs this as plain text to match situational queries.
  • Put notes and benefits up top
    Do not bury the substance below a tagline. Benefits ahead of features. Product facts as image-adjacent copy AI can index.
  • Cut brand-name repetition
    Lead with the reason to choose. Brand-name repetition in body copy wastes the semantic space AI uses to match queries.
  • Tagline-led opening
    A tagline communicates brand feel, not product facts. AI cannot answer "what does it smell like?" from "timeless elegance".
  • No audience or occasion
    AI queries are situational. Without an occasion, the page cannot match "best gift for dad" or "office-appropriate fragrance" queries.

Lead with the reason to choose. Spell the type in full, never abbreviate. Cut brand-name repetition from body copy.

02

Ratings & reviews

The schema and reviews that separate cited from invisible. AI does not just show a star rating. Review text appears to inform the use-cases and sentiment signals that engines surface. AggregateRating + Review schema is the single strongest signal across all cited pages.

Good · schema marked up
PRODUCTCO · BOSS Bottled EdT
★★★★★ 4.8 (2,400+ reviews)
fresh long-lasting office-appropriate
Review + AggregateRating schema marked up · tagged themes surfaced
Bad · no schema
COMPETITORCO · Competitor Fragrance
No star rating displayed. No review schema. No tagged review themes.

AI cannot surface a credibility signal for this page. Zero citations earned across six weeks of reports.
  • Mark up ratings prominently
    Display ratings on every retailer PDP and mark them up with Review + AggregateRating schema. This is the line between cited and invisible.
  • Tag the review themes
    Surface tagged themes ("fresh", "long-lasting", "office-appropriate") so AI can extract use-case patterns. These themes are what AI surfaces in answers.
  • Drive review volume on priority retailers
    Push recent, tagged reviews to priority retailer pages. More review signal = stronger citation probability.
  • No Rating or Review schema
    This is the single most consistent gap between cited and invisible pages. Without it, AI has no structured credibility signal to surface.
  • Reviews present but not marked up
    Unstructured review text is better than nothing, but schema is the superhighway. AI extracts tagged pairs first, not free text.

Reviews schema is the line between cited and invisible. It is present on every cited page and absent on every invisible one.

03

Contextual use-cases

Tell AI when to use it. AI queries are situational. "Best gift for dad", "best fragrance for the office", "fragrance for a first date." No page we analysed carries any "best for" or "when to use" copy. This is a universal gap and the fastest route to differentiation.

Good · best-for block
Marc Jacobs Daisy · Boots
Best for / occasions
• A playful, effortless fragrance for everyday wear
• Spring and summer occasions
• Daytime occasions such as lunches, work or casual outings
• Gifting, thanks to its versatile and widely-loved scent profile
BOSS Bottled or The Scent? Choose Bottled for a fresher, lighter daily use.
Bad · no use-case copy
Category page · parfumdreams.de
Occasion sits only inside review text ("kann zu verschiedenen Anlässen getragen werden") and never in a standalone Best For block.

No page in the analysis carries explicit "when to wear" copy. AI cannot answer "Parfum fürs erste Date?" from product description alone.
  • Add a "Best for / occasions" block
    Connect each product to the specific moments and scenarios it is made for. AI queries are situational. The page must answer them directly.
  • Add a "Choose this if..." line
    Contrast each SKU with the adjacent one so AI can answer comparison queries ("Product A or Product B?") directly from the page.
  • Occasion buried in review text
    Review text is not the same as explicit structured copy. AI needs a distinct "Best for" block it can clearly attribute and cite.
  • No cross-SKU comparison
    Comparison queries ("which is better for evenings?") go unanswered if the page never contrasts one SKU with another.

Add a fundamental reason to buy this over the hundreds of others. Say exactly who it is for and when.

04

Awards & certifications

Make third-party validation crawlable. LLMs prioritise verifiable, third-party validation when recommending. Crawlers cannot read badges or icons. Only readable text and structured data count. Currently, every validation on most PDPs lives in imagery and is invisible to AI.

Good · crawlable text
Recognition block · as crawlable text
"A modern icon since 1998 and one of Germany's best-selling men's fragrances."

Best-seller: Top 3 men's fragrance on Douglas.de · Heritage: Over 25 years in market
Claim added to Product schema (award / additionalProperty fields)
Bad · badge images only
Fragrance Foundation award banner
Award shown as image only: "AWARD CATEGORY: PACKAGING OF THE YEAR - PRESTIGE / POPULAR"

No alt text. No crawlable text. No schema. AI cannot read or cite this validation.
  • Surface as crawlable text
    Heritage claims, best-seller rankings, and editorial quotes must exist as plain text on the page. AI cannot read badge images.
  • Mark it up in schema
    Add awards to Product schema additionalProperty fields, not just the page body. This makes the claim machine-readable and verifiable.
  • Awards only as imagery
    Badge images without alt text are completely invisible to AI. If validation only lives in graphics, it does not exist for LLMs.
  • Empty adjectives without proof
    "Award-winning" without stating the award is unverifiable. AI requires specificity. "Winner of the 2024 Fragrance Foundation award" is citable. "Award-winning" is not.

Lead with verifiable proof, not empty adjectives. Crawlers can read text and schema, not badges and icons.

05

FAQ & question targeting

Answer the question on the page. AI queries are conversational questions. Pages with explicit Q&A become the direct source for the generated answer, and FAQPage schema lets engines extract clean pairs. Currently, no PDP in the analysis carries an FAQ section or FAQPage schema.

Good · 5-8 FAQs on the page
Daisy Eau de Toilette · theperfumeshop.com
FAQs
What does this fragrance smell like?
Is this fragrance suitable for everyday use?
How long does Marc Jacobs Daisy Eau de Toilette last?
What does this Eau de Toilette best suit for?
Can Daisy Eau de Toilette be worn in the evening?
+ FAQPage JSON-LD on every hero PDP
Bad · no FAQ anywhere
PDP tabs · parfumdreams.de
PDP tabs: Description · Scent notes · Application · Ingredients

No FAQ tab. No FAQ section. No FAQPage schema. The only question anywhere is a category H2 on the listing page, not the product page. AI cannot extract conversational Q&A pairs from this page.
  • 5-8 FAQs per hero SKU
    Seed questions from actual search queries in your market. Answer them directly on the PDP, not in a separate support section.
  • FAQPage JSON-LD on every hero PDP
    Supply pre-written FAQ blocks to retailers in every content brief. FAQPage schema lets engines extract clean Q&A pairs without parsing the full page.
  • No FAQ section at all
    AI queries are conversational. Without explicit Q&A, the page cannot become the direct source for a generated answer.
  • Questions only on category pages
    Category-level FAQs do not help AI answer product-specific queries. FAQ must live on the product page itself.

Seed questions from the actual queries in your market. Answer them directly on the page. Supply pre-written blocks to retailers.

06

Structured data & schema

A strong base already exists. Treat schema as table stakes, but prioritise on-page copy. Product schema lets AI verify price, SKU, availability and facts before recommending. FAQPage and AggregateRating are recommended but it is far more important to have the copy on-page first.

Good · full schema suite
Rich results test · BOSS Bottled PDP
✓ 13 valid items detected
Product · ProductGroup · Offer · BreadcrumbList · AggregateRating · Review + FAQPage (added via brief)

Variant URLs per SKU · price + availability per variant · per-market schema consistency
Bad · missing FAQPage
Rich results test · Competitor PDP
Product + Offer + BreadcrumbList present (strong base) but no FAQPage schema anywhere.

Rating/review markup missing or at ProductGroup level only, not Product level.

No per-variant offers. AI cannot verify individual SKU facts.
  • Required JSON-LD minimum
    Product {brand, name, description, offers (price, availability, SKU per variant), aggregateRating, review} + FAQPage. This is the minimum to brief into every retailer.
  • Carry rating at Product level
    Rating/review/description at the Product level, not only ProductGroup. AI verifies facts at the individual SKU level before recommending.
  • No FAQPage schema
    The most common gap. Product and AggregateRating schema is becoming standard. FAQPage is not yet standard, which makes it a differentiation opportunity.
  • Schema only at category level
    Category-level schema does not help AI answer product-specific queries. Schema must be at the individual product page and variant level.

Schema is table stakes. The copy is where you win. Build the copy first, then mandate schema in every retailer brief.

Worked example

What a fully-applied PDP reads like

Every pillar applied to one page. This is the target state for a content brief to any retailer.

Retailer PDP · Product Name · Format/Size
AI-Ready Copy
Description

A modern-classic Eau de Toilette built for everyday confidence. Crisp apple, bergamot and a green lift open bright and clean, before an aromatic heart adds warmth. Smooth base notes give it quiet depth and dependable all-day staying power.

Fragrance family: Aromatic Spicy
Top: Apple, bergamot, lemon
Heart: Cinnamon, clove
Base: Sandalwood, cedar, vetiver

Why you'll love it · Best for / Occasions
  • A modern icon, one of the world's most recognised men's fragrances
  • A crisp citrus opening that feels fresh, never heavy
  • Smooth spiced warmth with dependable all-day wear
  • Effortlessly versatile, desk to dinner, weekday to weekend
  • Best for: the office and professional settings
  • Best for: gifting, a widely loved safe choice
  • Choose this if you want a fresher, lighter daily signature
FAQ
  • What does this fragrance smell like?
  • Is it appropriate for the office?
  • Product A or Product B, which should I choose?
  • How long does it last?
  • Is this a good gift?
+ FAQPage JSON-LD · AggregateRating + Review schema · Product + Offer schema with per-variant SKUs
868 vs 25
Optimised pages get cited. Brand prestige does not. Two brands in the same category, same retailer, same price range. One earns 868 citations. The other earns 25. The difference is the six pillars. The playbook is table stakes, not competitive advantage, until every brand applies it.
Remember

Five things that define the work

01

Optimised pages get cited. Brand prestige does not move the needle without structure.

02

Reviews schema is the dividing line between cited and invisible. Present on every cited page, absent on every invisible one.

03

You shape retailer pages through briefs. Every play routes through a content brief to the retailer. Write them.

04

Crawlers read text and schema, not badges and icons. If validation lives only in imagery, it does not exist for AI.

05

Use-cases and FAQ are the whitespace. Missing on every page across every brand. First to add them owns the situational queries.

Want your PDP audit?

We will show you which of the six pillars your retailer pages are missing, and where fixing them earns the most citations.

Book a 30-minute session