# Beauty review schema for AI Overviews: shade, skin type and proof

Make beauty and skincare reviews readable to AI Overviews. Schema checklist for shade, skin type, AggregateRating and attributable review bodies. We make Idukki.

By Rohin Aggarwal · 2026-10-31

A shopper asks an assistant for a foundation for oily, medium skin. Three brands come back with shade notes and skin-type quotes. Your serum has four thousand reviews. The answer never names you. The proof sat in a widget the model could not trust.

**Quick answer**

- **AI Overviews quote attributable proof.** Named reviewers, dates and specific bodies beat anonymous star averages.
- **Beauty needs attribute depth.** Shade, undertone, skin type, finish and concerns belong in review text and product copy.
- **Emit Review plus AggregateRating** that match the reviews a shopper can read on the page.
- **Never synthesise ratings** to look denser in an Overview. Fake counts get ignored.
- **Validate before you ship.** Run the review schema validator on your top SKUs.

_Figure: Beauty shoppers ask by shade, skin type and concern. Schema and review bodies have to answer in the same words._

We make Idukki. On product pages it can emit Product, Review and AggregateRating JSON-LD from that product's reviews. The same library holds customer photos that show shade and finish in real light. For the wider AEO pattern, read [AEO for reviews: getting ratings quoted by agents](/blog/aeo-for-reviews-getting-ratings-quoted-by-agents).

## Why beauty reviews need more than a star average

A jacket review can live on fit and fabric. A foundation review fails without shade and skin type. Assistants and AI Overviews look for those attributes when shoppers ask in full sentences.

Google's AI features docs say no special AI markup is required. Clean Product, Review and AggregateRating data still labels rating, count and quotes for engines. The review body does the vertical work.

- Shade and undertone ("Warm beige 230", "runs peach on fair-cool skin").
- Skin type ("oily T-zone", "dry cheeks", "combination").
- Concerns ("redness", "texture", "SPF under makeup").
- Finish and wear ("dewy at hour eight", "does not pill under sunscreen").
- Routine context ("over retinol", "under mineral SPF").

## The schema shape that gets quoted

Two types carry the load. Review holds one attributable opinion. AggregateRating holds the rolled-up score and count. Emit both, computed from reviews that exist on that page.

| Type | Must include | Beauty-specific note |
| --- | --- | --- |
| Product | name, image, offers or review/aggregateRating | Name the shade or variant the page sells |
| AggregateRating | ratingValue plus reviewCount or ratingCount | Count only rated reviews for this product |
| Review | author, datePublished, reviewRating, reviewBody | Body should mention shade or skin type when relevant |
| reviewRating | ratingValue with bestRating / worstRating | Keep the scale honest (usually 1–5) |
| author | Person name (or clear verified label) | Avoid endless "Anonymous" rows |

_Working field checklist for beauty PDPs. Required properties follow Google's review snippet guidance._

> **Match the variant the shopper sees:** If the PDP is shade 230, do not mark up an aggregate from shade 110 alone. Say so on the page when syndication mixes shades. Misleading shade proof is worse than a thinner count.

## Beauty review schema checklist

1. Confirm every marked-up review is visible on the product page.
2. Confirm ratingValue and reviewCount match the on-page summary.
3. Confirm authors and dates are real. Drop placeholder names.
4. Scan ten recent review bodies for shade, skin type or concern language.
5. Add review prompts that ask for shade and skin type after delivery.
6. Keep imported marketplace or Business Profile scores out of product AggregateRating.
7. Omit AggregateRating entirely when a SKU has no rated reviews.
8. Run the [review schema validator](/tools/review-schema-validator) on hero and shade-variant URLs.
9. Ask your category questions in AI Overviews and note whether your rating is named.

_Figure: Checklist flow: visible reviews, honest aggregate, attribute-rich bodies, then validation._

## Collection prompts that feed AI-readable reviews

Post-purchase asks shape what shoppers write. Beauty brands that only ask "How was it?" get vague praise. Ask for the attributes assistants need.

| Ask | Why it helps |
| --- | --- |
| Which shade or variant did you buy? | Locks the review to the right PDP |
| What is your skin type? | Matches oily / dry / combination queries |
| What concern were you treating? | Gives quoteable problem language |
| How does it wear after 6–8 hours? | Supports longevity questions |
| Optional photo in natural light | Shows shade and finish without studio polish |

_Prompt patterns that produce attributable, attribute-rich review bodies._

> **Keep incentives sentiment-free:** If you offer a small reward for a review, do not tie it to star rating. Google's Product Ratings and review policies expect honesty. Flag incentivised reviews when a feed requires it.

## Customer photos as citable proof

Shade debates die faster when a photo shows the product on skin. Give photos clear alt text that names product, shade and setting. Keep captions in HTML so crawlers can read them.

Tag each photo to the product and shade it shows. A mis-tagged swatch trains both shoppers and models on the wrong item.

## Validate, then ask the Overview

Markup that fails Google's rules will not help an Overview either. Validate first. Then ask five buyer questions in Search with AI Overviews and in one other assistant. Log whether your brand, shade and rating appear.

Google says its AI features need no special AI file. Still, broken AggregateRating or hidden reviews read as weak evidence. Fix those before you chase llms.txt or extra feeds.

**Where to start:** Pick your top foundation or serum. Make AggregateRating honest, push shade and skin type into review prompts, and validate the live URL.

### Questions beauty teams ask

**Q: Do AI Overviews need special beauty schema types?**

A: No. Google's AI features guidance does not require special AI markup. Solid Product, Review and AggregateRating data plus specific review bodies is the practical path.

**Q: Should every shade variant have its own AggregateRating?**

A: Rate the product entity the page represents. On separate shade PDPs, keep each aggregate honest. Say when reviews cover related shades.

**Q: Can I mark up influencer quotes as reviews?**

A: Only if they are real reviews that meet Google's guidelines and appear on the page. Paid praise without clear attribution is a trust risk.

**Q: Where does Idukki fit?**

A: We make Idukki. It collects reviews and customer media, tags them to products, and can emit review JSON-LD.

### Sources
- [Google Search Central, Review snippet structured data](https://developers.google.com/search/docs/appearance/structured-data/review-snippet) — Review and AggregateRating rules; visible reviews; no aggregation from other sites.
- [Google Search Central, AI features and your website](https://developers.google.com/search/docs/appearance/ai-features) — No special AI markup required; Search ignores AI-only text files.
- [Google Search Central, Product structured data](https://developers.google.com/search/docs/appearance/structured-data/product) — Product and Offer fields for merchant listings and snippets.
- [Schema.org Review and AggregateRating](https://schema.org/AggregateRating) — Shared vocabulary for ratingValue, reviewCount and Review.
- [Idukki: AEO for reviews](/blog/aeo-for-reviews-getting-ratings-quoted-by-agents) — How agents weight verified, attributable ratings.

---
Canonical: https://idukki.io/blog/beauty-review-schema-for-ai-overviews
Tags: reviews, beauty, skincare, aeo, ai-overviews, structured-data, schema