27,769 reviews. No product schema. The most validated brand with the weakest data infrastructure.

Executive Summary

• Brand: Customised hair care brand — customers take a quiz, select hair goals, and receive a personalised formula
• Data infrastructure: Highest review volume in 55 audits paired with the broadest product schema gap. Customisation differentiation has no native structured data representation
• The pattern: Brands built on experiential differentiation (quizzes, customisation, personalisation) struggle to translate that differentiation into structured data AI agents read
• Key competitor gap: Generic hair care brands with basic product schema outrank Function of Beauty on standard queries because their data is at least crawlable
• Root cause: 2 of 3 audited products return no product schema, the Trustpilot 4.0/27,769 isn't reflected in on-site aggregateRating, no schema.org property exists for "personalised formula"
• Fix complexity: Medium — basic schema implementation is straightforward, expressing customisation in structured data requires creative use of existing fields

The brand

Function of Beauty is a customised hair care brand. Customers take a quiz, select their hair goals, and receive a personalised formula. The brand has been featured widely and built a loyal following on the promise that no two bottles are the same. The product is different for every customer — powerful for customer experience, structurally challenging for AI visibility.

The audit

We audited Function of Beauty's product data as part of a hair care group study. The audit covers structured data implementation, tag taxonomy, description depth, and external review signals.

The findings

The highest review count in 55 audits. 27,769 Trustpilot reviews at 4.0. The next highest count in the audit set is roughly 2,000-3,000. Function of Beauty has roughly ten times more external reviews than any other brand audited. A 4.0 rating at nearly 28,000 reviews represents a stable, high-confidence signal. This is not a brand with 50 five-star reviews from friends and family. This is tens of thousands of customers who felt strongly enough to leave feedback. On Trustpilot alone, Function of Beauty has more social proof than most brands accumulate across all platforms combined.

Missing product schema. Two of the three audited products return no product schema. No JSON-LD product markup. No aggregateRating. No structured price data. No product description in machine-readable format. For an AI agent processing a shopping query, a page without product schema is effectively a blank entry. The agent cannot extract the product name, price, rating, or description through structured data. The most reviewed brand in the audit set is structurally invisible to AI agents on most of its product pages.

The customisation paradox. Standard product schema (schema.org/Product) assumes a fixed product with consistent attributes. A shampoo has a name, a price, a description, ingredients, and a rating. Function of Beauty's shampoo has all of those — but they change based on the customer's quiz answers. There is no standard schema field for "personalised formula." There is no JSON-LD property for "quiz-based customisation." The attribute that makes Function of Beauty fundamentally different from Ceremonia or Crown Affair does not have a structured data representation.

Why this is happening

Even basic schema would help. Beyond the customisation issue, the basic infrastructure gap is significant. Product pages need schema markup regardless of whether the customisation angle can be fully expressed. Basic product schema — name, description, price range, category, and aggregateRating — would at least make Function of Beauty visible for generic hair care queries.

The Trustpilot signal isn't connected. Currently, an AI agent processing "best shampoo for damaged hair" cannot find Function of Beauty through structured data on most product pages. The 27,769-review Trustpilot score exists on a third-party platform. If it were reflected in on-site aggregateRating schema, it would be one of the strongest trust signals in the entire hair care category.

Personalisation is hard to schema. schema.org doesn't have a "personalisation" property. The answer is finding ways to express customisation through existing structured data fields rather than waiting for a new standard. Descriptions that explain the customisation model. Tags that cover the range of hair types and goals the quiz addresses. FAQ schema that answers "how does personalised hair care work?"

What Function of Beauty could do, in priority order

Phase 1 (quick wins):

• Implement basic product schema across the catalogue — name, description, price range, category
• Reflect the Trustpilot 4.0/27,769 in on-site aggregateRating JSON-LD on every product
• Add product descriptions of 80+ words that explain the customisation model in machine-readable form

Phase 2 (medium effort):

• Add tags covering the range of hair types, textures, scalp concerns, and goals the quiz addresses
• Implement FAQ schema that answers customisation questions in structured form
• Build additionalProperty JSON-LD fields for hair type ranges, goal categories, and ingredient choices the quiz exposes

Phase 3 (longer term):

• Develop content authority around personalised hair care as a category AI agents can cite
• Pursue editorial inclusion in "best customised hair care" and "best personalised beauty" roundups
• Build the brand's customisation story into a structured data model that other personalisation brands could adopt

Close

Function of Beauty raises a strategic question that extends beyond one brand. As AI shopping grows, how do brands whose differentiation is experiential — quizzes, customisation, personalisation — make that differentiation visible to AI agents? The brand with 28,000 reviews has proven the model works for customers. The challenge now is proving it exists to AI agents.