Beauty: ingredient data, INCI lists, and compliance across markets

Beauty's defining data headache is that its most important product data is regulated. Ingredient lists aren't marketing copy - they're legally mandated disclosures (INCI format in the EU under the Cosmetics Regulation), they change per formulation batch, customers genuinely read them (allergy sufferers, ingredient-conscious shoppers - a large and growing segment), and the claims built on them ("fragrance-free," "vegan," "non-comedogenic") each carry their own substantiation requirements that vary by market. Most beauty catalogs handle all this as a text blob pasted into the description. That's the headache; here's the structure.

INCI as structured data, not paragraph

The core move, pure Phase 2: the ingredient list becomes a structured field - ideally a list-type metafield of INCI names in declaration order (order is legally meaningful: descending concentration above 1%) - with the formulation version and effective date alongside it. What structure buys over the blob: searchability and filtering ("fragrance-free," "contains niacinamide" as real filters - for ingredient-conscious customers this is the shopping experience, and it runs on exactly the facet machinery from Phase 5); consistency at the brand level (one canonical INCI spelling per ingredient across the catalog - the closed-list doctrine; "Aqua/Water/Eau" chaos is beauty's version of the fourteen linens); change management (reformulations arrive as data updates with dates - Phase 10's versioning discipline, here with legal weight, because the shipped product and the displayed list must match per batch transition); and derivation - the claims layer computed from the ingredient layer where possible ("fragrance-free" as a check against the INCI list, not a vibe someone typed), which turns a compliance risk into a validation rule.

Claims and the per-market matrix

The claims vocabulary ("hypoallergenic," "clean," "dermatologically tested," "reef-safe") is where beauty marketing meets per-market regulation, and the two worst practices are winging it and hard-coding it per market. The structural answer from Phase 9: claims as structured attributes with substantiation notes, rendered per market through conditional rules, with the market-requirements matrix owned and reviewed (EU claims criteria, per-country enforcement flavors, and the labeling specifics - responsible-person details, PAO symbols, nanomaterial disclosure - each a data field plus a rendering rule, not per-market page copy). The discipline sounds heavy; in practice it's a one-time modeling exercise plus the annual requirements review, and it's dramatically lighter than the alternative - which is discovering per-market claim problems via a competitor's complaint or a platform delisting.

The discoverability dividend

Here's the reframe that makes beauty's compliance burden a competitive asset: the same structured ingredient and claims data that satisfies the regulation is precisely what modern discovery consumes. Ingredient-level queries dominate beauty search ("niacinamide serum without fragrance" - a fanout-shaped, constraint-laden query, Phase 6's home territory); AI assistants answering "is [product] safe for sensitive skin" extract from exactly these fields when they're extractable, and guess when they're blobs; and Google's product experiences increasingly surface ingredient facets directly. Brands that structured this data for compliance find themselves accidentally optimized for the most valuable queries in their category - while competitors' INCI paragraphs are invisible to every one of those surfaces. Compliance-grade data, discoverability-grade returns: one structure, both payouts. That's the pattern this phase keeps finding.