For beauty brands · 20–2,000 SKUs · £500K–£30M revenue
INCI
Beauty catalogs don’t lose on shade. They lose on a missing INCI field.
AI shopping agents filter beauty on ingredients, shade, volume, claims, skin type and regulatory metadata. When those fields live in your back-of-pack imagery and hand-written descriptions instead of structured metafields, the agent defaults to the brand that published structured data. Eight beauty-specific mistakes below — each one concretely observable in a real Shopify catalog.
Beauty median
47
/ 100 · grade F · 128 stores
The median beauty catalog scored by FlintmereBot. Updated periodically. Run your own scan to see where your store sits against it.
See my score →Eight beauty-specific mistakes
Each one is concretely observable. Each one is scored on every scan.
INCI list delivered as a back-of-bottle PNG
Your ingredient panel is a beautiful image — every brand does it. But agents can't read 'fragrance-free' off a 900×1200 JPG. A shopper asking ChatGPT for a niacinamide serum without fragrance gets the one brand that published their INCI as structured text.
Flintmere: Flintmere scans for an ingredients metafield or an in-description INCI block, flags image-only panels, and points you at the five to ten products with the highest search volume to convert first.
Check ·Structured attributes
Shade names only a human could parse
'Midnight Kiss.' 'Paper Crane.' 'First Dawn.' Beautiful merchandising; zero structured shade field. An agent answering 'nude lipstick' has no way to map Paper Crane to nude — so your 40-shade range might as well be one shade for agent-discovery purposes.
Flintmere: We check every variant for a structured shade-family field (nude, red, pink, berry, coral, brown, plum, bold) alongside your poetic names. Keep the romance; add the filter agents can match.
Check ·Structured attributes
Volume encoded as free text, not structured ml / fl oz
Your variants say 'Travel Size', 'Standard', 'Value Size'. Shoppers know what those mean. Agents don't. The structured size_volume metafield sits empty on 94% of variants, so price-per-ml queries — which agents run constantly — skip your entire catalog.
Flintmere: Flintmere flags free-text volume variants and offers a regex to extract numeric ml / oz values from existing variant titles — so you can bulk-populate the structured field without retyping each SKU.
Check ·Structured attributes
Claims written in prose, not a structured taxonomy
Your descriptions say 'vegan', 'cruelty-free', 'fragrance-free' in beautiful hand-crafted sentences. Agents matching 'vegan + fragrance-free cleanser' scan your description for those exact tokens, get inconsistent matches, and default to brands that publish a structured claims array.
Flintmere: We check for a claims or certifications metafield using the industry standard taxonomy (vegan, cruelty-free, fragrance-free, paraben-free, sulfate-free, dermatologically-tested, hypoallergenic) and flag products using prose-only claims.
Check ·Structured attributes
Period-after-opening (PAO) metafield empty on every SKU
EU + UK regulations require PAO on packaging but rarely make it into Shopify as structured data. Agents filtering for 'fresh, recently launched' or answering regulatory-aware queries skip products with no PAO. Your clean-beauty positioning depends on this field being present.
Flintmere: Flintmere checks for a period_after_opening metafield (standard EU spec: 6M / 12M / 24M / 36M) and flags every product missing it, grouped by category so you can populate by product archetype.
Check ·Data consistency
Skin type / hair type metafield not populated
A shopper asking 'cleanser for oily skin' depends on the agent matching a structured skin_type field. You mention 'oily and combination' in 40% of your descriptions. Agents can't parse it consistently — they return the three brands that ship structured taxonomies.
Flintmere: We check for skin_type / hair_type metafields using the Sephora-aligned taxonomy (oily, dry, combination, sensitive, normal; fine, thick, curly, coily, straight) and flag products where the field is empty or uses free-text synonyms.
Check ·Google category match
GTINs present on hero SKUs, missing on limited editions + collabs
Your hero line carries GS1 barcodes. The limited-edition collab you launched last month — the one with 30× the social buzz — has no GTINs. Agents demote it; shoppers asking about it by name don't find it in Amazon or Google Shopping. The hero line gets the halo; the collab disappears.
Flintmere: Flintmere flags every variant missing a barcode, validates the checksum on the ones you have, and groups the missing ones by collection — so you route the GS1 UK purchase decision by drop.
Check ·Product IDs
Seasonal + limited products still ACTIVE with zero inventory
Your holiday gift sets, the sold-out collab, the three shades you discontinued in January — all still status=ACTIVE with zero stock. Agents crawl them, send shoppers to dead pages, and your domain's agent-score drops. Limited drops are marketing gold and catalog-hygiene poison.
Flintmere: Flintmere cross-checks status × inventory × policy × publish date, ranks the dead pages by traffic cost, and gives you a one-click list of products that should be DRAFT, archived with redirect, or restocked.
Check ·Data consistency
Concierge audit for beauty
Send your shop URL. Three working days later, a written audit lands in your inbox.
From £197, the team reads your catalog product by product: a 1,500-word letter pointing at specific SKUs by name (beauty teams get the INCI, shade-family and claims-taxonomy reads as standard), a per-product fix CSV with the worst ten or twenty-five already drafted — shade family, structured volume, structured claims, skin type — a 30-day fix sequence, the right GS1 UK route for your product mix, and a 30-day re-scan. No video, no call, no slide deck. Three SKU bands — £197 (up to 1,500 SKUs), £397 (1,501–5,000), from £597 bespoke (5,001+).