How Noui Collective works

Our recommendation logic

Recommendations are deterministic. Your answers become a structured Skin Profile — skin type, sensitivity, primary and secondary concerns, preferences, tolerance to actives and budget. Products are then scored against that profile using their structured attributes: suitable skin types, targeted concerns, key ingredients, formulation flags, routine step, rating and price. Products that conflict with your stated sensitivities or avoidances are excluded rather than downranked, and every recommendation shows the reasons behind it.

Sponsorship and affiliate relationships never affect scoring. Sponsored placements and affiliate links are labelled wherever they appear.

Price comparison

Every product holds one offer per retailer per country, with price, currency, size, price per ml or gram, stock status and a retailer link. We show the lowest available price first and always show the full list so you can weigh shipping, loyalty schemes and bundle deals yourself.

Not medical advice

The skin analysis is a cosmetic preference questionnaire. It does not diagnose conditions and it is not a substitute for a dermatologist or doctor. Ingredient pairing guidance is editorial, manually reviewed, and written to help you use products sensibly. For persistent, painful or worsening skin concerns, please see a medical professional.

Built and working today

• — Structured product catalogue: brands, retailers, offers per country, ingredients and editorial content
• — Personalisation: full skin analysis questionnaire, derived Noui Collective Skin Profile, deterministic scoring and routine builder
• — Search & compare: keyword search, filters, sorting including value per ml/g, and multi-retailer price tables
• — Ingredient library with manually reviewed pairing rules and a routine compatibility checker
• — UK and US regional switching with local currency and country-specific retailer offers
• — Accounts, saved products, saved routines and stored price alerts

Integration roadmap

These features are designed and wired into the database, but need external integrations or custom development before they can be described as operational. Pricing, ratings and stock data in the current catalogue are placeholders for demonstration.

• Affiliate & retailer product feeds

  Requires: Awin, CJ, Rakuten, Sovrn or direct retailer APIs

  Needed for large-scale ingestion of thousands of SKUs, real stock status and real affiliate links. The catalogue schema, offer table and country fields are already built for it.

• Automated price refresh

  Requires: Scheduled job plus feed or licensed scraping partner

  Prices in the database are placeholder values generated for demonstration. Each offer row already carries a last-updated timestamp ready for a scheduled refresh.

• Price-drop alert emails

  Requires: Scheduled comparison job plus a transactional email provider

  Alerts are already stored per user, product, country and target price. Only the delivery job is missing.

• Conversational AI assistant

  Requires: An AI model integration

  The assistant currently maps language to structured filters with explicit rules. Free-form conversation, follow-up memory and explanation generation need a model.

• Photo skin analysis

  Requires: Image model plus explicit consent, storage and privacy handling

  Intentionally not implemented. Skin imagery is sensitive personal data and needs a legal review before launch.

• Reviews & ratings ingestion

  Requires: Retailer review APIs or a licensed review provider

  Rating and review counts are placeholders on the product record and swap out cleanly once a source is connected.