Sephora
Product matching by computer vision (skin tone) plus personalized on-site recommendation, synchronized between store and online
Sephora personalizes the beauty journey through computer vision (Color IQ, 14 million matches) and on-site recommendation: Sephora SEA achieved a 6-to-1 ROI on product-page recommendations via Dynamic Yield, with 82 experiences in six months.
Key points
- Product matching by computer vision (Color IQ) plus personalized on-site recommendation, synchronized store-to-online.
- In-house Color IQ and Dynamic Yield stack, unified by the Beauty Insider loyalty program.
- 6-to-1 ROI on product-page recommendations (Sephora SEA), 82 experiences in six months.
- 14 million Color IQ matches, evidence level B, confirmed status.
Objective
Reproduce online the personalized advice of the store: find the right shade, recommend the right products, and keep the thread between the in-store purchase and the web session.
The deployment
Sephora personalizes the beauty journey on two fronts. In store, Color IQ scans the customer's skin tone and translates it into a four-digit code drawn from a shade library built with the Pantone Color Institute; this code then filters foundations and concealers on mobile and online, once linked to the Beauty Insider loyalty account. Since its launch, Color IQ has generated 14 million matches in store. Online, Sephora SEA (Southeast Asia) entrusted Dynamic Yield with personalizing product recommendations and optimizing discovery points. Six months after the start, the team had set up 82 personalized experiences; product-page recommendations alone yielded a 6-to-1 ROI, with more than 6.50 dollars of direct revenue per dollar invested. Personalization extends to the no-results page, where contextual recommendations lifted the add-to-cart rate to 30% for returning visitors.
Results Proof B
The on-site ROI comes from a quantified Dynamic Yield case study (an interested vendor source); the 14 million Color IQ matches figure is reported by the press (Digiday) with a quote from a Sephora lead. Two aligned strands, but no consolidated financial result at the group level.
How it works
Documented architectureThe stack in detail
- plateforme Dynamic Yield on-site personalization engine: recommendations on product pages and the no-results page, optimization of discovery points (82 experiences in six months at Sephora SEA)
- outil Color IQ in-house in-store skin tone scan (computer vision), translated into a four-digit code linked to the customer account
- infra Beauty Insider loyalty program that unifies shade code, in-store purchases, and online browsing, the key to omnichannel synchronization
- integrateur Pantone Color Institute partner in building the Color IQ shade library
How it runs, concretely
For ops teams-
1Capture skin tone in store AI / retail team
Color IQ scans the skin and assigns a shade code, linked to the customer's Beauty Insider account.
-
2Filter products by shade AI / platform
The code filters foundations and concealers on mobile and online, to show only the matching shades.
-
3Personalize on-site recommendations AI / e-commerce team
On product pages and the no-results page, Dynamic Yield serves recommendations by context (similar products, bought together, automatic) and picks the best strategy by market.
-
4Synchronize store and online CRM team
CRM data feeds the online session from in-store purchases and behaviors.
The shade code and the purchase/browsing history, synchronized between store and online via the Beauty Insider account. If the CRM link breaks, the online session starts without context and the recommendation loses its relevance.
How your customers perceive this type of use
Sourced studiesLe paradoxe est documente des deux cotes : 71% des consommateurs attendent des interactions personnalisees et 76% sont frustres quand elles manquent (McKinsey, 2021), mais 75% declarent ne pas acheter aupres d'organisations auxquelles ils ne confient pas leurs donnees (Cisco, 2024). La « creepy line » est localisee : messages recus quelques secondes apres une recherche et suivi de localisation sont les pratiques qui mettent le plus mal a l'aise (Periscope by McKinsey, 2019).
Acceptance conditions
- La confiance dans le traitement des donnees precede l'achat : 75% ne achetent pas sans elle (Cisco 2024)
- Un cadre legal protecteur rassure : 59% des consommateurs disent que des lois fortes sur la vie privee les rendent plus a l'aise pour partager des informations dans des applications IA (Cisco 2024)
- La personnalisation elle-meme est attendue quand elle est consentie : environ la moitie des consommateurs (US 55%, UK 52%) disent s'inscrire souvent ou parfois a des services personnalises (Periscope by McKinsey 2019)
Red lines
- Le message declenche quelques secondes apres une recherche ou un achat : deuxieme ou troisieme cause de malaise selon les pays (Periscope by McKinsey 2019)
- Le suivi de localisation percu comme de la surveillance : 40% de malaise en Allemagne et au Royaume-Uni (Periscope by McKinsey 2019)
- Le mesusage des donnees personnelles par l'IA, devenu la premiere inquietude des consommateurs, a 53% et en hausse (Qualtrics 2025)
Sources: McKinsey & Company 2021 · Periscope by McKinsey 2019 · Cisco 2024 · Qualtrics 2025
How to replicate
Inference, not sourcedData prerequisites
- A unified customer account linking store and online purchases (CRM / loyalty)
- Fine-grained product attributes (shades, categories) to filter and recommend
- On-site browsing and purchase history
Org prerequisites
- A loyalty program as the data-unification key
- In-store scanning hardware if you want the shade component
- An e-commerce team running the personalization tool
Possible stack
- On-site personalization engine (Dynamic Yield, or equivalent) for recommendations and page optimization
- A dedicated product-matching component (shade, size) depending on the sector
The plan, step by step
- Step 1Unify the customer key: verify that the loyalty program links store purchases, profile, and online sessionsDeliverable: Single customer account with consolidated history
- Step 2Structure the fine-grained product attributes (shades, categories, complements) needed for filtering and recommendationDeliverable: Catalog with attributes usable by the engine
- Step 3Deploy the personalization engine on product pages and the no-results page, with a control groupDeliverable: First recommendations in production, measured against a control
- Step 4Roll out the recommendation strategies (similar, bought together, automatic) by page and by marketDeliverable: A set of experiences A/B tested in production
- Step 5Measure direct revenue per euro invested against the control, generalize what wins; the in-store scan component comes later, if the business justifies itDeliverable: ROI table by experience and a scaling plan
First step: Deploy recommendations on product pages and on the no-results page via a personalization engine, and measure direct ROI against a control group.
The same pattern, proven in other industries
Function of Beauty
personalized product formulation by algorithm from a quiz
Henkel
genAI content supply chain for one-to-one personalization at scale
Hims & Hers
AI-driven personalized treatment routing on intake data (MedMatch)
Sources
- S1 How Color IQ, Sephora's shade-matching skin care tool, boosts brand loyalty - Digiday Established press archive pending
- S2 Case Study: Sephora SEA Personalizes Beauty - Dynamic Yield Interested party archive pending
- S3 Sephora SEA Chooses Dynamic Yield to Personalize the Entire Customer Journey - PR Newswire Interested party archive pending
An error, newer info, a source?
This page lives on its accuracy. If a figure has moved, if the deployment has changed, or if you have a higher-quality source, tell us. Every sourced correction is verified before publication.