Resources
Beyond product listings, BeautyFeeds collects individual customer reviews at scale — star ratings, full review text, verified purchase status, and reviewer attributes like skin type and tone, structured and ready for analysis.
Each review is collected as its own record — not just an average star rating rolled up on the product. That means you get the full text, the reviewer's context, and trust signals for every single review.
Star rating, review title, and complete review body — plus structured pros and cons when the retailer collects them separately.
Verified purchase flag, incentivized review disclosure, staff-review flag, and helpful/not-helpful vote counts — filter out noise before you analyze.
Skin type, skin tone, age range, skin concern, eye color, and hair color, when disclosed by the reviewer — useful for segmenting feedback by who it's actually from.
Links to any photos or videos the reviewer attached, plus badges like "Beauty Insider" when the retailer surfaces reviewer status.
Every review carries product_id, sku_id, and product_url, so it joins cleanly against our product dataset for per-SKU sentiment analysis.
Reviews are collected alongside our product data across 55+ retailers and 7 countries, tagged with site_name, country, and language.
Current review coverage across every retailer in the reviews dataset, updated as collection grows. Each site links to a dedicated page with sample rows and the same 36-field schema.
| Site | Reviews | Share |
|---|---|---|
| iHerb (iherb.com) | 19,712,500 |
|
| Sephora (US) (sephora.com) | 5,072,878 |
|
| Sephora (UK) (sephora.co.uk) | 4,259,096 |
|
| Sociolla (sociolla.com) | 783,533 |
|
| Amazon (amazon.com) | 57,818 |
|
| The Ordinary (theordinary.com) | 43,679 |
|
Reviews data rarely stands alone — it's usually paired with product and ingredient data for one of these workflows. Each links to a full breakdown with a typical workflow and relevant fields.
Trust-signal fields (verified purchase, incentivized, staff) isolate genuine sentiment before any model runs.
Consistent fields across 55+ retailers make reviews a clean feature source for recommendation and NLP models.
Review volume and rating trends add a demand signal on top of pricing and launch-activity research.
See the full list on the Use Cases page, or sign in to app.beautyfeeds.io to try one against live data.
An actual review record from our Sephora (US) collection — nothing simplified.
"i have combo skin and this worked great. its exfoliating so it helped with my oily skin but still pretty hydrating where it didn't worsen my dry spots :)" — tagged with skin_type: comboSk, skin_tone: fairLight, not a verified purchase, not incentivized.
Preview the review set that matches your needs before purchasing — filter by site, country, rating, submission date, verified purchase, and recommended status, see the matching count, then buy exactly that slice.
This filtering happens on app.beautyfeeds.io — sign in, pick a site from the table above (or all of them), preview the matching count, and export or connect via API.
BeautyFeeds is built by CrawlFeeds, which also collects structured review data for hotels, companies, products, and apps. See CrawlFeeds Reviews Datasets for review data beyond the beauty category.
Download a free 50-row sample, or filter the full review set and see how many match before you buy.