About
How BeautyConsensus works.
Real opinions about beauty products are scattered across Amazon reviews, community threads and store pages. BeautyConsensus gathers them into one consensus page per product, so you can see what reviewers actually agree on.
Currently tracking 81 products across 13 brands, built from 718 reviews. The full corpus is available as llms-full.txt and mapped in llms.txt.
Where the reviews come from
Every review shown here is real, attributed, and links back to where it was published:
- Amazon: verified-purchase reviews, collected per product.
- Communities: public Reddit and TikTok mentions of a brand or product.
- Verified stores: first-party customer reviews from brand storefronts.
- Editorial: curated highlights, clearly labeled.
How the consensus is computed
Reviews are de-duplicated, filtered for substance, and focused on well-rated experiences. The rating you see is a plain average of the reviews shown on the page; we don’t invent, weight, or alter them. Because we feature well-rated experiences, a page shows a curated view rather than a product’s complete rating. If a product doesn’t have enough real reviews, it doesn’t get a page.
We also drop reviews that disclose they were incentivized (a free or discounted product given in exchange for a review) or that are bad-faith rating pressure rather than a genuine experience, since neither is a trustworthy signal.
Disclosures
BeautyConsensus curates and republishes real, attributed customer reviews of the featured brands. We do not write, alter, or pay for the reviews. Where BeautyConsensus has a business relationship with a featured brand, that relationship does not change which reviews are shown or how ratings are calculated.
What you see here is a curated selection of positive, well-rated reviews, not a product’s complete or overall rating. Every review is real and links back to its original source; incentivized and bad-faith reviews are excluded.
Built to be read
Every page is plain, fast HTML: no logins, no pop-ups, and no scripts needed to read a single review. Each page also carries structured data (schema.org) so search and research tools can read the same consensus you do.