Why reviews feel safe
Reviews are social proof in its most persuasive form: not just a crowd, but a crowd that seems to have bought the thing and come back to tell you how it went. When you cannot hold the product or meet the seller, that feels like the next best thing.
Ratings also pull later ratings along. In a randomised experiment on a social news site, adding an early positive rating made later positive ratings 32% more likely, and final ratings ended up 25% higher on average. A handful of planted reviews do not just sit there. They recruit real ones.
What a rating shows, and what it does not
What it shows
- How many reviews a listing has collected
- The average of whatever was published
- Sometimes, that the reviewer bought through that platform
What it does not show
- Whether reviewers were paid, given free stock or refunded
- Whether bad reviews were removed or discouraged
- Whether reviewers are the seller's staff or friends
- Whether a testimonial screenshot came from a buyer at all
There is a market for reviews
Economists Sherry He, Brett Hollenbeck and Davide Proserpio studied sellers who bought Amazon reviews in large private groups on Facebook and other sites. Buying reviews raised average ratings and sales rank, but the lift faded after about a month. When sellers stopped buying, their ratings fell and the share of one-star reviews rose, which suggests fake reviews were mostly propping up worse products.
Regulators have caught up. The FTC's rule on consumer reviews and testimonials, in force in the US since 21 October 2024, bans fake reviews including AI-generated ones, paying for reviews that say a particular thing, undisclosed reviews by insiders, company-run review sites posing as independent, and threats to get bad reviews removed. Courts can impose civil penalties for knowing violations.
In India, the Department of Consumer Affairs launched IS 19000:2022, a Bureau of Indian Standards framework for collecting, moderating and publishing online consumer reviews, on 21 November 2022. It applies to every platform that publishes consumer reviews and sets out ways to verify a reviewer, such as email, phone or SMS. It began as voluntary; the government said that once it is made mandatory, violations may be treated as unfair trade practices.
Both sets of rules cover businesses and platforms. Nothing moderates the testimonials a stranger pastes into your chat.
Who it fails, and how
Big platforms at least try to filter reviews. The harder case is the seller who has no platform: an Instagram page, a WhatsApp catalogue, a Facebook group listing. Their reviews are whatever they choose to show you, often screenshots with names cropped out.
Twenty perfect screenshots and no way to reach a single buyer is not twenty references. It is one source, the seller, speaking twenty times. Our guide to fake payment screenshots shows the same problem with money.
What makes a review worth trusting
We think a review earns trust in proportion to how hard it would be to fake. That comes down to two ties: to a real transaction, and to a real person who could be found and asked.
- Tied to a purchase the platform recorded, not just posted by anyone.
- From reviewers with their own history, not accounts that exist only to praise this seller.
- Specific: what arrived, how long it took, what went wrong and how it was fixed.
- Mixed. A real seller collects some three-star reviews, and how they answer them tells you more than the five-star ones.
- Best of all, from someone you know who bought from them, or a buyer you can contact yourself.
Questions
Is a verified purchase badge proof a review is real?
Are AI-written reviews illegal?
Isn't 4.8 stars from thousands of reviews safe enough?
Realr rests on the same idea as a good review: trust tied to a real person. Someone proves who they are once, and you can ask to see that before you pay.
Ask before you paySources
- FTC: final rule banning fake reviews and testimonials (14 August 2024)
- FTC: The Consumer Reviews and Testimonials Rule, questions and answers
- PIB: Centre launches framework to protect consumers from fake and deceptive reviews in e-commerce (IS 19000:2022), 21 November 2022
- He, Hollenbeck and Proserpio, The Market for Fake Reviews (working paper; published in Marketing Science, 2022)
- Muchnik, Aral and Taylor (2013), Social influence bias: a randomized experiment, Science



