Reviews and testimonials: what you may and may not say

Reviews occupy a strange legal position: they are usually written by customers, published on a platform, and yet the business is responsible for much of what surrounds them — how they were obtained, which ones are shown, and what the business does with them.

The four things that are clearly prohibited

• Fake reviews. Writing reviews as if a customer, or commissioning them. This now attracts direct civil penalties in several regimes rather than just regulatory guidance.
• Buying a specific rating. Paying for reviews whose content you control is the same problem in a different form.
• Suppressing negative reviews. Publishing only reviews above a threshold, or hiding negatives selectively, misrepresents the customer experience.
• Using a review as proof of typical results. An exceptional outcome presented as evidence of what customers usually get is a misleading claim even when the review itself is genuine.

What the recent legislation actually changed

The United States Federal Trade Commission published a final rule prohibiting fake reviews and testimonials, which took effect in October 2024 and allows civil penalties for knowing violations. It covers buying and selling fake reviews, using insider reviews without disclosure, suppressing negative reviews, and misrepresenting review metrics.

In the United Kingdom the Digital Markets, Competition and Consumers Act 2024 names fake reviews explicitly among the practices it prohibits, with unfair commercial practices provisions applying to conduct from 6 April 2025, and gives the Competition and Markets Authority direct enforcement powers including fines.

In the European Union, the Unfair Commercial Practices Directive was amended by the Omnibus Directive to add explicit rules on reviews, including obligations to verify reviews are genuine and to disclose how they are collected.

Soliciting reviews lawfully

1. Ask everyone, or ask a defined group and say what that group was.
2. Do not screen by expected sentiment — offering a voucher only to customers you expect to be positive changes who reviews.
3. Do not make a review a condition of receiving something, where the review content then influences whether they get it.
4. Do not offer anything whose value depends on the rating given.

Asking a happy customer is fine. Deciding in advance who is likely to be a happy customer, and only asking them, is the practice regulators call cherry-picking, and it misrepresents the distribution of customer experience.

Testimonials are not reviews

A testimonial is selected and arranged by the business; a review is published by the customer on a platform. The obligation differs, and the weaker standard applies to neither where the outcome claimed is exceptional.

Where a testimonial describes a result, the standard is whether that result is typical. If it is not, the testimonial needs to say so, because otherwise it is a claim about typical outcomes made through a real person.

This is a particular problem in health-adjacent categories, treated in health and wellness claims.

Handling a bad review

Answer it once, factually, without arguing about the customer experience. The response is read by future customers more than by the reviewer, and its purpose is to demonstrate how you behave rather than to win.

Where a review is false or defamatory, there is usually a platform process followed by a legal one. Pursuing either is a decision to be made slowly, because the process is public and expensive and the review is often less damaging than the fight.

The claims framework the whole area sits on is in advertising claims and substantiation. For paid recommendation, see sponsorship disclosure.

Review metrics are claims too

“Rated 4.8 from 2,000 reviews” is a factual claim, and it is false if the reviews were solicited only from people likely to be positive, or if some were removed for being negative. The number is not a defence; the collection method is the claim.

The same applies to thresholds. Displaying “4.8 average” computed from reviews above four stars, while others sit unshown, is a misrepresentation of the metric even though every individual review displayed is genuine.

Platform incentives run against you

Most platforms reward volume and rating, which pushes sellers toward exactly the practices the rules prohibit. Knowing this is useful: the pressure to filter is structural rather than a personal temptation, and the fix is to write the collection policy before the pressure arrives.

A published policy — everyone is asked, nothing is removed for being negative, removals are for rule breaches only — is both a compliance position and a marketing asset. Customers increasingly look for it.

Incentives that make reviews honest

If a review programme rewards volume, it will produce volume and it will produce pressure to filter. The structural fix is to reward participation rather than rating: ask everyone who bought, report the response rate alongside the average, and let the number be what it is.

Reporting the response rate alongside the average is a small disclosure with a large effect, because it tells the reader how much of the customer base the figure represents. An average from a tenth of customers and an average from most of them are different claims.