How Trustpilot's fake-review detection works in 2026
Trustpilot's detection in 2026 leans on four signals: review velocity, star-distribution anomalies, text similarity, and reviewer-profile consistency. Understanding them is the difference between a campaign that sticks and one that gets wiped. Here is what changed and how to stay on the right side of it.
Velocity is still the number-one trigger
A sudden spike in review volume remains the clearest signal. A profile that averages one review a month and suddenly gets twenty in a day is flagged before a human ever reads the text.
In 2026 the systems weigh velocity against your historical baseline, so the same ten reviews are fine spread over two weeks and suspicious in two hours.
Star-distribution anomalies
Detection increasingly compares your star spread to sector norms. A profile that is 100% five stars stands out against a category where the honest average is closer to 4.3.
A realistic mix — roughly 18% three-star, 41% four-star, 41% five-star — reads as genuine and blends into the category.
Text similarity and profile consistency
Near-duplicate copy across reviews is easy to cluster and flag. Every review needs its own structure, details, and tone.
Reviewer profiles matter too: accounts with no history, mismatched geography, or bursts of activity across unrelated businesses raise the risk score.
What this means for your campaign
The takeaway hasn't changed, only tightened: pace reviews to your baseline, use a realistic star mix, write unique copy, and use profiles that fit your market. A well-run campaign does all four by design.
Run a campaign built to pass detection
View packagesFrequently asked questions
Did Trustpilot get stricter in 2026?+
Detection is more baseline-aware, so velocity relative to your history matters more than ever.
Can reviews be removed weeks later?+
Yes — retroactive sweeps happen, which is why pacing and realism matter beyond day one.
How do I lower the risk?+
Spread delivery, keep a realistic star mix, use unique copy, and fitting reviewer profiles.