How the review rule check works
The audit runs one rule check: a fixed list of word patterns, each pointing at one Trustpilot or Google rule, with the matched words shown for every result.
100% risk-free. You only pay for the reviews that come down.
How does the review rule check work?
It is a heuristic: a fixed list of word patterns, each pointing at one Trustpilot or Google rule. A strong match makes a review worth flagging, a weak match means a person should read it, and no match means it likely stays. It reads only the words, so the platform decides every report.
A heuristic, and what that means
The check is a fixed list of 33 word patterns. Each one points at a Trustpilot rule, a Google rule, or both. When a pattern matches, the result shows the words that matched and the rule they point at. Nothing is scored as certain, and nothing is learned from past reviews.
The same code runs in every tool on this site, the free audit, the Chrome extension, the API and the embeddable checker. It reads words, not facts. That is why it can only say a review may break a rule. Trustpilot and Google make every decision.
The three verdicts
- Worth flagging
- At least one strong signal matched. The text on its own shows content a rule names, such as a promo code, a phone number or “never used them”.
- Needs a human look
- Only weak signals matched. The text points at a rule, but whether it breaks it depends on context, such as whether a name is a doctor who trades under it.
- Likely stays
- No signal matched. Most real complaints, however harsh, stay up, and this is the verdict for most 1-star reviews.
Every signal, by rule
This list is built from the code at every release, so it always matches what the check does. Strong means the text alone shows the content. Weak means it depends on context.
Trustpilot
| Rule | What the check looks for |
|---|---|
| Advertising or promotional |
|
| Personal information |
|
| Not based on a genuine experience |
|
| About a different business |
|
| Harmful or illegal |
|
| Fake reviews |
|
| Incentivized reviews |
|
| Who can write a review |
|
| Rule | What the check looks for |
|---|---|
| Spam and fake content |
|
| Conflict of interest |
|
| Off-topic |
|
| Personal information |
|
| Offensive content |
|
| Advertising and solicitation |
|
| Restricted content |
|
| Impersonation |
|
| Dangerous content |
|
| Illegal content |
|
| Sexually explicit content |
|
| Terrorist content |
|
What the check cannot know
- Order records: whether the reviewer ever bought from you, and when.
- Identity: who wrote the review, and whether they are a competitor, a former employee or a real customer.
- Anything off the page: replies, emails, support tickets, photos, or what happened after the review was written.
- What the platform will decide. It names a rule; Trustpilot and Google apply it.
How the rules are sourced
Every rule carries a short quote from the platform’s own page, the link to that page, and the date the quote was last checked. Each rule page on this site shows them. The last verified dates are below.
- Trustpilot: 8 rules, last verified .
- Google: 12 rules, last verified .
How the policy monitor keeps them current
Every day at 07:17 UTC, a scheduled job opens the 7 Trustpilot and Google policy pages the rules come from and compares their text with the last copy. When a page changes, it opens an issue with the difference. A person then reads the change, updates the rule, its quote and its date, and adds the change to the review policy change log. The check itself never changes without that review.
Related policies
How pages and statistics are written and sourced is in the editorial policy. How we research and date the rest of the site is in our research methodology. To run the check on your own reviews, use the review self-audit. The code is open: see the developer docs.
Find out how many of your 1-star reviews we can remove
Send your Trustpilot or Google profile link. The audit is free, and you only pay for the reviews that come down.
- Free audit
- Pay only for removed reviews
- Proof for every removal