How do you detect fake reviews in your own profile?
By the Ingenious AI team · Published 14 November 2026 · Reviewed 14 November 2026
Run the free review audit See pricing · See a sample provenance report
Short answer
You detect suspect reviews in your own profile by auditing the rows themselves: look for gating language, incentive markers, duplicated text, same-day bursts and AI-generation wording patterns, and check each suspect row for a transaction link. ReviewProof computes exactly those five signals on your imported reviews and traces every flag back to the row that triggered it — as signals for human review, never as automatic verdicts.
What is fake review detection?
Fake review detection is the practice of screening review rows for patterns associated with reviews that were bought, incentivised, gated, duplicated or generated, rather than written after a genuine transaction. Detection works on evidence inside the data — wording, timing, repetition and provenance — and its output is a shortlist for human judgement, because no wording pattern alone proves who wrote a review or whether they bought anything.
Start with your own file, not the public page
The public profile hides the very fields detection needs. On the page you see stars and prose; in your own export you can hold the date, the source, the reviewer handle and, crucially, whether a transaction sits behind the row. Importing your reviews into ReviewProof turns a vague worry ('some of these look off') into a countable picture: how many rows carry gating language, how many repeat another row's text, whether a burst day exists, and how many reviews have no transaction link at all. Detection that starts anywhere else is guesswork with better graphics.
ReviewProof computes every signal from the rows you import, never from a reputation guess about your business. The risk score is 100 × (gating × 1.0 + incentive × 1.0 + duplicate × 0.8 + AI-signal × 0.6) ÷ (2.6 × number of reviews), plus 15 points when a burst is detected, capped at 100. Each flagged review in the report names the imported row that triggered the signal, so a human can open that row, read it in context, and decide. The product labels these patterns signals for human review, not verdicts — and that restraint is deliberate, because a heuristic that overclaims would be its own kind of false evidence.
Read flags as a shortlist, not a sentence
Each signal has an innocent explanation somewhere. A burst can be a grand opening. Similar wording can be a coach-party that genuinely shared a phrase. Incentive language can be a reviewer describing a loyalty discount that had nothing to do with the review. That is why the report's discipline matters: signals are weighted, summed into a risk score, and presented beside the row so a person decides. The decision, whatever it is, is then logged in the provenance chain — which means your detection process itself becomes evidence of good faith if the profile is ever questioned.
In practice the routine takes minutes. Export or copy the reviews for the location, import them into ReviewProof by CSV or paste, and read the report top-down: total rows, flagged count, risk score, then the per-signal breakdown. Open the flagged rows themselves — the report traces each signal to its row precisely so you can do this. Decide moderation openly (kept, published, or flagged internally; there is no delete), check which reviews carry a transaction link, and publish or refresh the verified widget so the public claim matches the audited record. On Pro, export the evidence pack while the file is fresh.
Mistakes that undo the work
The classic mistake is deleting suspect rows on sight: it destroys the evidence, can breach platform rules, and under the FTC rule suppression creates a second violation while hiding the first. The second mistake is accusing a reviewer publicly on the strength of a wording signal. The third is auditing once, at crisis time, when a monthly import rhythm would have caught the pattern in its first month.
Two design choices keep the detection honest. First, nothing is scored until you import it: there is no scraping of your public profiles and no third-party score blended in. Second, the weights reflect what the FTC fake-review rule (16 CFR Part 465) actually prohibits — how a review was solicited weighs heaviest (gating and incentive markers at 1.0 each), while wording patterns weigh less (duplicates 0.8, AI-generation patterns 0.6). A burst of same-day reviews adds a flat 15 points because timing is visible in the dates you supply. You can reproduce every number in the report from your own file, which is exactly the point.
Worked example: a dental clinic that found its own problem first
A fictional worked example, written to show the numbers and the decisions in practice.
A dental clinic imported 214 reviews before a planned advertising push. The report flagged 9 rows with incentive markers — every one traced to a single month when a receptionist had offered a discount for reviews — plus 6 duplicate-text rows and a burst day holding 41% of the file. The clinic's risk score came out at 31. Instead of deleting anything (ReviewProof has no silent-delete path; the rows stayed, flagged internally), the clinic stopped the offer, retrained the desk, and kept the audit. Finding your own problem first, with the record intact, is the cheapest version of this story.
Your five-action checklist
- Export the location's reviews with dates, sources and reviewer handles
- Import them into ReviewProof by CSV or paste and read the per-signal counts
- Open every flagged row and judge it in context before deciding anything
- Check transaction links on the flagged rows — present, or honestly UNLINKED
- Record each moderation decision in the chain; flag internally, never silently delete
Frequently asked questions
What is the strongest single signal of a fake review?
There is no single proof. In ReviewProof's scoring the heaviest weights sit on gating and incentive markers (1.0 each), because they describe how the review was solicited — which is what the FTC rule prohibits.
Can detection prove a review is fake?
No. Signals are heuristics that shortlist rows for human review. Provenance — a transaction link or its absence — is the harder evidence, and even that speaks to backing, not to the writer's honesty.
Should I delete reviews the audit flags?
No. Keep them, flag them internally, and log the decision. Silent deletion is suppression, destroys your evidence, and is itself prohibited conduct under the fake-review rule.
How often should a business run detection?
On a rhythm, not on a panic: a monthly import for most locations, plus an extra audit before campaigns, listings or any moment the profile will be examined by others.
Keep reading
Related guides from the ReviewProof library:
- How can you prove a review is real?
- What is review gating, and how do you stop doing it?
- What counts as an incentivised review?
- What does duplicate review text tell you?
- What is a review burst, and when is it innocent?
- Can you tell if a review was written by AI?
- How is the ReviewProof risk score calculated?
Start with the tool itself: open the free audit desk, see how ReviewProof works, compare Free, Pro and Agency plans, inspect the sample provenance report, or browse all ReviewProof guides.