Guide
How we verify reviews
The signals, tools, and human judgement we use to keep reviews on Ratecrest real.
1. Why this matters
A review platform is only as useful as the reviews on it. If most of what you read is fake, the platform is worse than useless: it actively misleads people into bad decisions. We take that risk seriously, because we know what is at stake. Nigerians lose money to fake reviews on other platforms every week.
This page explains, in plain language, what we actually do to keep reviews real. We do not claim to catch everything. No platform does. We do claim to work harder at it than most, and to be honest with you when something gets through.
2. The signals we watch
When a review is posted, our systems look at a wide set of signals before, during, and after publication. The list below is not exhaustive (publishing the full list would help bad actors avoid it) but it covers the main areas.
2.1 The reviewer
- How long the account has existed.
- How many reviews the account has written, across how many businesses.
- Whether the account has been flagged before, and the outcomes.
- Whether the account is verified (NIN, BVN, or ID document).
- Patterns in posting time, frequency, and device.
2.2 The review itself
- The text. Reviews that are very short, very generic, copy-pasted from elsewhere, or written in obviously machine-generated language get extra scrutiny.
- The rating. A 5-star or 1-star review from a brand-new account is more likely to be a plant than the same review from an account with a long, varied history.
- The photos, if any. We check for stock images, photos used on other businesses, and metadata that does not match the claimed experience.
- Edits. Quietly editing a 5-star review into a 1-star review weeks later is a known fraud pattern.
2.3 The business being reviewed
- Whether reviews are arriving in a normal pattern or in sudden bursts.
- Whether multiple reviewers share devices, networks, or other fingerprints.
- Whether the same review text is appearing on competitors.
- Whether the business is offering rewards in exchange for reviews (review trading is a serious breach of our Terms of Use).
3. The verified reviewer badge
Reviewers can choose to verify their identity using a Nigerian identification document. We support NIN and, where applicable, BVN-linked verification. Once verified, a small badge appears next to the reviewer's name on every review they post.
Verification is optional. Honest reviewers should not be forced to share government IDs to be heard. But the badge is useful, because it tells readers that the person behind the review is real, and that we have evidence of who they are if a dispute escalates.
Verification details are stored under the strict rules in our Privacy Policy. Business owners do not see your NIN or BVN. They only see whether the badge is present.
4. What happens after you post
Reviews are published right away. Holding everyone in a queue would slow honest reviewers to a crawl while doing little to stop determined fraudsters. Moderation happens after publication.
Within seconds of posting, automated checks run. A review that matches several high-risk signals is held back from search results and other prominent surfaces while a human takes a closer look. Most reviews clear this in minutes. A small number get removed, and the reviewer is told why.
Over the following days and weeks, the review is re-evaluated as new information comes in: new flags, more posts from the same account, patterns that only become visible at scale. We do not freeze a verdict on day one. We follow the evidence.
5. What we cannot detect
We are honest about the limits of any verification system.
- We cannot tell whether a single review describes a real experience just by reading it. A skilled fake reviewer can write convincingly. A real customer can write badly.
- We cannot always tell when staff at a business write 5-star reviews from personal accounts, especially the first time.
- We cannot detect every coordinated campaign on the first day. Coordinated campaigns become visible as patterns emerge.
- We cannot see what happens off the platform. Cash-back deals offered through WhatsApp groups in exchange for reviews are caught when their effects appear on Ratecrest, not when the offer is made.
Where we fall short, we want to know. The flag link next to every review is the fastest way to tell us, and we read every flag.
6. What we do when we find a fake
The action we take depends on what we find.
- A single suspect review. Removed, the reviewer is told, and an appeal is allowed.
- A pattern of suspect reviews from one account. The account is suspended or closed, and the reviews are removed.
- A coordinated campaign attacking a business. All the campaign reviews are removed, the accounts behind them are closed, and we may publish a short note on the affected business's profile explaining what happened.
- A business gaming the system to inflate its rating. The fake reviews are removed, the business loses paid features, and a public moderation notice may be added to the profile. Repeat offences end the business's presence on Ratecrest.
- Anything illegal. The matter goes to our senior team and, where warranted, to the relevant authorities under Nigerian law.
7. How you can help
Verification works best when our systems and Nigerian readers work together.
- If a review reads as obviously fake, flag it. Be specific in the flag. "Reviewer never used this business" is more useful than "Suspicious".
- If you are asked to write a review in exchange for a discount, a refund, or anything else, tell us. Asking for incentivised reviews is a breach of our rules, and we want to know about it.
- Verify your account if you are comfortable doing so. Verified reviewers carry more weight in moderation decisions.
- If you spot a business asking customers to "tell us privately first if you would leave a low rating", that is review gating. Flag it.
8. Transparency and accountability
We aim to publish an annual transparency summary covering flag volumes, content removed, accounts actioned, and patterns we have spotted. This sits alongside the Content Moderation Policy.
If you think we have got a specific decision wrong, every removal can be appealed once, and a different member of our team looks at the appeal. The detail sits in the Review Policy.
9. Questions about verification
The contact form is the right way to reach us. Write "Verification" in the subject so the right team picks it up.