"Blockchain-verified reviews" sounds like the kind of phrase invented to sell something. Sometimes it is. But underneath the buzzword there's a simple, useful idea, and it's worth understanding exactly what it can and can't do before you trust any badge that claims it.
We built this into Reputfy as Trust Shield, so we'll explain it the way we'd want it explained to us.
The problem it's trying to solve
Reviews can be changed after the fact. A platform can edit them, hide them, reorder them. A business can display a curated selection on its own website and call it "our reviews". A reviewer can change their story. And readers have no way to check any of it.
Verification is about one narrow question: is this review the same as it was when it was collected, and did it exist by a certain date?
That's it. Not "is it true". Not "is it fair". Just "has it been tampered with".
Step 1: A fingerprint for each review
Take the text of a review, plus its rating, date and a reference to where it came from, and run it through a hash function called SHA-256. Out comes a 64-character string like:
9f2c1e…a4b7
This string is the review's fingerprint. Two properties make it useful:
- The same input always gives the same fingerprint.
- Change anything at all, even one comma, and the fingerprint changes completely.
You can't work backwards from a fingerprint to the review. You can only check whether a given review matches it.
Step 2: Bundle the fingerprints together
Publishing thousands of fingerprints one by one would be slow and expensive. So they're combined in a structure called a Merkle tree. Pairs of fingerprints are hashed together, then pairs of those, and so on, until you're left with one fingerprint at the top: the root.
The root depends on every review underneath it. Change one review and the root changes. And for any single review, you can prove it's part of the tree with a short list of the neighbouring fingerprints along its path to the root, without revealing the other reviews.
Step 3: Anchor the root somewhere public
Now the root needs to be stored somewhere no single company controls, with a trustworthy date attached. That's where a blockchain comes in.
Reputfy uses OpenTimestamps, an open standard that commits fingerprints into the Bitcoin blockchain. Once a root is included in a block, it's effectively impossible for anyone, including us, to change it or backdate it. Anyone can check the timestamp with free, open-source tools. It doesn't depend on our servers being up or on trusting our word.
What you can verify
With the review, its path to the root, and the anchored root, anyone can check:
- Integrity. The review hasn't been edited since it was sealed.
- Existence by a date. It existed no later than the time the root was anchored.
- Membership. It was part of a specific batch. If a business claims to show "all reviews from March", you can check whether the batch matches what's shown.
That third point is quietly powerful. It makes it much harder to bury bad reviews and present the rest as the full picture, which is exactly what review gating and review suppression try to do.
What it doesn't prove
This is the part most marketing skips.
- It doesn't prove the review is honest. A fake review, fingerprinted, is a perfectly verified fake review. Verification proves nothing about who wrote it or whether it happened. Detecting fakes is a separate job, covered in how to spot and remove fake reviews.
- It doesn't prove the reviewer was a customer. Some systems link reviews to a verified purchase or booking. That's a different check, and a badge should say clearly if it includes it.
- It doesn't make reviews permanent or public. Only fingerprints are anchored. The review text lives wherever it was posted. If it's deleted there, the proof still shows that it existed and what it said, but only to someone who has a copy.
- It doesn't apply backwards. Reviews written before sealing started can only be sealed from the date they were collected.
Why bother, then?
Because trust in reviews is falling, and fakes are now a legal issue as well as an ethical one, as we cover in the FTC fake review rule, explained. A business that can say "here are our reviews, unedited, and you can check that yourself without trusting us" is offering something real. It's a small, honest claim, and small honest claims are rare in this industry.
It also matters more as AI assistants summarise reviews instead of showing them. If the source material can be checked, so can the summary. We touch on that in how AI assistants pick local businesses.
Try breaking one
The fastest way to understand this is to change one letter in a review and watch the fingerprint fall apart. Every review on Reputfy's public business pages shows its fingerprint, and our homepage has a sealed example you can try to tamper with. You can also write a review and see the fingerprint it gets.
