Revdict.
Trends

Singapore's regulator made 45 businesses apologise for buying fake reviews

Andrew Keymaster · Founder, Revdict 5 min read
Singapore's regulator made 45 businesses apologise for buying fake reviews — Revdict
Table of contents
  1. What did the CCS announce?
  2. How many businesses were involved, and what happened to them?
  3. Why is the six-month apology the interesting part?
  4. What does this say about how review selling works?
  5. Could our data have spotted a case like this?
  6. What should owners and customers take from the case?
  7. Frequently asked questions
  8. Is buying fake reviews illegal in Singapore?
  9. Did the CCS fine anyone?
  10. Were the businesses named?
  11. Can Revdict tell whether a business bought reviews?
  12. What should I do if I find a suspicious review on my own listing?

Singapore's Competition and Consumer Commission (CCS) announced on 30 September 2026 that 45 businesses had given formal undertakings after buying fake reviews from a single provider. Each must remove the fake reviews and publish an apology for six months. The provider has to stop selling the service, apologise too, and donate its proceeds.

The case is a rare look at a review-selling operation from the regulator's side, and it answers a question owners keep asking: what happens to the customers, not just the seller?

What did the CCS announce?

The CCS investigated a provider of fake reviews, identified in the public summary only as "T", and the businesses that bought from it. According to a law firm's summary of the media release, T supplied reviews presented as genuine customer feedback to businesses that had agreed review numbers with it. The text was either written by T or supplied by the business itself.1

The regulator found that consumers could reasonably have been misled about how popular or well regarded those businesses were. It did not publish a count of fake reviews, and the summary does not name the statute it relied on. Treat any figure beyond the ones below as unconfirmed.

How many businesses were involved, and what happened to them?

T served about 100 businesses. The CCS looked at 47 of them in a first phase, and 45 of those gave undertakings. Two declined, and the CCS says it will keep investigating them and may take firmer enforcement action.1 A second phase will cover other businesses that may have used the same provider.

Party Outcome reported
45 businesses Undertakings: stop buying or posting fake reviews, find and remove existing ones, report progress to the CCS, strengthen compliance, publish an apology for six months
2 businesses Declined to give undertakings; investigation continues
T (the provider) Stop providing the service, publish an apology for six months, donate proceeds to a charity of its choice
Other providers and businesses Invited to come forward; cooperation before a formal investigation will be "viewed favourably"

The summary mentions no fines. An undertaking is an agreement to change behaviour, which is not a finding of guilt in court, and the businesses involved are not named in the material we found.

Why is the six-month apology the interesting part?

The apology is aimed at the audience that was deceived. The orders require it on websites, social media and physical premises, so the customers who relied on the inflated ratings are the ones who see the correction.

For a small business that is a heavier cost than it sounds. A fine is paid once and forgotten. A prominent notice saying you bought reviews sits next to your name for half a year, and it is easy for a competitor to screenshot. The CCS is also asking for the existing reviews to be found and removed, which means the owner has to audit its own listing and report back.

What does this say about how review selling works?

It shows the shape of the trade. One provider, roughly a hundred customers, and a negotiated number of reviews per client. That is a supply chain, and how review farms price and deliver their work follows the same logic: the buyer picks a volume, the seller delivers it, and the text is either generated or handed over by the client.

It also shows why regulators are going after the buyers. The seller is the obvious target, but the harm lands on consumers through the businesses that displayed the reviews. The UK takes a similar view of businesses that fail to stop fake reviews, as our write-up of the CMA's five open investigations explains. Singapore adds a concrete remedy: remove, report, apologise.

Could our data have spotted a case like this?

Not by itself, and we should be clear about the limit. Our analysis scores public review records for patterns consistent with manipulation. It cannot see who paid whom, and a flagged pattern is not proof that anyone bought reviews.

What it can show is how common the underlying patterns are. As of November 2026 we have analyzed 901 places and 656,518 reviews. Of those places, 604 (67%) came out clean, 243 (27%) suspicious, 35 (about 4%) likely manipulated, and 19 (about 2%) fall in a separate group where businesses appear to be soliciting reviews in organized drives. Across the whole review pool, 136,999 reviews (21%) carried at least one flag. The most common flag by far is a low-activity account, with 121,015 reviews; same-window bursts account for 19,906.

Those are signals, mostly weak ones in isolation. A purchased batch tends to show up as several of them at once on one listing, and no single one proves a purchase.

What should owners and customers take from the case?

For owners: if a vendor offers you a review count, the offer itself is the risk. Both the buyer and the seller were pursued here. The safer route is asking real customers, a topic we cover in our guide to asking customers for reviews legally.

For customers: a high rating with a tidy volume of reviews is not evidence of quality. Check the dates and the reviewer accounts, or paste the Google Maps link into our free fake-review checker and read the evidence it lists.


Frequently asked questions

Is buying fake reviews illegal in Singapore?

The CCS treated it as an unfair practice that could mislead consumers, and required 45 businesses to stop. The public summary does not name the specific legal provision, so check the CCS notice itself before relying on a particular section.

Did the CCS fine anyone?

The summary we reviewed mentions undertakings, apologies and a voluntary donation by the provider, not fines. Two businesses refused to give undertakings and remain under investigation, where firmer action is possible.

Were the businesses named?

Not in the summary we read, which refers to the provider only as "T". We have not named any business here and do not infer who the 45 are.

Can Revdict tell whether a business bought reviews?

No. Our score flags patterns consistent with manipulation, such as bursts and low-activity accounts, and links to the public evidence. It cannot show who paid for what.

What should I do if I find a suspicious review on my own listing?

Report it through Google Business Profile and document the pattern. If you bought reviews in the past, removing them and stopping the practice is the direction regulators here ordered.


  1. Allen & Gledhill, "CCS takes action against provider of fake reviews and businesses procuring reviews" (28 October 2026), summarising the CCS media release of 30 September 2026. ↩ ↩

References

  1. CCS takes action against provider of fake reviews and businesses procuring reviews — Allen & Gledhill (2026)

Tags

fake-reviews singapore ccs consumer-protection review-buying enforcement undertakings

About the author

Written by

Andrew Keymaster · Founder, Revdict

Andrew Keymaster built and runs Revdict, a free forensic checker for Google Maps reviews. Its scoring model has examined 420,533 public reviews across 443 places as of August 2026, flagging the patterns that show up in manipulated review records — same-day five-star bursts, throwaway reviewer accounts, and near-duplicate wording. He writes here about how those patterns are detected, what platforms will and won't act on, and what the data says about how common fake reviews actually are.

Fake-Review Signals Across 825 Places: Our Data, October 2026 — Revdict

Trends

Fake-Review Signals Across 825 Places: Our Data, October 2026

As of October 2026, 570 of 825 analyzed Google Maps places come back Clean and 28 reach Likely Manipulated. Here is how the verdict split, score distribution and signals changed since our July release.

Andrew Keymaster · Oct 19, 2026 · 5 min