How to spot a fake or bought review

Short answer

Read the distribution rather than the average. Bought reviews cluster — several arriving close together, short, generic, five stars, often praising the business by full name in a way real customers rarely do. A perfect score across a handful of reviews is far weaker evidence than a slightly imperfect one across hundreds.

Reviews are the most-used signal in hiring a tradesperson and the easiest to manufacture. You cannot verify them individually, but the patterns are visible if you look at the right things.

What to look at

  1. Volume before score. Five stars from six reviews tells you almost nothing; 4.6 from four hundred tells you a great deal.
  2. The shape of the distribution. Real businesses accumulate a few threes and fours. An unbroken wall of fives on a business with real volume is unusual.
  3. Timing clusters. A quiet history with a sudden burst of five-star reviews in one fortnight is the classic signature.
  4. Specificity. Real reviews mention the job, the suburb, what went slightly wrong and how it was handled. Bought ones praise professionalism in the abstract.
  5. Reviewer history. Accounts with one review, created the same week, reviewing businesses in unrelated cities.
  6. How the business answers criticism. A measured reply to a bad review is worth more than the review itself.

What this site cannot check, stated plainly

We hold review counts and averages. We do not hold individual reviews or their dates. That means our ranking cannot see when reviews arrived — forty earned steadily over six years and forty earned in a fortnight score identically here, and the second is far more likely to have been bought.

We publish that as a limitation rather than implying a check we do not run. It is also why review volume is weighted on a log scale rather than linearly, and why volume counts for more than score: neither is manipulation-proof, but volume is considerably more expensive to fake convincingly.

Signals that survive review manipulation

The reason this site checks ABN registration, contact-detail agreement and independent corroboration is that those are hard to fake and nobody thinks to. A business can buy fifty reviews far more easily than it can make two independent mapping sources agree about its address. Across 36,230 businesses, 94.1% are corroborated by two or more independent sources.

Common questions

Can businesses pay to remove bad reviews?
Not legitimately from the major platforms, though they can dispute reviews that breach platform policy, and some are removed that way. What happens more often is dilution — a burst of new positive reviews pushing a bad one off the first page, which is why reading the distribution beats reading the top few.
Is a five-star rating a bad sign?
Not by itself. It is weak evidence when the count is low, and unusual when the count is high. The question is not whether the score is perfect but how much evidence sits behind it.
Why does this site rank on review volume rather than rating?
Because a rating without volume is noise. The median business in this directory has 29 reviews; treating a 5.0 from four as equal to a 4.6 from three hundred would rank whoever asked their family. Volume is log-scaled so that going from 10 to 100 reviews matters far more than 400 to 500.

Read next