How can you tell a real reputation from a manufactured one?

The trail is the tell. A genuine reputation is built from real transactions, so it has a natural shape: reviews that arrive at the pace of real custom, a spread of ratings rather than a flawless wall, reviewers who are real people with histories, and the same broad picture across more than one place the business cannot edit. A manufactured one has to fake that shape, and faking it leaves marks. None of those marks is proof on its own. Several of them together, across different categories, is when you slow down.

And underneath all of it sits one question that does most of the work, whether you are checking a supplier, a competitor, or your own claims before you make them: what can they show you that they did not produce themselves?

The one question that does most of the work

A screenshot, a star rating, a follower count: every one of those was produced by the business, on its own device, and chosen by the business from whatever looked best. That is not proof. It is a claim with good production values. Real proof is the evidence a business does not control. A neutral platform it cannot edit. References you pick and contact yourself. A spread of outcomes across all its customers, not the one in the testimonial.

So the test, for any reputation including your own, is this. Could they show, in an afternoon, the real job behind ten reviews picked at random? An honest operator can, and is usually glad to. A manufactured reputation has nothing behind the numbers, which is the entire reason it is fragile. Asking is not rude. It is the one thing the whole machine is built to make you skip.

Read the shape, not the headline number

• The pace. Sort the reviews newest first and look at the rhythm. A genuine business earns feedback gradually, so a long quiet stretch broken by a sudden cluster, especially a wall of five stars from a standing start, is a question. The limit: a real promotion, a viral moment or a single "please review us" email also cause a spike. The pattern is a prompt to look closer, not a verdict.
• The spread. Real customers do not all give five stars. Large studies of millions of reviews find that trust and sales actually peak when the average sits around four-point-two to four-point-seven, and fall as it approaches a perfect five, because people read flawless as too good to be true. A profile that is all five stars with almost no middle is the shape of a manufactured sample. The limit: a genuinely excellent small business, or a very new one with few reviews, can sit high and honest. High is not the flag. Unnaturally flat is.
• The reviewers. Thin, brand-new accounts that have only ever reviewed one business, or that have rated dozens of unrelated things across the world in a week, are a question. The limit: plenty of real, privacy-minded customers keep bare profiles, and a small business's earliest reviewers are genuinely first-timers. No history is a question, never a verdict.
• The detail and the cross-check. Do the positive reviews name the actual job, or could they describe anything? Does the same business look broadly the same on a second platform it cannot curate? One glowing profile that contradicts a mixed picture elsewhere tells you which one was managed. The limit: different platforms attract different customers, and you cannot tell a human review from a well-written one by feel alone.

The suspicion scorecard, so you do not overreach

Mark each check green, amber or red. Treat a case as genuinely suspicious only when several reds land across different categories at once: clustered timing and templated wording and thin reviewer profiles and one outlier platform. That is how regulators and the research models work too, on combinations of signals, never on a single smoking gun. One red on its own is a coincidence waiting for an explanation. Four reds, pointing the same way, is a pattern.

Go deeper on any one check

Each of these has its own page, with the full checklist and the limit of every signal:

• How to spot fake reviews
• How to spot bought followers and fake engagement
• How to spot fake streams and views
• How to read a star rating, and when a perfect score is a red flag
• How to check a business, supplier or contractor before you pay

And one signal sits next to the reviews: the badges. Trust marks, accreditations and "as seen in" explains which ones mean something, which are bought, and how to make any badge verify in one click. New to any of the terms here? The glossary defines every one in a sentence.

The limit, because every check has one

You cannot detect a faked reputation by reading harder, and that matters more every year. Researchers have found that reviews written by current artificial intelligence are, in one 2025 preprint (not yet peer-reviewed), hard to tell apart from real ones to both ordinary readers and the automated detectors built to catch them. So treat any tool that scores a review as fake or real as an input to your judgement, never a verdict, and lean on the things software cannot easily fake: the shape of the pattern over time, and the proof a business can or cannot put in your hand.

What a real reputation looks like

This is the reassuring half, and it is the more useful one. A genuine reputation has a natural spread of mostly good reviews with some honest criticism, answered like a grown-up. It shows up consistently across more than one independent platform. It carries specific detail, the named job, the real date, the actual problem solved. And the operator behind it can produce the working: references you can ring, a track record you can confirm, photos or records that line up in time and substance with what the reviews say.

The checks on this site are built so that the genuine operators pass them easily and the manufactured ones cannot. If you are the genuine operator, that is the point. The same trail that lets a careful customer trust you is the trail a faked reputation can never lay down.