Can Google spot a fake review?

You’re looking at five one-star reviews for a local small business. Are they real or generated by a 'bot farm' 2,000 miles away in the hills of Montana?

Learn how Google evaluates negative reviews to separate truthful customer complaints from fakes, helping businesses protect their online reputation.

But first, a story...

Once upon a time, long ago and far away, I worked for The Owner.

He was real friendly, and offered me a job in a market where other employers thought that me being past age 40 was a deal-killer. The Owner described how he was asking me to build a new department inside his business, with a different name than his business, and it would be very successful.

And his office manager was real friendly, because she came walking over to my desk instead of sending email to explain where to find blank timesheets, and why my tax forms would look different than what I was used to. 🚩Red flag warning: concerning behavior

Several months go by. Work, work, work.

The Owner comes around and says he's got this WEIRD problem that he just can't figure out. On the web, a bunch of people have left nasty Google reviews with low stars. He has NO IDEEEA who those people are. He asks if I can talk to Google and have those lies taken down.

I say SHORE! Because, the guy who offered me a job doesn't deserve some crazy talk online!

I rummage around and find the Google steps for flagging a bad review and submitting it for removal. I start going down the list. Google is not quick on this; it takes days to start getting replies from the mega-corporation. And you have to go through their interface: No telephone calls or emails.

Each request is denied. I search and search for a reason why.

The Owner is disappointed. He insists that the negative reviews are all FAAAAKE.

There is one exception, a person who The Owner says used to work there, and about whom The Owner says "He was a CONNNNTRACTOR", like that explains everything. But I don't catch his meaning. 🚩Red flag warning: Concerning behavior

Eventually, we decide that it's a mystery and move on to other things.

Years later, a random email conversation with the HR department tipped me off that The Owner was scamming me. I resigned, and he withheld a small mountain of my wages.

I also discovered the bad reviews were either from former employees or from people with solid reputations.

And the personal visit by the office manager when she talked about timesheets and tax forms left no paper trail.

And the “different” tax form was a  1099.Some companies use an IRS 1099 form to retroactively change an employee into a contractor. A contractor has little to no legal rights.

Google's AI explanation:

It’s one thing to suspect a review is fake, but it’s another to realize that Google’s systems "knew" something was wrong...

Google uses a "behavioral fingerprinting" model:

1. The common thread (network mapping)

Google doesn't just look at the review; it looks at the graph of connections between the reviewer and the business.

Shared infrastructure: Even if they used their own phones, employees who logged into the office Wi-Fi or used a company-issued device are "tethered" to that business entity.

"Review rings." If different people all worked at the same place (based on their GPS data) and then all happen to leave a review, the probability of that being a coincidence drops to nearly zero.

2. Geolocation history and entity recognition

It's like a digital fingerprint or a "security camera" that notices 20 people entering a store through the same back door at midnight.

If a reviewer’s phone spent 8 hours a day, 5 days a week at your boss's office for three years, Google categorizes them as Staff, not a Customer.

Conversely, for a legitimate service business, Google expects to see the reviewer’s device present at the business around the time of the transaction.

3. Linguistic fingerprinting

Just as we can tell when a smile is fake, Google looks for over-enthusiasm or repetitive phrasing that real humans rarely use.
Google’s Natural Language Processing (NLP) can detect when different reviews share the same sentence structure or "vibe."

4. Why the denials were vague

When you requested the removals, Google’s replies were likely generic (e.g., "Does not violate our policies") because of an internal mandate: They do not want to coach scammers. If Google told your boss, "We're keeping these because we know they are former employees based on their GPS history," it would teach bad actors how to hide better next time (e.g., by turning off GPS).

The irony of your situation:

While you were being asked to "fix" his reputation, the system was actually protecting the truth because it recognized the insider nature of the complaints.

Since you mentioned he was scamming you as well, those vague denials from Google were actually the algorithm working exactly as intended—preventing a business owner from silencing legitimate grievances from people who knew the truth.”

‘So what?’

People just want to find a good deal

Google's goal isn't just to rank websites. It's trying to be sure that you don't accidentally buy a lemon.

Filtering bad and good reviews prevents "review bombing" where competitors try to hurt a small business.

Targeted concerns how the wealthy and well-connected target regular people, and how people are pushing back.

Report problems

Contacting me is a much more efficient and honest way of addressing issues than bringing a lawsuit against the author(s). Bringing a lawsuit instead of asking for a revision reveals the motives of those who abuse the legal system to intimidate citizen journalists while seeking to retain their employees’ stolen wages.

I may be reached via the Contact page or on Bluesky or Facebook.

NOTE: This website has always included my contact information.

Comments?

Please use the button below to comment on this article...