Spotting Fake Reviews: 5+2 Patterns You Can Expose in Under 60 Seconds

Monday morning, 8:47 am. A tax adviser from Munich-Maxvorstadt opens his Google Business Profile – and sees four new single-star ratings. All four arrived within 90 minutes, on Sunday evening between 10:15 pm and 11:45 pm. Not a single reviewer has ever written another review. The texts? Short, generic, without a single detail about deadlines, annual accounts or the team.

This is no coincidence. This is systematic.

The question is not whether fake reviews occur – they do, systematically and with increasing sophistication. The question is how quickly you recognise them and what you then do concretely.

Why the problem is bigger in 2026 than it was in 2022

Four years ago, fake reviews were mostly cheap and obvious: poor German, zero context, accounts from Eastern Europe. Google has largely swept that wave away.

What comes now is more sophisticated. Considerably.

AI-generated reviews sound like real people. Coordinated waves spread the pattern over two weeks so that no individual review stands out. And the market for "negative SEO as a service" – paid attacks on competitors' profiles – was one of the fastest-growing grey areas in digital marketing in 2024/2025, according to a Statista survey on online reputation management.

What this means for you: the old detection rules are no longer sufficient. Anyone who only watches for spelling mistakes and foreign-sounding names misses 60% of today's fakes.

The 5 classic fake patterns – and how to spot them instantly

Pattern 1: The one-time reviewer

The reviewer's profile has existed for three months and has left exactly one review – yours. No profile photo, no "Local Guide" status, no history. Sometimes a generic name like "Max M." or "Anna K.".

Micro-example: A barber in Berlin-Mitte receives a 1-star review from an account created the same day. Profile picture: the default Google avatar. Total reviews written to date: 1.

This alone is not sufficient proof. But it is a strong first signal.

Pattern 2: The time cluster

Several negative reviews pile up in a short period – often a weekend or a Monday morning. Organic reviews spread out over weeks. Coordinated attacks do not.

Three 1-star reviews in 90 minutes are not a coincidence. Six in 48 hours from accounts all less than 30 days old are even less so. For more on coordinated attack waves and their local SEO impact, read our piece on review fraud by competitors.

Pattern 3: The generic text

Real customer reviews contain details. A staff member's name. A specific dish. A waiting time. A room you walked into.

Fake reviews sound like this: "Very poor service. Not to be recommended. Absolutely disappointed." Or the opposite: "Great place, top quality, will definitely return!" – without a single concrete detail.

What few people realise: this generic style is not laziness but intention. The more specific a text, the easier it is to refute.

Pattern 4: The geographical anomaly

Google displays a rough location for each reviewer when they actively write reviews. A reviewer whose profile activity places them in Dortmund who suddenly gives your café in Freiburg a 1-star review – that simply does not add up.

Even more suspicious: multiple reviewers with an identical geo-signature who all contribute to the same profile within one week.

Pattern 5: The language pattern

Translation-influenced sentence structure, unnatural comma placement, phrasings no native speaker would write. "The quality of the services was truly very poor and the staff displayed no professionalism whatsoever." – that sounds like Google Translate from Romanian.

But beware: poor language alone disqualifies no one. The combination of language pattern + one-time reviewer + generic text is the real warning sign.

The 2 new patterns from 2025/2026 – that almost no one knows yet

Pattern 6: The suspiciously polished AI review

What we have been noticing over the past six months: a new generation of fake reviews sounds startlingly authentic. Correct grammar, structured composition, even a pseudo-concrete detail ("Waiting at the checkout took over 20 minutes"). What is missing: typos, colloquialisms, genuine emotion.

AI-generated texts have a characteristic smoothness. No sentence is too short. No sentence is too long. Everything is balanced – and that is precisely what is suspicious. Real people write either very briefly and emotionally or very lengthily and detail-obsessively. These middle-of-the-road texts with perfect structure and zero linguistic personality – those are worth double-checking.

Pattern 7: The coordinated mini-wave

This is the most sophisticated tactic we are currently seeing. Instead of five reviews in a single day, three to five reviews arrive spread across seven to fourteen days. Each account is two to four weeks old. Each has one or two harmless positive reviews at other locations – clearly intended to simulate legitimacy. The texts vary slightly but follow the same sentence structure.

This is negative SEO on a subscription basis. And it works because Google's automated system is optimised for individual reviews – not for this kind of distributed pattern.

We see this regularly in the data dashboard: sectors such as dentistry, law and hospitality are disproportionately affected.

Genuine vs. fake – the comparison table

Check in 60 seconds – the operator workflow

Four steps. No law degree required.

Step 1 – Open the reviewer profile (10 sec.): Click on the profile name. How many reviews does the account have in total? When was it created? Is there a photo?

Step 2 – Check timing (10 sec.): When did the review arrive? Are there other suspicious reviews in the same time window? Use the Sternehero review analysis tool to make time clusters visible automatically – instead of scrolling manually through the review history.

Step 3 – Analyse the text (20 sec.): Does the review contain a concrete detail that a real customer could know? Or is it so generic it could apply to any business?

Step 4 – Make a decision (20 sec.): Do three or more of the patterns above apply? Then reporting is appropriate. Does only one apply? Gather more context before acting.

What to do once you have found a fake review?

Sequence matters. Do not start the wrong way round.

1. Screenshot with timestamp. Preserve the current state – review text, date, reviewer profile URL. If the review is later removed, you will have no evidence. If it is not, you need the documentation for the report.

2. Document the profile. Number of the account's reviews, creation date, profile photo, location data. The more complete, the stronger the grounds for reporting.

3. Select the reporting reason. Google distinguishes between several violation categories: conflict of interest, spam, misrepresentation. Choosing the right category significantly increases the prospect of a successful review – and it is not always obvious.

4. Submit and track the report. Google reports have a tendency to vanish into thin air if you do not follow them up. How to submit a report correctly, step by step, is explained in our complete reporting guide 2026.

Sternehero bundles these four steps in a single dashboard: screenshot upload, profile documentation, reporting reason selection and real-time tracking of processing status. Some providers charge per review – which says nothing about whether the detection logic behind it is any good. Sternehero uses prepaid credits at a fixed price per submission that do not expire.

Use case: owner – Italian restaurant, Hamburg-Eppendorf

A restaurant owner from Hamburg-Eppendorf had a supplier dispute at the start of 2025 – an invoice was contested, and the business relationship ended badly. Three weeks later: three new 1-star reviews, all within four days. All accounts less than two months old, all without a profile photo, all with interchangeable texts containing not a single concrete detail about the restaurant.

The owner documented all three reviews with Sternehero, selected the appropriate reporting reason for each and submitted the reports. Result after 6 days: The platform removed two of the three reviews. The third remained – the text was generic but not clearly in breach of the guidelines. For that review, the owner used the reply generator to respond professionally without appearing defensive.

This is the honest reality: not every report leads to removal. To be frank – no tool can guarantee that every reported review will be removed. What a good tool can ensure is that your report is submitted completely, correctly categorised and with full traceability. That makes the difference.

Use case: agency – coordinated wave at an eye clinic

A local SEO agency from Frankfurt manages more than 30 medical facilities. In March 2025, the Sternehero multi-client dashboard registered suspicious review activity at an eye clinic in Wiesbaden: five new 1-star reviews in ten days, all from accounts with an identical structure – two to three positive reviews at other locations, each three to five weeks old, similar profile layout.

The agency instantly recognised the pattern as a coordinated mini-wave (Pattern 7), documented all five cases together and submitted the reports via the white-label dashboard. Three of the five reviews were removed within two weeks following the platform's own review. The clinic retained its 4.3-star overall rating – without the wave it would have dropped below 4.0, with direct consequences for local pack visibility.

For agencies protecting multiple clients simultaneously, exactly this kind of consolidated working is the decisive lever. More on this on the page for agencies.

When fake suspicion is NOT warranted

This belongs to intellectual honesty: not every bad review is a fake review. This mistake costs business owners energy, credibility and sometimes even customers – because they respond to legitimate criticism with accusations.

Four cases where fake suspicion is usually wrong:

• One-time reviewer with a concrete experience: An account with only one review, but a text that names your employee by name and mentions a specific date – that is very likely genuine.
• Negative review shortly after a public incident: If your business was discussed negatively in a neighbourhood forum or on social media, organic 1-star reviews sometimes follow from people who were never personally a customer – but whose review nonetheless does not violate Google's guidelines.
• Disgruntled former employee with a genuine experience: Painful, but real. As long as the text contains no false claims, the platform will not remove this review.
• Poor experience you have no internal record of: If you have no evidence that the person was never a customer, a report is very unlikely to succeed.

If you are unsure whether a review is reportable – the most common questions are answered in the Sternehero FAQ, including notes on typical processing times and the relevant guidelines.

Conclusion: recognising fake reviews is not luck – it is a system

An estimated one in seven Google reviews is not genuine. Anyone who ignores this loses local pack visibility, the trust of potential customers and ultimately revenue – without ever knowing why.

The 5+2 patterns in this piece are not an academic construct. They are what we see every day in reports: time clusters, one-time reviewers, AI-smooth texts, coordinated mini-waves. Anyone who knows these patterns recognises a fake in 60 seconds.

The problem: recognising is not enough on its own. The report must be complete, correctly categorised and traceable – otherwise it disappears into the support void. Sternehero bundles documentation, reporting reason selection, submission and real-time tracking in a single dashboard. German hosting, GDPR-compliant, prepaid credits with no expiry – for individual businesses as well as for agencies with 30+ clients.

Agencies and individual business owners from the DACH region already use Sternehero – and most start by reporting a single suspicious review.

Sternehero is a software tool and does not provide legal services within the meaning of the German Legal Services Act (RDG). Responsibility for the accuracy of the information provided lies with the user. The decision to remove or retain a review rests solely with the respective platform; removal cannot be guaranteed by anyone. The use cases, rates and effects mentioned are illustrative and do not constitute a promise of success.