Not written in the stars

09. October 2018, Editorial

Whether a customer ends up buying a product or not is often down to the online reviews. But making them up is all too easy. How to spot a fake review and what really tells you whether a restaurant deserves its five-star rating – a quick guide.


a quick guide
Anleitung zum Erkennen
bald keine Falschrezensionen mehr
Fake Review
Transparency for the internet

“Brilliant coat. It keeps you warm and the colour matches the pictures 100%. The negatives: the fabric is pretty sensitive, so you have to take it to the dry cleaners if you get any stains or marks on it.”

Real or made up? Anyone who checks the product reviews on the internet these days is hard pressed to tell the difference between genuine and fake reviews. This applies in equal measure to fashion chains, restaurants, hotels and online stores run by high street drugstores. But it is precisely these user-generated reviews that are supposed to give customers a place to start. So is it worth making a purchase or not?

Nine out of ten customers read customer reviews first, taking them at face value

But it actually goes beyond that: almost half of consumers purchase a product after reading positive reviews and around 30 percent of shoppers are ready to part with more of their hard-earned cash if the reviews are really good. After all, fake reviews do not always have to be bad – positive reviews are just as likely to be fabricated. According to a study conducted by the Harvard Business School, around a fifth of all the reviews published on Yelp – both good and bad – are fake. The people behind the fake reviews are becoming increasingly clever, making a name for themselves with authentic reviews before going on to write up their fakes. But it is still possible to spot a fake.

Warning, fake!

Jeff Sakasegawa works for Sift Science, a start-up that specialises in identifying and tackling online fakery. His job title of Trust and Safety Architect says it all, really. When it comes to fake reviews, he knows his stuff. He reveals that there are six characteristics that give away a fake

One: they link to external sites.


Two: their authors write multiple reviews one after the other.


Three: they spice them up with flowery language – generally using more verbs than nouns.


Four: they tend to have extreme opinions, either loving or hating something. They virtually never weigh up the relative pros and cons.


Five and six: technical terms crop up a lot in fake reviews as well as specific words such as “recommend”, which is a real favourite.

Be it a private individual or a small company, anyone able to interpret the signs correctly can spot fake reviews. But sites like Fakespot and Trustwerty, which use algorithms to check reviews, offer addition certainty. Large companies such as Amazon take a similar approach. In some cases they use moderators who look at the authenticity of a contribution, using location and user history for clues. Other platforms such as AirBnB only allow verified users to leave feedback. And in spite of all this, people still manage to slip through the net.

Prison and fine for faking reviews

Only a short time ago, an Italian man was sentenced to nine months in prison and an 8,000 euro fine for faking reviews. Before being rumbled, he sold more than 1,000 fake reviews to hotels over a period of several years using a fake identity. Sometimes fake reviews are about nothing more than increasing bookings or sales figures, but sometimes they are a way to make competitors look bad. So it is hardly surprising that an entire industry has sprung up around fake reviews, with the sole aim of preventing them.

Transparency for the internet

Researchers at Aalto University in Finland worked on a type of algorithm that could author particularly authentic-sounding fake reviews – they did this so they could program a supplementary AI application at the same time, which was able to identify these clever fakes with a high degree of reliability. Another method, one that programmers Petar Slovic and Ivan Ciric back, is related to blockchain. The technology behind the Bitcoin boom also makes it possible to create verifiable digital identities, with data sets covering date of birth and address, as well as info on locations and purchasing history. Because once in blocks, data sets can no longer be changed without anyone noticing, suddenly making it possible to simply check whether the author of a review has actually purchased a certain product or eaten in a certain restaurant.


All Public Art is using blockchain to confirm the provenance of art works. Slovic and Ciric want to open up their Review.Network for online reviews too. Verified users write real reviews, earning themselves REW tokens, a currency similar to Bitcoin. More than 100,000 people have already signed up in advance. To help make the internet what it was to start with: a place where people go for help and find honest opinions.


And before we forget: the review at the start ticks all the boxes for the genuine article. But it’s still a fake. The coat does not exist. Well, at least that’s what we think.

Similar Articles

Trends in 2021

What’s on its way, what’s here to stay?


What a load of rubbish

How sustainable packaging can work for online retail.


Take a deep breath

...and let all thoughts settle. Also companies are turning to mindfulness techniques.


Biene auf Blüte

It’s a gift!

Win customers with free samples.