16/02/2012 09:47 GMT | Updated 17/04/2012 06:12 BST

TripAdvisor's Fake Review Problem

An Irish hotel group has been caught asking its employees to post fake reviews on, says The Irish Times.

It wasn't the Carlton Group's brightest idea, and it's earned them nothing but opprobrium and bad publicity in Ireland. In July 2010, the chain sent an email to around 30 employees, revealing that a plan had been agreed whereby managers would nominate five people from each of the group's ten hotels to post the fake reviews.

The hotel says that the plan was never implemented and that no fake reviews went on TripAdvisor, but it's a barely credible response - especially as former and current employees have provided emails where the hotel chain called on them to "proactive" in managing TripAdvisor reviews; many staff members who expressed concern were told the policy was going ahead anyway. An further email in November asked managers to delete July's previous email. No further mention was made of the plan to mislead customers, and the policy was quietly shelved.

In any case, the damage is done: the internet has a long memory, and the story is likely to pop up any time a potential customer Googles the hotel group's name.

Fake online reviews are not particularly hard to spot. When they're glowing, full of marketing speak, and from a new commenter, this should raise a flag for website owners and users. On the other hand, some hotels and restaurants have complained about malicious and unfounded reviews, some of which come from rival companies.

Anyone can post a review on TripAdvisor, which uses a "crowdsourcing" model to present users with an array of opinions on hotels and restaurants. But some businesses have been known to recruit freelancers to write positive reviews.

Last month, the UK Advertising Standards Authority criticised TripAdvisor, warning the site "not to claim or imply that all the reviews that appeared on the website were from real travellers, or were honest, real or trusted." Last year, researchers at Cornell University have created an algorithm which claims to pick out the fake reviews from the real with about 90% accuracy.

Two years ago, following obnoxiously poor customer service, I posted a bad review of a restaurant on my food blog, The response followed what has become a wearily predictable pattern: first, some regular users of the site commented. Opinions were diverse, with many expressing shock at my experience. Then, a wearily predictable pattern of abuse began: one commenter, new to the site, called me a liar. More piped in: clearly, I'd made up the whole story for the sake of nothing more than a malicious libel (despite being an experienced journalist with no logical reason to lie and expose myself to a massive legal bill. But not every statement is worthy of rebuttal). Some glowing reviews appeared on, also from people who'd never commented before.

The suspicious comments came from fake email addresses. We later found a number of reviews on TripAdvisor praising this business, and another restaurant run by the same management, from the same reviewer that had used a fake email address on Cheapeats. Rival businesses in the area were slated by this reviewer.

Des Doyle, Online Editor of, Ireland's largest online dining guide, says that suspicious reviews are easy to spot. "We put a red flag beside any untrustworthy review," he explains. runs a "credibility score system", which gives more credibility to users who regularly contribute to the website. "A suspicious reviewer may have contributed, at most, one or two reviews, and often the second review is simply to cover their tracks. You can almost tell if there's an ulterior motive: they'll tear the restaurant apart, or praise it effusively."

Fake reviews are easy enough to spot on We once had a new commenter recommend a restaurant once. They used a fake name but put in their real email address; when we Googled the email, it turned out to be from the head chef of the restaurant in question. Another Dublin restaurant is notorious in the Irish food blogging community for posting positive reviews from the same IP address - a counter-productive approach that has widely damaged their credibility.

The problem isn't just confined to restaurants. Last year, a book on the menopause by author Jill Shaw Ruddock received 52 five-star reviews on Amazon all at once. All were from names that had never posted reviews before, and all were overflowing with glowing adjectives.

Doyle believes that the changing nature of public relations makes fake reviews inevitable. "Restaurants feel that they have to defend themselves. But aggregate review websites allow people to read a large number of reviews and get the average viewpoint."

A bad restaurant meal is just one evening, but a bad hotel or campsite can ruin an entire holiday. So how can a consumer protect themselves against bogus reviews? It's easy enough. First and foremost, read more reviews, and discard the completely positive or completely negative. Most genuine reviewers will find something wrong, however small: the restaurant was lovely but was a little too noisy, the hotel was great but reception was always very busy. Check out if the reviewer has written anything before. Watch out for positive reviews which name the wonderful owner.

Ultimately, the crowd-sourcing model works because, although there will always be lies and spin wherever there are words, the wisdom of crowds allows the truth to speak for itself.