Dottourism.com

User generated content and Hotel review sites

15 November 2006 | Dominic Sawyer

The Sunday Times has entered the ongoing debate concerning false reviews found on hotel review websites and follows an article published earlier this year in the New York Times. In recent years, travellers have switched their attention from printed guides compiled by professional inspectors to the numerous user reviews that can be found on a burgeoning number of hotel review sites. It is no surprise that investigations found hotel owners inflating their ratings, with the most blatant giving themselves glowing reviews. The Sunday Times article did not point out that the problem affects all types of rating/review sites - a recent example being speculation that a London PR firm has rigged the YouTube system to get listed on the ‘Most Subscribed to’ page.

The success of hospitality review sites have created their own problems. Trip Advisor has over five million reviews and moderating such a high number is no easy task. Trip Advisor claims that all the reviews are read but how can they determine if a review is fake or not? The Sunday Times reporters submitted extremely favourable reviews to hotels that had previously been scored consistently poorly. These reviews were published, and in one instance a London hotel review score was increased from 1.2 to 4.2 with the addition of four favourable reviews.

As the number of reviews continues to increase and hotels discover and adopt tactics to improve their ranking, the hotel review sites will need to adopt strategies to protect the validity of their, albeit user-generated content. I’m not aware of current strategies in place and it is a fine balance, but I’m sure reviews that are submitted that contradict previous ones markedly could be flagged for further investigation. The New York Times article writes that Trip Advisor were involved in the creation of a fraud detection algorithm. Perhaps each newly-placed rating could be cross checked against the number and score of previous ratings for a particular hotel. Other factors could also be taken into account, such as the reviewer’s history of contributions.

For example, if a user appears to negatively review hotels in a localised area then this should be easier to detect than a user who creates a separate identity for each review. Perhaps, Trip Advisor could distinguish certain reviewers that post regularly, in a similar vein that Amazon highlights ‘top’ reviewers that could aid users to attribute weight to reviews. According to a discussion between Travolution and The Sunday Times, the latter believes Trip Advisor’s ‘processes to catch persistent hoax reviews does work - it is the single bogus reviews that are ones more difficult to catch.’

As these sites continue to tackle the issue of bogus reviews and any erosion of user trust, you’d hope reviews and ratings play one part in the user’s decision-making process and that where ratings are numerous, it is more likely any scam reviews would be drowned out in a sea of common consensus from authentic reviewers. The temptation for hoteliers to work the system will not go away but on the flipside there is a risk of having reviews removed, an increase in disappointed and duped customers and of course the moral case for misleading site visitors.

There are plenty of highly charged comments, and some amusing, at the bottom of both The Sunday Times article and the story picked up by the Scotsman!