Tuesday, 6 January 2015

TripAdvisor fined over 'false reviews'

Not our headline, the Financial Times's. On 22 December Italy's antitrust regulators fined Tripadvisor €500,000.

We would have more sympathy for TripAdvisor if, in its defence, it cited the precise methods and resources it devotes to combating review fraud. Instead their spokesman said it was a “force for good” and its systems were “extremely effective in protecting consumers from the small minority of people who try to cheat our system”.
We are getting a little tired of hearing this line trotted out time-and-again. What exactly are TripAdvisor's 'systems'? (TripAdvisor: please feel free to comment on this blog post, and 'No' we don't want to know the precise details of your 'secret algorithm', just sensible businesslike guidelines). 

For now we don't know what systems TripAdvisor deploys. And we suspect they don't either. When directly asked by Evan Davis on Radio 4, this was Steve Kaufer's response:

Evan: "Steve, how reliable are your reviews?"

Steve: "We have, you know, dozens of people, an entire department frankly, set up to focus on automatic detection of reviews that are suspicious. We have enough reviews so that there's this image of honesty in numbers. No-one is going to pay attention...to the review that is all terrible."

See the full transcript here.

Steve, we'd love to know more about that 'entire department' and their focus on 'automatic detection'. How do hoteliers contact that department, for instance?

When TripAdvisor was censured by the ASA in 2102 the ASA said:

"TripAdvisor said, because trust was such a key component in the site’s ongoing success, they had invested heavily in systems, processes and resources to identify and minimise fraudulent content: they provided details of their anti-fraud systems in confidence. 
We noted that reviewers were asked to agree to a declaration that their review was their genuine opinion of the hotel and that they had no personal or business affiliation with the hotel, or been offered an incentive to write a review for it. We also noted that reviewers were not asked to similarly confirm that they had no competitive interest in the place they were reviewing, or were posting a review on behalf of a competitor or other interested party, and we did not consider that agreeing to a declaration in itself would necessarily prevent non-genuine reviews from being posted on the site.

Notwithstanding that, we understood that reviews could be placed on the site without any form of verification, and that whilst TripAdvisor took steps to monitor and deal with suspicious activity, it was possible that non-genuine content would appear on the site undetected.

We noted that TripAdvisor allowed hoteliers a ‘right of reply’ to critical or negative reviews posted on the site and that they believed that users of the site had a healthy scepticism as a result of their experience of review sites more generally. However, we did not consider that consumers would necessarily be able to detect and separate non-genuine reviews from genuine content, particularly where a hotel or other establishment had not received many reviews, and nor did we consider that a hotelier’s response in itself would go far enough to alert consumers to, and moderate, non-genuine content."

The ASA was effectively saying that TripAdvisor's 'anti-fraud systems' were so ineffective as to make their claim to publish 'trusted reviews' unsupportable. We think it is high time TripAdvisor told its millions of users (not to mention its investors) what those systems are.

No comments:

Post a comment

HelpHound is all about feedback, so please feel free to comment here...