I met a family member (like 2nd cousin or further away) at some wedding a few years ago. I only talked to him once, but he was very clear about this situation. He owned a hotel, and was trying to track down the writers of bad reviews on Yelp. Primarily, as the owner of the hotel, it was his job to find out what's wrong about the hotel and improve it.
But he started to investigate some of the 1-star reviews, and came to the conclusion that these people didn't exist. He dug through all the receipts, he tried to track down and correlate the time with when these reviews popped up and tried to track down any issue. His ultimate conclusion was that a number of the 1-star reviews on his hotel were simply fake.
The discussion continued to talk about tort law, and how he's unable to even get a person to sue. He can't sue Yelp, because Yelp wasn't the author of those posts. So he was basically helpless to defend against 1-star reviews. In any case, if people can fake a 5-star review, they can also fake a 1-star review. And based on what I discussed with this man years ago, it seems like it really is happening right now.
Its probably easier for a Hotel-chain to leave bad reviews on their competitors. There are only so many hotels in a given area, and Yelp easily allows you to list all of the competitors in your region. Its probably less of a thing on Amazon, but I don't see any reason why it wouldn't exist there either.
> The discussion continued to talk about tort law, and how he's unable to even get a person to sue. He can't sue Yelp, because Yelp wasn't the author of those posts.
Yelp doesn't need to be a party to the lawsuit for a subpoena to be issued to them. You sue the poster as John Doe, send a subpoena to Yelp for the poster's account information and IP address, send a subpoena to the poster's ISP for the account information of the user assigned that IP address at that time, and keep following that trail. Eventually you might find some person who was paid to post the fake reviews, and you can use the legal system to find out who paid them. (In fact, they'll probably cooperate with your investigation in exchange for dropping them as a party.) There's no guarantee of success -- you might find the trail ends up with a no-logging VPN, an open Wi-Fi hotspot, or some other multi-user shared environment where the activity can't be linked to a specific person -- but it is possible to try.
A heuristic that I've used quite a lot recently is to look at three-star reviews and see if the reviewer's gripes are relevant to me. Works quite well, but I guess it is only a matter of time and we'll soon start seeing fake three-star reviews.
And eventually when people start to favor two and four star reviews those will start being gamed as well, and then, just like in a math problem where all the variables magically cancel out, all the fake reviews will exist in perfect equilibrium and average review scores will reflect only the real reviews! Genius!
How would you even implement a verified purchase? For a hotel maybe you could forward an itinerary (which can be faked) but there’s probably no receipt for a random taco stand that only accepts cash. Granted the latter does not care as much as the hotel but whatever they do would need to cover everybody.
After thinking about this problem for... umm... 30 seconds...
Allow users to write reviews, and then send a message to the hotel owner saying "X has posted a review of your business. Do you accept that X was a guest on X date?"
The hotel owner is forced to verify the stay before he gets to see the review.
Of course he'll know all about the outliers who had a bad experience and complained at the hotel and might want to lie about the guest staying, but such guests are usually very motivated and willing to prove their stay with receipts, photos, etc.
With a receipt system to authenticate real customer and an opt-in inscription to the service by restaurants/hotels, it could make it.
It could be linked to an order/invoice number the owner could check or enter in the system when the client check in or check out. The client would then be required to enter that number to authenticate he's a real consumer.
That's not how Yelp works, though. It's not opt-in or even opt-out. I'd guess most businesses reviewed on Yelp have no relationship with Yelp.
And ones with primarily bad reviews (which may not be their fault: cheap motels in particular seem to attract bad reviews from people who expect too much for the price) could just refuse to acknowledge any reviews.
Plus I think users are attracted to the illusion that the business doesn't know who they are. People would be less inclined to post reviews if they knew they'd be emailed directly to the business owner for sign off.
How would Yelp know if you made a purchase? You’re not buying anything through them. Lack of a verified purchase describes pretty much every review long before there was a Web although there were of course gatekeepers like newspaper editors.
But he started to investigate some of the 1-star reviews, and came to the conclusion that these people didn't exist. He dug through all the receipts, he tried to track down and correlate the time with when these reviews popped up and tried to track down any issue. His ultimate conclusion was that a number of the 1-star reviews on his hotel were simply fake.
The discussion continued to talk about tort law, and how he's unable to even get a person to sue. He can't sue Yelp, because Yelp wasn't the author of those posts. So he was basically helpless to defend against 1-star reviews. In any case, if people can fake a 5-star review, they can also fake a 1-star review. And based on what I discussed with this man years ago, it seems like it really is happening right now.
Its probably easier for a Hotel-chain to leave bad reviews on their competitors. There are only so many hotels in a given area, and Yelp easily allows you to list all of the competitors in your region. Its probably less of a thing on Amazon, but I don't see any reason why it wouldn't exist there either.