[WSJ] Yelp Reviews Brew a Fight Over Free Speech vs. Fairness

On the flip side though isn't there a whole industry now to posting positive reviews on review sites and other internet forums to pump up a business?

If we are cracking down on dishonest negative reviews shouldn't a similar effort be made to crack down on dishonest positive reviews?

 

Right now the businesses are the ones paying for it substantially. This solution is far cheaper and gives them the chance to protect their sales. You tell me what's more expensive, losing 30% of your sales or spending a few hours to get rid of fake reviews or a few dollars to make genuinely unhappy customers happy?

 

If you look at the reviews on Amazon, you'll notice they put in "verified amazon purchase" which shows whether the user account actually purchased the item they are reviewing, however Amazon still allows people who haven't purchased to review an item.

Another way those selling on Amazon get around this is that they just pay the reviewer to purchase the product, leave the review and then return it or just let them keep it. Positive reviews are valuable and it's a cost of doing business so paying someone to buy the product and leave a positive review can be written off as a marketing expense.

 

It's not stupid to be duped by the unscrupulous posing as sincere. People are generally trusting. Fake sincerity appears to us to be truth until the real truth is absolutely unavoidable.



And if it works for marketing crap, it will work for marketing crap politicians too.

 

Oh I agree that your solution is better than what we have now. But I think there are better ways by incorporating ID verification. It creates the potential of a libel lawsuit as an incentive to not post fake reviews - it helps stop the process before a business has to waste time on it.

 

I can see it, depending on what impact it has on the little stars and how commoditized the service is. I think there is an outsized psychological impact from small changes in the aggregate score. My wife's business sees it in click-thru data, with 4 stars getting a lot more than 3.5 stars. You can also see it in BBS thread reviews -- 3 stars, that's a stupid pissing contest between two factions; 4 stars, solid thread I should read. And if its a service with many competitors all doing the same thing, the number of stars might be the only data-point by which a customer picks a counterparty.

So, I imagine some Yelpers probably search for carpet cleaning and scan down the list from best to worst, and there's 50 places in geographic proximity. Knock a half-star off an outfit and he could fall 10 places down the stack and not get a phone call.

Or you want to call this place, but you'll check it out on Yelp real quick. Seeing the 3 stars upfront, your opinion is anchored to the idea that service is spotty. You go on to read, and some of the bad reviews are poorly argued, so you mentally discount them, but others are legit. Maybe the math without the poor bad reviews would have calculated 4 stars, but you've had an anchoring effect, so you mentally upgrade to only 3.5. Better than you thought, but not good enough to do business with, so you're on to the next guy.

 

That could a solution. I guess the question is whether Yelp would be willing to do it and risk depressing their response rate of legitimate reviews as people may not be willing to take that added step.

I think an optional verification process which user can opt-in to in exchange for some sort of incentive would be a good idea, with those reviews being less likely to be open to challenge.

 

There are competing marketplaces in which customers can do their shopping. It used to be the Yellow Pages, where companies would pick names that start AAA so they can be at the top of the listing. Google Maps is another one, which I often use to find a commoditized service provider, where you can pick by geography (incidentally, they also have reviews, but no complicated algorithm to say what's valid or invalid). If Yelp didn't exist to drive customers to this guy's carpet cleaning company, those customers would use some other marketplace, and one in which he might compare more favorably or arguably fairly. Maybe you could say Yelp provides the fullest information about businesses of the available marketplaces. That's going to help some businesses and hurt others, depending on how good service is and how well they can game the system.

 

I think the answer is an algorithmic approach to evaluating the credence of a rater, rather than any outright outing over ID.

ID is subjective, and I'm still not sure "unmasking" people does so much with so many disparate online identities, and few touch points between the physical world and the digital world. Nevertheless, the mere threat of legal action would stifle dissent, and negative views, not something anybody wants.

So just like Uber rates the drivers AND the passengers, you can have a similar approach. Evaluate the credibility of a rater by how often they have posted, join date, number of positive vs negative reviews (if they tilt towards 100% in either direction, that's a minus), and distance to the business in question. Weight the ratings of a business accordingly, with high importance placed on the ratings of quality raters. This algorithm would also easily find "fake" raters, and force some special form of authentication for those particular users for a high likelihood of violating the terms of service---you're more likely to punish the right people, instead of punishing the entire ecosystem for a few users.

This does sidestep the issue of libel, but at the same time it's a fix that placates both sides. I lean towards free speech, especially on the Internet, and Yelp does have a lot of great raters---so it's something that appeals to me, rather than this whole libel deal.

 

This would be a great idea - you could also promote those people's reviews more highly or list the other peoples' as "unverified" or something. It would be a good step that's purely voluntary. Sort of like twitter's "Verified" account status.

 

Yelp being more of an intermediary, like a for-profit BBB, where users have to verify their identity and purchase to the yelp but not face external pressures of getting sued for defamation.

I don't think it would even take that much more people to verify. Just an automated process (upload receipt) that kicks in when people call bull**** on a review.

/yes it's like Sweet Lou's idea and with more hooks.