Yelp Founder Says "No Extortion — Just a Misunderstood Algorithm" 120
Early last year, a story in the East Bay Express reported that review site Yelp's ad sales force was using hardball tactics that amount to extortion — essentially, suggesting that negative reviews would remain prominent on the Yelp page for a particular restaurant or other business, unless the business bought advertising through Yelp, in which case Yelp could "do something" about the negative reviews. In a recent interview with the New York Times (the questions seem rather softball, but they do address this issue), Yelp co-founder Jeremy Stoppelman says it just isn't so, and blames unhappiness by business owners with the review site on the site's "automated and algorithmic" review-filtering system, which he describes as "counterintuitive." Stoppelman also says that Yelp's advertising salesmen have no connection to that filtering system, which doesn't quite answer the question of whether the salesmen claimed to be able to influence the reviews displayed, as some business owners allege. Updated 22:09 GMT by timothy: As reader AKMask points out below (now corrected above), that's the East Bay Express, rather than the East Bay Examiner.
maybe (Score:1, Insightful)
Sales people are liers. Maybe thats why.
A Misunderstanding? (Score:5, Insightful)
Advances in technology, such as our precious intertubes, will always bring advances in criminal activity. I hope no one hears them yelp as they go out of business when their next round of venture capital begging fails miserably. If they want to deal on business' reputations they'll need one of their own.
Re:Sham (Score:4, Insightful)
You might want to reconsider that. My wife and I use Yelp on the iPhone as one of our primary sources of information on how to get stuff we need when we're on the road. If you're not in Yelp, you're probably missing a lot of walk-in opportunities. If you're doing well on local advertising and reputation, great, but if you'd like to have more new customers, Yelp is actually a good way to make it happen.
The problem with services like Yelp is getting good information out of them--unfortunately, the main motivation people have for writing reviews on the web is that they are pissed off, and so that tends to work against any place that ever has a pissed off customer, which is pretty much every place. So we tend to look for patterns in the bad reviews to see what they tell us, and also patterns in the good reviews to make sure they're not fake. It's worked out pretty well for us.
Re:Sham (Score:5, Insightful)
Regardless, online reviews are pretty pointless. They're anonymous and easily gamed by anybody on the planet. They're about as reliable as bathroom stall graffiti.
Re:A Misunderstanding? (Score:5, Insightful)
Read the original stories about Yelp's tactics and practices. I can only hope they end up with both civil judgements and criminal convictions.
I almost believe him (Score:5, Insightful)
Re:A Misunderstanding? (Score:3, Insightful)
I expect a more nuanced response to his denial. One that might, perhaps, refute it. As it stands that comment could refer to any company except for the side-splitting double meaning of "yelp" and one word from the article title, suggesting that you have perhaps put a rash need to appear witty above the more significant issue of fraud versus innuendo and human pattern-matching.
Re:Sham (Score:5, Insightful)
Regardless, online reviews are pretty pointless. They're anonymous and easily gamed by anybody on the planet. They're about as reliable as bathroom stall graffiti.
Do you have a better option for someone just passing through who just wants a nice place to eat? The only alternative I'm aware of is the back of the phone book. With online reviews, I can check a few sites and mostly figure out whether it's the kind of place I'm going to like. Back of the phonebook is a total crapshoot.
Self serving story much? (Score:2, Insightful)
Re:Stoppelman doesn't get it (Score:5, Insightful)
The fundamental problem is that such algorithms are based on taking a small bit of information and extrapolating a result from it. It's fundamentally the reason why benchmarks are often both gamed and a very bad way of actually understanding the products being measured. The answer to the problem has consistently been shown to be not to withhold information about the way a benchmark is made, since invariably people will find a way to reverse engineer the algorithm and game it anyways; the answer is to further refine the benchmark to take more and more samples until the point that even if the benchmark is incomplete at measuring things, anything that tried to game the benchmark would still be very close to meeting what the benchmark is meant to represent.
AFAIK, that's primarily what Google has done with the page ranking algorithm. If Yelp is really worried about having a good review system, they should focus less on trying to hide how their algorithm works and more on improving their algorithm to guarantee it works. In the end, they'll remain ahead of any competitors so long as they can maintain a consistent lead on quality through such efforts. Any other mindset really is based in a belief that one has some sort of monopoly that can't be replaced. It's one of the reasons why complacent middle management, which is primarily a byproduct of large corporations, kills most corporate monopolies.
Re:Sham (Score:4, Insightful)