There's been a lot of disturbing news about Yelp recently. Yelp, if you're not familiar with it, is a website which allows people to review stores, restaurants and other companies. The problem is that some store owners post bogus positive reviews and, so some allege, post negative reviews on their competitors.
If you’re interested in game theory the Yelp example is bound to be fascinated to see how this scenario plays out. Adding to the problem is that Yelp is removing bad comments should a store advertise on Yelp – or so some say. As someone who is affiliated with stores on Yelp I have to say that I haven’t experienced that.
This posses an interesting IA problem: How should Yelp, and by extension other review oriented sites, deal with this problem? You can never stop people from creating multiple accounts but you can give added weight for community participation and then give added weight to accounts which give more and more evidence of legitimacy.
There are sites which I’ve reviewed that have dozens of reviews. Each of these glowing reviews are from accounts with only one review. This makes the reviews more than a little suspicious. One way to stop the abuse is for sites to have no more than a few reviews from single review accounts. The second way was alluded to earlier – give added weight to accounts which are considered more legitimate by counting the number of reviews, friending other users and through general community involvement (at its simplest this can be done by noting the times the account logs in, page views and time spent in the community.)