Joey deVilla has a great post about his experience trying to get a recommendation for a moving company. It sounds simple enough, ask you’re blog readers to respond with recommendations for local services. In reality its hard to get reliable information, in Joey’s case the recommendations were tainted by an apparent fake recommendation and legal/intimidation against apparently true recommendation. Not a great representation for user generated recommendations…

We looked at this when I was with Trader Media, and again with Refresh on Freshrevie.ws. The hard part with user generated recommendation is to evaluate the quality of the recommendation. Its a similar situation with Ebay, and their approach is a user rating system based on past transactions.  Its hard to do a rating a recommendation engine since the past reviews have less baring and there is no money being exchanged to qualify the review.

I think the solution is to have some sort of context around the person writing the review. Facebook is a great environment for this, since you generally know your friends and your friend network helps define and validate you. Gigpark is doing a great job of getting user generated recommendations, by focusing on the user relationship first. They’ve done a good job of connecting users in Facebook and on their own website. The challenge they might face is that people have a limited friend network and they don’t have a complete businesses listings.

The big business directories, like Yellow Pages have very complete listings but haven’t been willing to expose their advertisers to user generated recommendations. Yellow Pages has just recently launched user reviews inside the Yellow Pages Facebook application. This should reduce their exposure to fake reviews since Facebook spends alot of time tracking down and deleting fake accounts.