Driving the right people to the right places is a hot topic at the TechCrunch 50 conference. Mobile technologies has improved to facilitate the advancement of location based product recommendations. Companies like GoodRecs are allowing users to explicitly rate restaurants, books, and nightlife by giving a thumbs up-thumbs down type rating indicator.

Survey Bias. I can see how location based recommendations would be extremely useful, but explicit recommendations introduces survey bias and the only people praising or condemning the products are the loud users. Studies often show that users that dislike or disapprove are more likely to rate a product then those enjoying the product.

Contextual Differences. Additionally, its often unclear what context the person is in when the recommendation is given. For example, I may rate a product poorly because its not what I was expecting. However, it could be great for someone with different expectations. In order to discover these expectations, we’d need to know what is the higher level intent of the user. Accurate recommendations must take into account the context of the user voting for the product. While Baynote is able to uncover a user’s context on a website, this information is difficult to extract offline.