Looking for a restaurant tonight Start-up Bizzy says you cannot trust your friends, who have different tastes than you. Nor Google or Yelp, which are far too generic. Bizzy's better recommendation engine, CEO Gadi Shamia says, does a Netflix on your tastes, You tell it what you like, and it finds other places you're also likely to appreciate based on hidden signals in your data.

For example, if you like loud restaurants over quiet ones, or if your top criteria for a dining establishment is the attractiveness of the waitstaff, the Bizzy engine will return results that work for you, even if you're never specified these ideas. You just tell Bizzy the places you do like, and its algorithms do the rest.

Bizzy doesn't actually know which restaurants are loud. It just knows what you like and what other people like who like the same things you do. The more people create preference connections between commercial establishments, the better the results should be. And although Bizzy relies on its users rating places to generate good results and it lets users share their finds, it's not a social network. You don't have to tell it who your friends are or give it your social graph. Rather, says VP Ryan Kuder, "There's a huge untapped potential in figuring out what the commercial graph looks like."

Bizzy asks 20 questions to get to know you.

Assuming the recommendations engine works as advertised, the Bizzy team would have to be exceptionally dunderheaded not to make a nice profit from referring users to dining establishments they like. This will be the company's main business. The site will also be able to recommend shopping and entertainment venues, but Kuder realistically notes, "You make eating decisions three times a day."

They're also considering adding a deal or coupon service for businesses, which could work (this was the original idea for the product, before they zeroed in on the recommendation system). Kuder says that Bizzy will never juggle recommendation order based on commercial deals. "Our promise to users is that recommendations will be pure," he says, but neither will the company shy away from the fact that it's a commercial platform.

Bizzy may link out to Yelp, as Bizzy is not, strictly speaking, a competitive reviews site. Yelp, however, is a competitor for mindshare, in that it remains a go-to source for people seeking dining recommendations, and because it has the raw data necessary to create a Bizzy-like recommendation system.

The site doesn&39't yet have enough data to provide solid advice.

The necessary mobile apps for Bizzy should be along shortly--iPhone first, followed by Android.

The Bizzy business model is smart and I believe it's conceptually robust. It should be tolerant of a less-than-perfect implementation, although user uptake may be slow due to the strong lineup of existing restaurant-finding systems that Bizzy hopes to displace or supplement.

The only thing I don't like about the Bizzy product itself is that it's a potential date killer, as it will tell you what you will like, not what would be best for you and your dining companion. However, the team is aware of this, and has on its development roadmap a function that will let multiple Bizzy users find the intersections of their personal recommendations to locate the venues that are most likely to make everyone happy at once.