New local search engine Grayboxx has spent over a year developing a methodology and collecting information that offers “implicit” recommendations or endorsements of local businesses. These are represented as scores, called “neighbor recommendations,” and offer an alternative to more familiar review sites such as Citysearch, Yahoo Local and Yelp. Grayboxx founder Bob Chandra says this approach allows the site to offer qualitative evaluations for many more business categories and locations than can traditional review sites that require people to actually write reviews.
Grayboxx is being somewhat vague about the process by which it generates the scores — the algorithm is called “PreferenceScoring” — citing that as its “secret sauce”:
Put simply, our recommendations come from other people in your neighborhood. When folks buy things from a store or make reservations at a restaurant, they are expressing their interest in local businesses. We add up and make sense out of these indications. Our PreferenceScoring™ approach provides recommendations in over 4,000 kinds of businesses.
The site is launching in Burlington, Vermont to illustrate that it can offer content where others cannot. Other cities will roll out over the next several weeks and months. You need a Burlington ZIP code to get in — try 05401.
Here’s my October, 2006 write-up of Grayboxx after a discussion with Chandra and here’s a more recent post after seeing the working demo this past Friday.
Related Topics: Search Engines: Maps & Local Search Engines
Since I work for a Vermont company (Maponics), I find it pretty cool that they are emphasizing Burlington Vermont. However, I’m astonished that they don’t offer neighborhood level searching. There are some unique aspects to Grayboxx’s business model, for sure, but with neighborhood boundaries and map data available, most local search companies tag their listings with neighborhoods so users can search using local names. Perhaps that’s in the works there?