Startup Grayboxx is out of beta, launching a nationwide local search service Jan. 16 that recommends and ranks businesses through an algorithm rather than just through user-generated reviews, which is the modus operandi for competing services.
Local search services such as Yahoo Local, IAC's CitySearch or Yelp generate a list of businesses and let users write reviews to rank them. While these user-generated reviews provide useful information, there is a gap between the number of available reviews and the number of local businesses in the United States.
To fill this void, Grayboxx uses a technology system called PreferenceScore to tell users which businesses are most popular with their neighbors, founder Bob Chandra told eWEEK.
When people buy goods or make reservations at a restaurant, they indicate their interest in local businesses via a name, phone number, street address or city. These references, also known as implicit endorsements, are references people make to businesses in user data sources such as an e-mail calendar or cell phone address book.
Grayboxx gets this data from a variety of third-party vendors that access it, but it does not acquire, track or keep any personally identifiable information about its users, relying instead on the frequency with which the business information is referenced.
PreferenceScore tallies this data and uses it in congress with user-contributed reviews, culled from third parties or created directly on Grayboxx. Instead of relying solely on results from one of the traditional local services that offer a sprinkling of user-generated reviews, a search on Grayboxx yields dozens of ranked businesses, each of which often has anywhere from five to 100 neighbor recommendations.
With this approach, grayboxx.com can offer 185 million business rankings and recommendations in more than 4,000 Yellow Pages categories across the country.
Chandra said he came up with the idea for Grayboxx after becoming frustrated looking for a cobbler in Berkeley , Calif. , to repair a boot. Yahoo Local returned more than 500 results in no particular order, making it difficult for him to make a decision.
He realized with Yahoo Local that there are plenty of reviews for certain categories such as restaurants, but not so much about plumbers, patio furniture stores and the thousands of other businesses people want recommendations for but don't take time to write reviews about.
That disappointing discovery fueled PreferenceScore, which Chandra expects will do for local search what Google's PageRank site reputation ranking system does for top-line search.
Google changes its search algorithm regularly, causing tectonic shifts in PageRank scores, driving search engine marketers batty in the process. Grayboxx will also update PreferenceScore often to keep businesses from "gaming" the system to improve their scores.
At launch, Grayboxx has no advertising on its search site, perhaps a welcome breath of fresh air from the search engines of Google, Yahoo, Microsoft and Ask.com.
However, Chandra said Grayboxx will incorporate paid advertising over time to generate revenue. These could be specially denoted results on a sidebar or letting businesses pay for more complete profiles.
Grayboxx is going national from beta at a time when local search is expected to enjoy a boom in 2008, according to research released from The Kelsey Group on Jan. 10.
Merchants will be able to take part in the consumer review process so they can have a voice in the feedback about their businesses, as user-generated reviews increasingly come to the fore.
Also on tap are increases in local search inventory, more search through specialized devices and greater interest in local automotive and real estate markets, the research firm said.