Are most of the businesses in your address books ones that you would endorse? How about the mentions in your calendar of restaurants that you are meeting colleagues at? How about the businesses in your photo software that you have bothered to tag? Or the ones you mention in emails, various Web sites, and other sources?
Generally, the answer is “Yes. I like them. That’s why I use them.” And that’s the premise behind Grayboxx, a new, Bay Area-based company that is “four to eight weeks” from being funded, and hopes to mine user data from a variety of sources to come up with “most popular” ratings (although in this case, “most popular” is translated as “most mentioned.”)
Company founder Bob Chandra thinks the system is going to prove vastly superior to the ratings and review sites, like CitySearch, Judy’s Book, InsiderPages, Yelp and Kudzu. Unlike the others, he claims, it achieves a critical mass of reviews, without changing the essence of user meaning. Aside from popular categories like restaurants, most businesses on other ratings and review sites have just one or two reviews, and most don’t have any, he says.
True, Google Base and MSN Live Expo have gotten closer to crticial mass by aggregating reviews from several of the sites. That helps. But they don’t always have enough, either.
So Grayboxx sounds like a good starting point. But by no means is it a perfect solution. Speaking personally, my Palm tends to only contain service providers, such as painters, electricians and termite killers. The only restaurants that I tend to list are a few upscale restaurants that require reservations – usually in cities where I travel on business, rather than where I live.
Although I have reviewed them on the ratings sites, I certainly have no need to list the Mexican restaurant that I go to every Friday night in my address book — Fidel’s. I also don’t list Linda’s Homemade Yogurt down the street, where I finish my pigout. If you want to follow in my footsteps, gastronomy-wise, you won’t get satisfaction from reading my Palm.
For that matter, I don’t actually tag the names of businesses in my PhotoShop, either. The best you will get out of me is a photo outside “The Plaza Hotel” in New York, where Redford and Streisand once stood. It would be kind of an aberration for me to tag, say, a Marriott Courthouse in Overland (even though I am a proud member of Marriott Rewards).
But you got to start somewhere, right? And the Grayboxx solution is certainly inventive – assuming it doesn’t alienate people by taking their address book and email info — even in aggregate. The company also gets brownie points with me by hiring a marketing exec that has had some exposure to local sales – former SBC Senior Director of National Advertising Doug Threet. That’s kind of unusual for a Bay Area startup these days, hats off to them. (Innocent question: why would a VC invest millions of dollars in a local concept company that hasn’t hired an executive or consultant without any kind of local experience, or even exposure to a local media company?)
With this in mind, I had a phone conversation with founder Bob Chandra. Chandra says I am probably atypical in my limited mentions of local businesses. In Grayboxx’s early testing, most address books alone yield three-to-five businesses apiece, yielding aggregated review counts of 25-100 for businesses in the Bay Area, where the company has been doing some prototypes. On Yelp, he says, you might see just one or two.
Chandra says his ultimate ambition is to collect a base of 50 million reviews. He also plans to go vertical, with such products as a “home contractors” page and a “business to business” page – both areas of particular strength. His goal is do an “effective job” of providing ranked results in 3,000 categories; 600 of which he hopes “to provide at least 200 recommendations.”
To reach his numbers, Chandra is currently leaning heavily on a national data collector (whose identity he won’t publicly reveal, but if what he told me checks out, it is an interesting one). Looking forward, he envisions a number of additional sources for data, including the aforementioned photos, but also such sources as free directory assistance queries. One imagines there is real potential with that as well.
So what do I think? The company’s solution is kind of left field. And it isn’t going to be nuanced enough to plan a date. Based on sheer number of mentions, you’d probably end up at Applebee’s.
But Chandra disagrees. To paraphrase an old record cover, to him, basically, 50,000,000 Elvis fans can’t be wrong – especially when they are weighted with some unique Grayboxx algorithms for negative comments etc. “Please see our correlations with quality (user rankings and critics picks),” he says. “I invite you to conduct searches against Yelp or Yahoo.”
What else do I think? Assuming that Grayboxx gets the right partners, and its concept goes over, I think it may well get over the serious hump of “critical mass.” And then it could be perfectly positioned to build its own review database, since it will be operating from something. I also think its chances are somewhat stronger in vertical categories, where it can go up against approaches like LinkedIn’s new Yellow Pages.
I just hope it doesn’t take its algorithms too seriously. This isn’t about computers. Ultimately, social networks really are about individual taste. There’s nothing in Grayboxx that helps me with that. Yet.