Wanted: An iPhonebook
- TAGS:search, white pages
- IT TOPICS:Business Intelligence, Internet, Macintosh & Apple, Personal Technology
There are times when a plain old paper telephone book works better than an Internet white pages lookup. Anyone who's ever been frustrated by an unsuccessful white pages lookup knows exactly what I mean.
If you don't know the exact spelling of the person's name, if you're looking for Joe Smith but his telephone is listed under spouse Jane Jones, or if you're trying to get in touch with a "Bill" who lives in the Maple Ave. area of either Cambridge or Boston, you know how frustrating it can be to get a white pages search engine to spit out the right listing. You can do multiple searches and still come up empty. Sometimes it's easier to let your fingers do the walking.
The dead tree browser
Then there's the good old fashioned telephone book. It's only organized one way: By last name. But for some tasks, the browsing is much easier. I'm an analog kind of guy. Rather than sparring with a search engine, using logically rigid searches to return discrete bits of search results in a back-and-forth dance, there are times when I simply want to roll my finger down the page, scrolling my way down until I find the right listing. For fuzzy lookups, white pages Web sites should offer a browsing option. And what better browsing model could there be for that user interface than the iPod?
After one frustrating search experience I decided to speak with Alex Algard about this idea. Algard is CEO of WhitePages.com, which last year made $61 million by offering free listing lookups to the public. Whitepages.com claims to have over 180 million listings covering 80% of the adult populaton in the U.S.
Finding the needle
Algard acknowledges the need to do discovery-type searches, but says most people still prefer a more targeted approach that presents just the correct listing, when possible. "You’re trying to find a needle in a haystack and you don’t care what the needle is. Our [customers] are on a mission to find the people they’re looking for. It's critical that we float up to the top of the results set the person they’re looking for."
Actually, I do care what the needle is. I just can't find it easily because the data I have - what I know about the subject - doesn't fit neatly into the limited search fields I'm offered.
Algard is quick to point out the limitations of my dead-tree white pages search mechanism. It's usually out of date before it's printed. It's only updated once a year. It only lists landline telephone subscribers. It offers only names, street addresses and landline numbers. Each book covers only a local area.
In contrast, Whitepages.com has a national scope, allows searches by telephone number, name and address, and uses data from multiple sources to add contextual information to the listing results, such as the person's age or professional work number. Algard is also allowing people to add other data, such as their cell phone numbers, to their listings.
Whitepages does use some fuzzy logic to try to guess the right result. It will do "automatic wildcarding" to match nicknames to formal names and match you up to the right city if you misspell "Philadelphia," for example. Or if you can't find "Robert Mitchell" on Main Street but you know I'm a reporter you might find my name associated with Computerworld - that is, if my publicly listed business telephone line showed up in the online white pages databases, which it doesn't.
In the real world, online searches can quickly lead to frustrating dead ends. If you search for Robert Mitchell at my Gilsum, NH address you won't find me. The phone is in my wife's name. You also won't find the name of my neighbor, Joe, for the same reason. The reverse is true for my primary residence in Keene. That phone is in my name, not my wife's.
A better browser
Whitepages.com has a nice feature that allows you to generate a listing of all "neighbors" for a given listing. But not every home shows up in that list. Until recently, if you knew my number at home but want to use reverse lookup to double check that you have the correct number you'd come away confused. Instead you would be fed the name and address of the previous owner of that number. It took more than six months for the correct information to be updated to Verizon's SuperPages and propagated out to other white pages search databases.
Online white pages searches fail because they still rely heavily on that same information provided by the carriers. A white pages search is based on incomplete, outdated, and often incorrect information from multiple databases, which are assembled on the fly like a giant jigsaw puzzle. All too often, key pieces are missing.
White pages searches don't always work. So why not enhance the ability to browse for those listings that are more of a discovery process than an exact matching science? I'd like to see an iPod-like interface that lets you scroll through a filtered set of listings. In this way I could ask for all of the McLoughlins and McLachlans in Boston. Or all of the listings located on Arch Street in Keene, NH.
While PCs lack the iPod's dial interface that allows quick scrolling through groups of songs, a roller button on a mouse could be used in the same way. If you scroll faster it would let you skip ahead from the "A"s to "B"s or "C"s. In this way you could quickly cruise a street, neighorbood or town for all listings, or review all of the McLoughlins and McLachlans in a given area.
You can do this today with the whitepages.com neighbors search - sort of. You must start with an address and then ask to see neighbors on that street. You can include surrounding streets as well. But results are presented in discrete chunks that require clicking "Next" several times to browse the entire result set.
Unfortunately, those approximation searches want you to start with a fixed address or listing name first, and they don't always work. A search on "Island Street, Keene, NH" returns no neighbor results, but a search for neighbors to "115 Island Street" comes up with 30 possible listings.
Guessing an address can put a monkey wrench into the works. A search on "100 West Street, Keene, NH", for example, pulls up only six correct listings. Twelve others are on the wrong street - or in the wrong town. Most are missing entirely from the result set.
Coming soon?
Algard agrees that there's plenty of room for innovation when it comes to this type of searching online. The Google model has had great success, he says, but adds: "There's an opportunity to go above and beyond that model, which we haven't done yet."




