People Search Engines and the tail

Web Worker Daily has a review of some People Search Engines, where (as usual) “People Search Engines” consist of the new hotness only – engines that have just appeared, get all of their data from web crawling, and (in many cases) meet the MDRP (minimum daily requirement of pastel) plus other buzzwords necessary for Techcrunch to love on ’em.

In this review, the author searches for three folks with medium-to-very-large levels of net presence, and judges that while none of them are all that useful, Wink is the “best.” I think Wink’s a solid product as well, but this is a self-fulfilling prophecy: it turns out that if you search for three people who have LinkedIn profiles and Twitter accounts, an engine that’s expert at finding people who have LinkedIn profiles and Twitter accounts (i.e. social profiles with real names) will do great. (The author’s impressed with the remnant links on the page – those are Google results, coming from one of Google’s embeddable search products.)

When it comes to people without any sort of self-defined net presence, these engines fall flat. Search Wink for all Ruthfields, for example, and you get a bunch of LinkedIn profiles (and someone who’s using our name for some reason… I can’t access her acct on Bebo so I don’t know for sure). People in the family who didn’t go through the trouble of creating a LinkedIn profile? They don’t exist.

When the engines do access web-crawled data that isn’t self-defined – like Pipl, or of course Google – you can get a broader swath of data, but the seemingly random nature of that information brings you back to the motivation – why are you searching for a person in the first place?

If you’re stalking gathering information, then you want as much as you can find and want to pick-and-choose the most interesting data (and might be willing to pay $10+ for more information). If you’re looking for someone from your past, same deal. In either case, more information is better, and so (assuming reasonably accurate matching) Google or the ilk do fine.

But if you’re looking to connect – which, fortunately for us at WhitePages, we know is far and away the most common case – then what you need is comprehensive data, not just the things crawlers happen to find online in whatever subset they like best, and you need contact information that doesn’t require you to join a third-party service. (You also need accurate data, but that’s for another post.)

So it’s not surprising that these engines aren’t that impressive (yet?), when online data is their most common source. Try them against the members of your family or outer circle who don’t have Facebook profiles and they’re even less impressive.

(BTW, thanks to the folks who said nice things about WhitePages in the comments. I don’t believe they all work here…)

Whozat? – an early look. Ow, that hurt my eyes.

A few weeks ago, Whozat? burst on the scene as YAPSE (yet another people search engine), with a press release that said

Whozat?, The People Search Engine, was recognized by the prestigious TechCrunch20 competition among the 100 hottest new start-ups on the planet. In a blind study of the relevance of search results for ordinary people, Whozat? beat every other search engine by a large margin.

(Other parts of the press release change “on the planet” to “on Earth” (thx for the clarification) and “in the world.”)

I’m always interested in YAPSEs, so I figured I’d apply for the beta and wait. The signup page is a bit, umm, colorful (see the face in the ?), but why not:

The best part – which you can’t see in the screenshot and is no longer there – was the <BLINK> tag around the “Recognized…” line. I’m very sad it’s gone – who sees the BLINK tag anymore?

Today my beta account became active, and here’s what I got (after logging in from a Windows dialog)

OK, still talking about TechCrunch20 which has become TechCrunch40, and there’s no way to log out. But fine, I search – type in a first and last name, wait 15 seconds, and get

I’m sort of at a loss here. Three background colors,  weird phrases in tags, dropdowns (the ones on the top right) without content, colored backgrounds used to strange effect (two are actionable, one is not), strange use of images when text will do, keywords to vote on which are clearly meaningless – and, if you look at the text results, a remarkable number of spam/stolen content sites below the fold.

When I search for my father, I do get this:

For the record, George W is not my father.

There are some clever bits – the “select the one you are looking for” actually seem like names. Well, that’s one. But then you have dropdown content like

and you know that there’s just a whole bunch they haven’t thought through yet.

Anyway, these guys may be presenting at TechCrunch60, so I may be stealing their thunder. Hey – good luck. I’m going to look more when it isn’t a Friday afternoon…

People Search is The New Hotness

Yesterday, Twitter sent me mail advertising its new “People Search” feature.

It’s new feature season and we’re starting with People Search.
This new Twitter feature is great for finding more people to
follow because it searches profile information such as name,
location, bio, and url. Come on by and find out if your friends
are already Twittering and you just didn’t know it!

As a techie geek, I’m obligated to love the Twitter, and I do – I tell it when I have a headache, I see what people who I never see personally are doing with their Xboxes, I click more tinyurl‘s then I ever used to do. I don’t understand why I would use both Facebook status and Twitter, but if I think hard, I’m not sure why I would use either, so I try not to think.

But here’s the thing – searching for Bob’s Twitter profile isn’t a People Search, anymore than searching for e-mail Bob sent you is a People Search. If People Search means anything, it has to mean trying to find important and canonical data about a person – practical contact information, personal history or genealogy, social network filtering, professional data, pedophile-related tag spam (I kid!), etc. It can’t mean “does Joe use service X,” even if service X has a tiny bit of profile information available other places. You wouldn’t call deliberately searching for my blog a people search, for example: you’d call it a blog search.

I’m not actually annoyed by this – I just find it interesting to see how the “People Search” term has taken off, such that it’s either a marketing checkbox or an easily-typed, if inappropriate, moniker for new features. It’s a good time to be in the People Search business – and, perhaps, a good time to look like you are.

Spock’s Scary Signup

Jeremy Zawodny’s post on the Spock signup process prompted me to write about another scary step – when you want to claim your profile.

I searched for myself, found myself (that’s a relief), and clicked to claim my profile and said I didn’t have an account. I got the page below:

In this case, I’m being asked to provide my login credentials (over HTTP, not HTTPS) for another site, and there’s no information at all.

Does Spock store these credentials? Does it use them to crawl anything private on my account? What if I change those credentials? And is “sit back” supposed to make me feel relaxed as you crawl through my data? I’m not relaxed.

It’s likely that their use is benign (although they’re still spoofing me on LinkedIn), but there’s no way for me to know that. Most of the users of the Twitter API, for example, message appropriately.

Those ****s you see above? Fake password that I tried on day of launch. Page 404’d anyway.

(Mentioned this to Dave McClure, who’s on their advisory board, during Gnomedex: hoping they do something better here.)

(Disclaimer: my employer and Spock are both in the People Search business. Rising tide lifts all boats.)