A short experiment in people search

Besides speaking about our shinny app we will use this blog to review applications and/or services that we consider to be in some way connected to what we are trying to achieve with uberVU. We will review apps that deal with semantic web, vertical search, open API’s etc.

This week let’s talk about people search! It is a hot topic for us since uberVU will soon integrate social networking data and we work on a way to match profiles among different social platforms. And people search seems to be a good way to achieve that. For my experiment I chose 2 apps: Spock and 123 People. Spock is the creation of Jaideep Singh and Jay Bhatti, two engineers from IBM and Microsoft. Spock got a lot of attention from the press in 2007 and was nominated among Top 10 Web Start-ups by Wired magazine. 123 People, on the other hand, is an Austrian-based start-up backed by i5invest fund. So …. let’s start rolling the ball.

How it works

Well! You input a name and the service gathers information from all over the cloud: social networks, wikis, google search etc… And it displays all the information on a single page that you can then edit by adding details about a person.

Are you looking for someone famous?

One thing I noticed and were pretty disappointed about was that search usually works for famous people. It works for Paris Hilton or Bill Gates, but if you search a “not so famous” dude you would probably find nothing. The same about searching for your own name.
And this is a big big problem. Because the value of this service, for me at least, is the ability to extract information from the cloud about a person that I know almost nothing about. If all you can do is dig information about super-starts why not use Wikipedia or simply Google?
The fact that the most searched people are stars also makes me believe that displaying telephone and address is a bad idea. You can not find the phone number of Steve Jobs.

Context, context, context

Context is everything. From my experience I rather search for Microsoft than for Bill Gates. And I expect that the service to be smart enough to make the difference between a company and a name. Spock works great in that regard. It finds people connected to a company, an event etc. 123Pople on the other hand is not so precise. Search for Microsoft and the results will be very strange. I finds some pictures and then it displays some people with Microsoft in their names etc. Not very intuitive.
But although Spock does such a good job at identifying context it has no idea about filtering. I mean: I want to find people connected to Microsoft that are attending a conference in my area. That will be super cool.

Conclusion and future development

People search is an interesting area to keep under observation. Spock is clearly a more mature product but it needs more features until it could go mainstream. What could make me become a fan of any of these products:

• long tail indexing. I would appreciate if I could find coherent information about a rather not so popular person.
• API for integration of the product with web apps like on demand CRMs.
• matching. We have so many connections embedded in social networks comments, blog posts. I would love that someone organize this social stream.
• add follow option. What I mean? If I am searching for a person I’ve met a few days ago at a party it would be so cool to find where this guy publishes information and to have the possibility to follow him on some of the services I use:Facebook, Twitter or RSS feed.

What is your experience with people search? Is it something you regularly use?