Improving Your PeopleFinder

Vivek Deshmukh over at Boxes and Arrows (Joost thanks again for the pointer!) has an interesting post on a People Finder Application he designed. Our company calls this 'who is who'. 'YellowPages' is also a common name for this application.

Vivek's post is about how they improved their people search based on user input. His approach is straight-forward, but nonetheless very interesting.

For one, I like the approach they take to peoplesearch. Basically they note that many applications have a people finder component to it. It's not only the regular stuff you would find in 'who is who' (name, phonenumber, location, emailaddress, etc.). You need people finder for meetings, visitor registration, etc too. This relates well to 'Identity and Access Management', but extended to non-colleagues, like visitors, too.

To improve their people finder they started collecting data from the way users query, specifically the incorrect entries. Based on the data they learned how they could help users find people more quickly, based on patterns. Eventually they were able to ask the user 'do you mean...?'.


This is pretty straight-forward, I hear you say. True, but how often do we improve our systems by simply collecting data on how users actually use our Intranet and other corporate applications?

By the way, (StepTwo also pointed to this post.)