Wednesday, December 15, 2004

Better recommendations in candidate selection

Two related items came to my attention recently. First was a great (but quite technical) article from 3 German academics – Färber, Weitzel and Keim entitled “An automated recommendation approach to selection in personnel recruitment” and then seeing information on isdd’s OmniQ products in recruitment.

Both seem to tackle the same problem – most HR departments have far too many CV’s and finding the relevant ones is difficult and time consuming.

There are, as I see it, two main ways of tackling the problem – first you can ask job specific questions which can rank applicants for a particular role (of course you might have numerous types of tests which give you further ranking data) or the system takes a previously provided CV and uses analytics (hopefully a lot more intelligent than Boolean searching) to rank the applications. From what I can see isdd and Färber, Weitzel and Keim have seeked to do the latter.

Without seeing the isdd product in action I am not going to be as bold as commenting specifically on it. However the website suggests that it is used instead of a dedicated front-end system. Their phrase “OmniQ will automatically collect all emails and attached documents sent to the company from prospecting graduates” (taken from the PDF on it used in the graduate market) fills me with horror. I suggest it may be better to have the technology built in to existing applications.

Alternatively the industry could press for the adoption of standards such as the HR-XML resume schema that would enable far more accurate searching within CV’s.

This area is one to watch.