Slashdot Mirror
Data Mining Moves To Human Resources
theodp writes "Just when you thought annual reviews couldn't get worse, BusinessWeek reports that HR departments at companies like Microsoft and IBM are starting to use mathematical analysis to determine the value of each employee. At an undisclosed Internet company, analysis of (non-verbal) communications was used to produce a circle to represent each employee — those determined to generate or pass along valuable info were portrayed as large and dark-colored circles ('thought leaders' and 'networked curators'), while those with small and pale circles were written off as not adding a hell of a lot. 'You have to bring the same rigor you bring to operations and finance to the analysis of people,' explains Microsoft's Rupert Bader. Hey, who could argue with what Quants did for finance?"
I think if this is used purely to select 'valuable' people by-the-numbers it could be disastrous. If, however, it turns into a data mining exercise, it could enhance the ability of the HR department to ferret out resources and organize teams or research projects. The first thing they should probably do is test out the theory: HR alone, HR + data mining, data mining alone, to see if there is any empirical evidence for making such a move. Maybe they'd then find out that combined, the human + machine performs better because the machine can manage gobs of data more efficiently and the human is more capable of adding qualities that normally aren't encoded into bits. Overall though, it should be tested before implementing. : P