Thursday, May 21, 2009

Google - HR prediction algorithm

After losing some top execs to competitors like Twitter and Facebook, Google has reportedly developed an algorithm that helps them predicts which employees are likely to leave, and when.

In times of economic turmoil, everyone thinks business goals are all about 'trimming the fat' and cutting costs. And that's partly true - as a first step, any business worth their salt needs to be remain profitable and able to compete. Clearly, that means getting rid of any factor that is costing more than it reaps. So tough decisions get made, these impact on individuals, and can - in turn - affect morale and productivity.

After a frenetic period where we've seen many previously untouchable organisations go bust and others bailed-out, there's a more proactive approach emerging - and this is taking the form of workforce planning.

If you're unfamiliar with this term, workforce planning in its simplest form is a matter of anticipating the how many people and what skills are needed in what areas of an organisation - and setting out a plan to meet needs. It's the opposite of the reactive HR approach that puts an ad in the paper when an employee leaves.

So right now, many businesses are not recruiting. So why is workforce planning relevant?

The answer is that workforce planning also happens within an organisation. Employers look at where they need people, and where they don't need people. They look at what skills they have, and what skills they need. They look at unprofitable areas of the business, and work out strategies to minimise pain for affected employees.

And this explains part of Google's move. Google will not want to lose key talent. High performers are an intangible business benefit that will be lost from them, with their competitors only standing to gain. So it makes sense to work out who at Google might leave, and when. But how do you work out strategies to keep them?

There's plenty of evidence to suggest that businesses that manage, engage and inspire their top talent, are much more likely to engage them in tough times, retain them once the upturn comes. There's a huge competitive advantage to be claimed via this method, but this algorithm is a bit like knowing your mum's birthday date, but not neglecting to ask what she'd like as a gift.

If I was looking at this sort of problem for Google, I would ensure extensive qualitative research (workshops, interviews, and the like) was conducted. That way, you can get beneath the skin of a problem and assess they 'why' as well as the 'who' and 'when'.

Statistics and data will only ever give you a wide but shallow pool of information. It's only through using a talented interviewer who can cut through the surface to uncover unconscious beliefs, drivers and ideas of individual employees that you can formulate strategies to affect real engagement. An algorithm, no matter how advanced, could never do this.

As a sidenote, this Google announcement reminds me a little of a PR exercise perfected by Ryanair (low cost airline). They released a story in February this year announcing considerations of a one pound charge for customers to use the toilet.

Customer outtake: Ryanair are obsessed with keeping costs low.

Customer outtake here: Google is obsessed with the perfect predictive search tool.

No comments:

Post a Comment

Say what you like - honest - but don't be abusive hey, I'm not pretending to know it all.