The Human Side Of Data-Driven Recruitment
Late on a Sunday night in August 2017 US Navy destroyer the USS John S McCain smashed into the side of a large tanker filled with nearly 12,000 tonnes of oil off the coast of Singapore. Ten seamen lost their lives in the collision and it took many years before the warship was rendered sea-worthy again. A remarkably similar accident happened just two months previously to the USS Fitzgerald which was struck by a large container ship off the coast of Japan. Seven US sailors died.
The ships involved in both collisions were large and well-fitted with radar and navigation systems. The USS John Mccain also had GPS tracking, automatic identification systems (AIS) and a state of the art Integrated Bridge and Navigation System (IBNS). How could such collisions have happened?
In response, Aron Soerensen, head of maritime technology and regulation at the Baltic and International Maritime Council said:
“Instead of looking at the instruments, you have to look out the window to see how the situation actually evolves,” he explains. “Maybe today there’s a bit of a fixation on technology.”
When you make a hiring decision on a candidate, do you use your head or your heart? Have you ever taken a punt on a candidate just because you had a good gut feel about them, even though their CV wasn't quite up to scratch?
I’ve worked with companies of all shapes and sizes, and whereas larger companies with armies of recruiters have highly systematized recruitment processes, for smaller companies, gut-based recruitment is pretty much all they’ve got.
Ever since ATSs became available, the drive in recruitment has been primarily twofold:
And so, the next logical step is to try and automate the decision-making process further and use metrics from previous campaigns (or current employees) to guide, and ultimately to automate, recruitment campaigns of the future.
Data-driven recruitment has primarily helped in increasing the efficiency and consistency of screening, but, as counter-intuitive as it may sound, it has also allowed for a more personalized relationship with each candidate. Using assessment metrics to screen candidates allows recruiters to feedback to candidates with a high degree of specificity as to why they did or did not get the job – or perhaps where they may be better suited. So ironically, more sophisticated talent tech has the potential to make candidate engagement more personal, or dare I say it… more human.
领英推荐
[As an aside, I should also point out that assessment data on successful hires can also be useful further down the line in their own personal development]
But, a point we’re all too shy to admit, it also helps make sure nothing slips through the net. Recruiters are human and have good days and bad days like anyone, so there are aspects of a candidate that are sometimes missed or ignored – to the benefit or detriment of that candidate’s application. I’ve written previously about bias, but even the most objective recruiter can’t be expected to get it right all the time. Automation can help ensure no candidate gets ignored and that the more transactional monotonous tasks are taken care of, to a high degree of accuracy and reliability 24/7.
Perhaps most importantly though, data-driven recruitment has opened up the whole field of recruitment marketing. You can’t really do any sort of marketing campaign without data about who your message is targeted at, and the accumulation of candidate data has enabled much more targeted recruitment campaigns as well as less formal ‘reach outs’ (ie suggestions of potential passive candidates based on their Linkedin profile or previous applications etc).
But data-driven recruitment does rely on assumptions which can sometimes be called into question.
Firstly, it assumes that more data equals more accurate decision making. What this really means is, more variables equals more complex decision-making. But as we all know, not all data is good data and one still has to be judicious about what factors are being taken into account – or perhaps more importantly which factors are being ignored. As the saying goes, junk in equals junk out. So, for example, if the recruiter is using data based on their current team, they're only ever going to hire clones. Whilst this approach can give the team a sense of cohesiveness, it can make it too homogeneous and prone to groupthink, which doesn’t really do much for diversity or innovation. Sometimes, a bit of tension is good!
One of the benefits of data-driven recruitment is that it can remove references or allusions to protected characteristics (age, disability, gender, etc) from an application allowing recruiters to assess the candidate purely on ability. But the downside is that this makes it more difficult for the recruiter to proactively recruit a more diverse team. If the only candidates that meet the criteria are from the same demographic group, perhaps because the job ad is worded in a way that excludes a particular demographic or protected characteristic, the recruiter will never know.
A common assumption for all of us is ‘has done’ = ‘can do’. We look at CVs for experience of X (or rather, run algorithms to find it on a CV or Linkedin profile) and assume that a candidate that has done X in a previous company can do X for us. But is this always true? Is X something that can be taught (I’ve written previously about the difference between knowledge, skills and attitude)? How important is X for the job and is recruiting for it just reinforcing a bias? To put it in statistical terms: to what extent it correlation reflecting causation?
Lastly it assumes that everything is measurable. I may be a bit old fashioned, but I believe recruitment should have a human component to it. I understand this can be a luxury when it comes to volume recruitment, but I still think that if you’re going to spend the majority of your waking time during the week working alongside someone, you ought to have seen the whites of their eyes first. As good as online assessments are, they can’t truly predict chemistry and all the other soft stuff that goes into making us who we are, and some sort of face-to-face (or 'face-to-Zoom') interaction is needed.
With ever-increasing computer literacy and internet availability it is now easier than ever to apply for jobs – and that’s before taking into account ‘easy apply’ options on most job boards. Sifting and screening CVs can be really challenging, I know, especially for big brands doing volume recruitment; and whilst data-driven recruitment has a lot to offer in marketing and screening candidates, I’d hate for it to take the ‘human’ component out of recruitment. Sometimes, we all just need to look out of the window.