Dan Collier is the CEO of recruitment platform Elevate. In this article, he discusses how intelligent machines could potentially revolutionise the recruitment process.
While most employers would agree that hiring someone who’s the right fit for their team is as much about chemistry as their CV, for smaller companies this decision is far too important to leave to a hunch.
Making the wrong hire is always a costly mistake, but for a young tech business with limited funds and investors keen to see returns, it can be disastrous.
One obvious way to derisk the process is to hire skilled contractors on a project basis rather than take on full-time staff.
Not only does using contractors offer an employer greater flexibility, experienced professionals are particularly valuable to young firms that are still perfecting their technology and need to get the job done as quickly and painlessly as possible.
Hiring bright young grads and allowing them to learn on the job is all well and good if you’re Google – but for the CTO of a tech startup with a product to launch, using a contractor is a quicker and safer bet.
Intelligence versus instinct
But even employers who plough through dozens of CVs before choosing someone tend to rely heavily on gut instinct.
While this approach can get it right sometimes, it is a poor predictor of how a candidate will perform on the job.
However huge leaps forward in data science mean that computers are moving from just helping directors choose who to hire, to actually choosing staff themselves.
The idea sounds fanciful, and many company bosses would question whether a computer programme could ever make such a fundamentally ‘human’ decision as choosing who is the right fit for their team.
But the reality is intelligent machines – which use complex algorithms to identify the traits shared by a company’s most successful staff and then measure candidates against them – can often choose new recruits better than the employer.
The key is predictive analytics – complex algorithms which deconstruct the characteristics and experience the employer says they’re looking for in the job spec, as well as looking at who they’ve hired before and noting which types of employees have gone on to be star performers.
Armed with this knowledge, the systems suggest to the employer what they’ll need to pay to attract candidates of the level they want. This is crucial not just for budgeting, but also in ensuring the company strikes the right balance between finding sufficiently qualified people and making sure they don’t overpay.
Rise of the machines
Now for the really clever bit.
As the candidate applications come in, the system assesses the suitability of each would-be employee, scoring them on their previous employment and education, building up a detailed picture of their character and rating their chances of thriving in the role they’re applying for.
In other words, the technology learns which are the best types of people for any given company to hire, and uses data from multiple sources – CVs, social media and personality tests if they’re available – to give a highly accurate ranking of applicants.
To my mind, this suggests that algorithms will soon be able to make better recruitment decisions than bosses.
This is not to say the final decision should rest with the machines. The idea of supercomputers as ultimate arbiters of a person’s career progression is more than a touch dystopian, and all sensible employers will want to review the software’s recommendations before offering anyone a job or a contract.
Such technology is already transforming the way many companies hire, and dragging the recruitment industry into the data science era.
The combination of vast computing power with the software’s ability to learn which candidates will perform best in any role in any company allows it to apply the same decision-making criteria as the hiring manager, but better, because they’re backed up by data rather than guesswork.
And as someone who has himself founded a tech company – and needed to build a team that could get my product to market before someone else did – I choose data over guesswork any day.