Ever wondered why it is so hard to find the right people for your business? Why so many hires turn out to be the wrong person? If you are expecting me to say ‘because it’s hard’ then you are mistaken. According to the latest Harvard Business Review (May 2014), the real reason is we’re just not very good at it.
The short but punchy piece by Nathan R. Kuncel and Denis S. Ones from University of Minnesota and David M Klieger at Educational Testing Services is entitled ‘In Hiring Algorithms Beat Instinct’. In some ways it is quite scary. It succinctly reveals just how bad we humans are at choosing the right candidate for a job. But what makes it really scary is the weight of evidence they have to show this. They cite 17 different studies demonstrating that even a simple equation outperforms human decisions every time. This is true in every case and in every algorithm they looked at. Nor does the number or level of the candidates involved matter.
The conclusion is we are more easily distracted by things that might only be marginally relevant than we realise. We use information inconsistently. We are easily thrown off course by inconsequential data such as a compliment or a remark on an arbitrary topic. This undoes all the good work that went into establishing the job parameters in the first place.
The conclusion is clear. We are just better off leaving the whole process to an algorithmic system based on a large number of data points. So a well set up recruitment system (such as Saba Recruiting@Work), giving you access to powerful machine-Learning algorithms and predictive capabilities will make far better decisions and quickly provide measurable return on investment. All you have to do is trust it!
There is the rub of course. It is hard to imagine that some managers would not find a way round a numbers-only hiring system. They will not believe a mere machine can beat all their accumulated knowledge. What can replace a piercing look into a candidate’s eyes? But the numbers are compelling. To not use a recruiting system with the best possible algorithms, at least narrow the field as far as possible, could soon become a thing of the past.