Predictive Recruitment: do we still need recruiters?

Recruiters have always studied CVs, analysed career pathways, and evaluated their candidates’ personalities, motivations, aptitudes and competencies. For an equally long time, interviews have been held (in a more or less standard way) and decisions about people’s futures have been made…

But for several months now, a new player has transformed the landscape of recruitment: the “Machine”. Data, algorithms, and machine learning are all revolutionising the way recruiters operate compared to not that long ago…

A question arises (and is perfectly legitimate): will predictive recruitment bring about – in the short or longer term – the end of the role of the recruiter?


#1 – What (really) is predictive recruitment? 

Predictive recruitment refers to a method of recruiting that relies on the use of models that allow reliable prediction of:

  • On one hand the capacity for success that people have for a given position in a particular company,
  • On the other hand, the lifetime of these people in the position (still for a given position in a particular company).

The processes of predictive recruitment rely on the exploitation of all the data available to recruiters (the profiles of people already in positions, their personal and demographic characteristics, the career pathways and behavioural data of candidates…) as well as the power of algorithms capable of sifting through and organising this vast mass of data.

The objective of this approach is to make a qualitative leap in both assessing the match between the candidate and position AND to be able to measure the ROI (Return On Investment) linked to the implementation (what is the concrete return of putting this process in place? Is it economically viable compared to the cost of implementation?).


#2 – What concrete results can you expect from predictive recruitment processes?

When this process is intelligently applied, the most commonly seen results are as follows:

  • An average reduction of delays in the pre-selection stage of the order of 50 to 75%.
  • A reduction in the number of candidates interviewed of between 30 to 50%.
  • An increase of performance of new recruits of 15% on average (compared to candidates selected through traditional recruitment processes).
  • A reduction in turnover after one year of up to 50%.
  • An average reduction of recruitment costs of 20%.


#3 – Will the “machine” replace recruiters in the future? 

In a number of sectors (equally in recruitment as in terms of medical diagnostics for example), it has been found that decisions made by algorithms are on average 25% more relevant that those taken by “panels of experts”.

So, the question may arise (and some people are certainly are asking it): “Do we still need recruiters? …and if yes, for how much longer?”


#4 – Man and Machine are both intelligent…but in different ways.

Today, the “machines” already have more than enough power to surpass human beings in a number of activities that until recently were considered “typically within man’s domain”. As early as 1977 for example, Deep Blue (the nickname of a machine invented by IBM) beat world chess champion Garry Kasparov – hands down. In 2011, Ken Jennings – the most talented Jeopardy player of all time – was forced to bow before Watson (another intelligent system designed by IBM). Today, Google’s driverless cars are already criss-crossing Silicon Valley’s roads in California…you have to admit it’s a little worrying, don’t you?

On the other hand, Google made headlines in 2012 by announcing that one of their data processing centres was – after examining 10 000 000 YouTube clips – capable of identifying a cat…with almost 75% accuracy! This can also seem impressive.

Until we remember that a four-year-old child is capable of doing the same thing without a single mistake! ^^

It’s easy to see – just from these few examples – that man and machine are not best suited for the same types of tasks…


#5 – From “predictive recruitment” to the “augmented recruiter”

In fact, what you have to keep in mind is that when we talk about “predictive recruitment”, it’s that it is to help recruiters. Basically, no-one is capable of managing thousands – let alone hundreds – of pieces of data at the same time, of developing a relevant summary to be able to make a fully informed decision at the end of the process, while giving sufficient consideration to every aspect.

That is exactly what the algorithms are designed to do. For information, a human being is capable of memorising “7 plus or minus 2”, in other words between 5 and 9 things at once, what we call the “memory span”. It must be recognised that from this point of view, we’re a long way from being able to match the machine.

Are the algorithms infallible even with all this? No. Certainly not. Each model is only intelligent to the extent of the rules that make it up, and will always have a certain margin of error. Do they perform better than the best human recruiter? Yes, assuredly and unequivocally. Refusing to accept this fact would be like refusing to admit that the earth is round.


#6 – What is the role of recruiters in predictive recruitment? 

Even if the predictive recruitment process is capable of managing – better than us – an efficient preselection (for example, identifying in 1 second from among 500 candidates the 10 who – objectively – have the strongest chances of success in a given position, in a specific company), none of them are able to create a real relationship with the finalists, and/or to be able to sell them an opportunity in an engaging way.

In fact, every time you have to make a complex decision involving emotional or interpersonal aspects, human beings surpass machines, by far. And that’s the whole point of predictive recruitment: to offer recruiters the chance to take back the time that is so often missing to be able to dedicate themselves to the thing that ONLY they can do well: concentrate on the relationship with their candidates!


#7 – Who will the future belong to?

For recruiters, there are two possible positions:

  • There are those who – like a deer transfixed in a 4X4’s headlights on the edge of the woods – see and will always see a threat in predictive recruitment. These are the people who will hide behind the argument “Hey…we’re dealing with human beings here! Surely we’re not going to automate our preselection!”. While in reality every study on the subject has proven the superiority of algorithms over subjectivity (no matter how enlightened it may be). Because in the end, what really matters? To take satisfaction in thinking that we’re smarter / more competent / beyond perception bias, OR to recruit people who are capable of performing well AND being satisfied AND committed in the long-term?
  • And then there are the others, those who understand that we’re standing at a turning point in the history of talent management, and that predictive recruitment is a real opportunity. For them, there is no question of being disempowered, quite the opposite in fact. They’re more mature and have – from a psychological standpoint – renounced the “fantasy of omnipotence”. They’ve understood that they have everything to gain from relying in part on systems capable of managing a level of complexity way beyond humans, and using these to help them make better decisions. These people are also those who understand that the machine will never be able to match them…and that’s without doubt why they’re at ease with predictive recruitment: they don’t feel threatened at all, in fact the opposite.


#8 – Why it’s crucial to jump on the bandwagon…

The idea – at the current level of understanding – is no longer to argue about whether predictive recruitment can really be relevant (which has already been amply demonstrated), but rather to be able to make the most of these predictive recruitment systems, respecting both the interests of the companies but also those of the candidates.

What recruiter today – for example – is capable of coming up with 5 to 20 alternative positions to a candidate who isn’t suitable for a position? Who really takes the time to do this these days? Those who fully understand the power of predictive recruitment…and certainly not those who refuse to see that it’s changing for the better.

Will certain recruiters disappear? To my mind, yes. Personally, I’m convinced that those who refuse to take the opportunities offered by technology – in the name of frozen ideological positions – are a dying breed…a bit like the taxi drivers who refuse to see that their job has changed, or bookshops that are trying to introduce taxes on Amazon instead of concentrating on what they bring their customers in terms of relationship and experience (which no algorithm has – for the moment – been able to provide).

On the other hand, those who know how to identify opportunities and who understand that their day-to-day activity – as well as that of their candidates – can be improved thanks to suggestions based on taking into account thousands of variables; they will be the big winners of the digital transformation.

Why? Since, freed of their fears, they will know how to put all their empathy and their aptitude into creating relationships or their capacity to influence (in short, the things that make us human!), into activities that will doubtless never be within the reach of the machine (…or at least not in the near future), and which are the activities that have a real, definitive power to transform businesses.



David Bernard

CEO @ Co-founder @AssessFirst


Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: