The use of algorithms during recruitment: is it really the panacea for all ills?

28 August 2019

Companies invest millions of euros in recruiting people to do their case and allowing them to achieve their growth goals. In an attempt to reduce the costs and risks of this process – for example by taking on candidates who do not meet the required profiles or who leave the new company within a few months, a very widespread reality among young people known by the term ‘shopping around’- numerous companies in recent years are resorting to algorithms, trusting in their effectiveness, according to some studies.

For example, a recent research conducted by the University of Toronto in Canada that reviewed over 17 studies on the various assessments of candidates made by humans and algorithms, showed that recruitment done through algorithms is 25% more effective (su three parameters such as: assessment of the boss, number of promotions and learning skills following a training) compared to the assumptions made by traditional recruiters, regardless of the position for which the interview is done.

The reason for this success? An algorithm is able to attribute the correct weight to the information present on the CV (such as, for example, the weight attributed to the degree, post-graduate courses, previous experiences, and so on) while the human being could decide to base his choice on less relevant aspects of the CV, such as giving a higher rating to a candidate only because he / she studied at the same university as the recruiter.

But are algorithms really a panacea for all ills? As discussed in the Digital HR & People Analytics course, organizations like Amazon have decided to reduce the use of algorithms in the recruitment process for some “undesirable” and potentially discriminatory effects. For example, the Amazon algorithm, based on the historical performance of various employees, continued to attribute lower scores to female candidates than men due to the historical numerical disparity between men and women. Similar problems are found in other companies with algorithms that systematically attribute more positive evaluations to white candidates than black candidates, perpetuating historical stereotypes that new technologies should help eliminate.

The scholar Peter Cappelli, in a recent contribution published in the Harvard Business Review, underlines with a further example the potential risks related to the exclusive use of the algorithms to carry out the recruitment of candidates. For example, it is known that the distance between work and home affects the turnover of candidates: candidates who have a greater distance to travel to reach their workplace are also those who also have a higher dropout rate. This could induce an algorithm to give a lower score to these candidates than those who have a shorter distance. However, the choice of where to stay depends on many factors, first of all the cost of buying / renting the apartment which is linked to individual income. Penalizing those who live far from work could mean excluding those who live in peripheral and more disadvantaged areas from the recruitment process, thus penalizing those belonging to minorities.

What solution can be adopted? In many companies, mixed solutions are increasingly common where algorithms are used to make a first and accurate selection of the best profiles, leaving managers with the task of making the final choice on a narrow range of candidates. Similarly, several companies are investing to expand the databases on which the algorithms work, adding thousands and thousands of observations to those already existing with the aim of “training” the machines to make more accurate judgments.

Other scholars propose more innovative solutions to curb the issue of rising recruitment costs. If selecting new talent is so difficult and expensive, a possible solution could be to reduce the need for recruitment, not with a view to preventing or limiting the growth of the organization, but rather with a view to favoring a higher rate of retention indoor. To do this, it is advisable to invest in the onboarding and socialization process – topics covered during the Recruitement and Onboarding course – so that new hires are enabled to be productive from day one and feel integrated into the new reality: almost 50% of new hires leave the company before the first anniversary. A further effective strategy is to promote any free positions in the internal market to encourage greater staff mobility. This strategy has proven to be very valid in reducing the turnover rate in companies such as Credit Suisse which, analyzing the various causes of staff turnover, found that for people who frequently change roles within their organization, they are more reluctant to leave it. Today the company internally posts over 80% of the candidates and also proposes positions to specific candidates when it sees opportunities for growth.

Authors: Gabriele Morandin and Marcello Russo

Co-Directors of Studies of Master in HR & Organization

Some ideas included in this contribution are taken from the following article: Cappelli, P. 2019, “Your Approach to Hiring is All Wrong”, Harvard Business Review

 



Apply

Back To Top