Assessment Tips

Is an algorithm better at selecting candidates than recruiters? Can an algorithm replace a personality test or cognitive ability test? Can a recruitment algorithm predict work behavior? Read the advantages and disadvantages here.

A new hype for selecting applicants: not the recruiter but an algorithm determines whether someone fits well with an open vacancy. But does this really work?

How does such a recruitment algorithm work?

A large group of employees is needed to create a recruitment algorithm. These employees provide the data with which the algorithm can do its work. Without this large group of employees (within the same company) the algorithm does not work. Compare it to big data: algorithms cannot do their job without large data sets. For example, from a large group of employees the CV and the motivation letter are analyzed by the algorithm. Some organizations even go further and have their applicants play a game. Based on the scores of these employees (for a well-functioning algorithm, more than 500 employees are required per company) algorithms are developed. The scores of a new applicant are then compared by the algorithm with the best scoring employees.

The online assessment: personality test and ability test

However, when the algorithm has selected a candidate, companies will still personally interview the candidate. Often the candidates also completes a personality test and an aptitude test or cognitive ability test. There is therefore less time gain during the recruitment process than expected. In addition, it is (still) the question whether recruitment algorithms are accurate and / or if they can be predictors of work behavior. Sound scientific evidence (in the form of scientific publications) is not yet available at this moment.

What are the advantages of a recruitment algorithm?

• The algorithm applies the same rules to everyone, a human recruiter can be subjective;
• With large amounts of candidates, the algorithm can save time.

What are the disadvantages?

• The development of the algorithm and the matching dataset is human work and therefore very pricey;
• Cannot be used for specific functions (of which there is only 1 per company);
• An algorithm can overlook things, it only does what it was developed for.

Are you looking for more information about the use of recruitment algorithms for the selection of your future employees? Then contact us here.