Ein Algorithmus wählt den besten Kandidaten aus
Do you sometimes no longer see the forest for the trees? If it is up to the PhD student Colin Lee , it will soon be a little easier for recruiters. Lee received his PhD from the Rotterdam School of Management at Erasmus University in Rotterdam with a unique algorithm for recruiting. Based on the CV, he can predict with 81 percent certainty which applicants the recruiters will invite to an interview.
The research is unique because Lee’s algorithm scrutinized over 440,000 real digital resumes and applications, and actually had them reviewed by professional recruiters. Lee used the recruitment data, which made a large number of Connexy’s customers available to him. The computer model takes into account various variables, such as work experience, age, distance to the place of work and level of education. The software also checks factors such as ‚did the applicant apply in time?‘ and ‚Does the applicant already work for this employer?‘
If recruiters only use the resume, the algorithm predicts the selection of experts with 81 percent reliability. If the recruiters also take the letter of motivation into account, the forecasting accuracy drops to 69 percent. There also seems to be a difference between what recruiters claim and what they actually do. „Recruiters sometimes seem to choose based on intuition, but there is a common thread,“ Lee says. Unfortunately, the data regarding the motivation letter could not be included in the investigation due to data protection reasons.
Big data is already used to a greater or lesser extent by employers and employment agencies who process large amounts of resumes. Intelligent tools in the ATS (applicant management system) are used to filter applicant data in order to create an initial assessment. It’s about data from internal databases, but it’s also looking for public data on social media. „Lee’s algorithm can be a nice addition,“ said Jan van Goch, Vice President Marketing EMEA at Bullhorn .
Predict future performance
Researchers expect that the use of big data on the job market will become even more valuable in the future. „You can measure the performance of applicants who have been accepted and add them to the algorithm. Based on this data, an applicant’s future performance can be easily predicted simply by scanning the candidate’s uploaded resume,“ concludes Lee.