Predicting candidate uptake on individual online vacancies and vacancy portfolios

The internet has undoubtedly had an substantial effect on how organizations and job seekers behave on the labor market, which has been beneficial for both job seekers and organizations. However, despite these benefits, it also comes with difficulties. Organizations might observe both applicant excess and applicant shortage on their vacancies. The problem of either applicant excess or shortage has been addressed by previous studies, which frequently conclude that this problem is inherent to the process of online recruitment. The usage of analytical techniques might reveal new insight into how organizations can account for this problem. This paper therefore studies how the number of job- applications on online vacancies in a particular week, which is referred to as the application rate, can be predicted and controlled. To answer this question, a dataset originating from a large Dutch organization was used. This dataset contains recruitment outcomes over a period of three years and just over 5,000 unique vacancies. This study trained multiple machine learning models on predicting the application rate. Furthermore, it analyses the predictability of the total number of weekly applications over the entire vacancy portfolio, and how both the application rate and the total number of applications is affected by the usage of online marketing campaigns.