An age-old debate that rages forever is the decision as to whether recruitment is a science or an intuition-based skill.
A few years ago, experts would have said that there is no way to predict the behavioral intelligence of a human. In a way it’s true, there is no traditional algorithm to select the right candidate. But there could be a self-learning intelligent model that can be built for every single job opening.
The movie “Moneyball” explained how few age-old baseball scouts strongly used their intuition and predictive skills to judge whether a person is a good player or not. In modern recruitment practices, intuition and guesswork play a huge role in the selection of the candidate. The bitter truth is that people have hardwired systematic biases in how they evaluate candidates. The more experience a recruiter has, the more intuition comes into play.
While skill-based assessment tests proved to be a good solution for a brief period of time, it is beginning to become ineffective for senior resources. This is because the same skill-based tests can be cracked by any junior resource/student who has an intention to learn fast. Knowledge is available everywhere, so learning a skill that someone else possesses has become very easy.
What needs to be assessed is the behavioral ability of the resource. To break it down further, there are close to 200+ traits that play a crucial role in selecting a candidate. Below are the traits that play a vital role in deciding the candidate’s fit.
The best possible way to measure these traits is by presenting the candidates with a situation and noticing how they would react. A good example of a situation-based judgment question can be found below.
Question Source: Xobin
The above question checks how a candidate reacts to the given scenario. There is no right or wrong in selecting an option, we are just measuring how the candidate would react to a particular situation. Usually, a group of questions is mapped to these traits and we can get individual scores for all the traits. A 90-minute test could help us get scores of 35 such key traits that influence the decision of that particular role.
However, it is very tough to have a standard algorithm because every company defines a job role differently. For example, expectations from a “Sales Executive” in Company A could be different from a “Sales Executive” in Company B.
The most preferred solution is to build a Machine learning model that trains from the historic recruitment decisions of the company and helps predict the best suitable candidate for the role at present.
For example, the below workflow was used (now deprecated) in Xobin for a brief period of time, but it does the job of explaining the concept in layman terms.
The catch with this is that it requires at least 100 complete data sets to train the model. While big enterprises don’t have a scarcity of data, smaller companies find generating this data very challenging. However, several pre-trained models are now built, which do a pretty decent job of helping hire senior resources that save the recruiters time and get a huge nod from the business heads as well.
The purpose of such systems is not to replace recruiters, I personally feel recruiters can never be replaced. But, such systems can help them make more scientific decisions in a more efficient way.