Artificial Intelligence, Hiring & You

techhiringblog_jan20172Recently, I read an article. The author of this article, Recruitment industry will DIE in 2018, threw some light on how Machine Learning is picking up and thus, how many organizations are opting to use it to make their hiring seamless. We, recruiters, need to evolve and prepare ourselves for this change if we want to remain relevant. Some very good examples were cited by the author and; as expected; there were mixed reactions from the readers. Few agreed and others had a different perspective on this.

This article forced me to think about my future. I also had discussions about it with many professionals who are well informed and insightful about the recruitment process. Eventually, we decided to conduct an experiment with few of the open positions that we had. Our thought was to evaluate candidates on different parameters, just like how a computer program would. We created a tracker (an excel spreadsheet) which captured various fields that we have been taking into consideration while screening profiles like relevant experience, skills, qualification, salaries, notice period, etc. We also assigned weightage for each field after talking to the hiring managers. The ‘must have’ criteria were given more weightage and ‘should have’ & ‘could have’ criteria were given a little less weightage. Here is a snapshot of that tracker.tracker_export

Let me admit, initially, it was not working exactly the way we assumed. We were trying to build a spreadsheet to imitate how a human brain decides on the suitability of candidates for a given job description. You can imagine how daunting that task could be. A lot of thinking, brainstorming, and analysis went into it. It was really very challenging for us to reach an agreement. Finally, we decided to go ahead with this tracker which had many excel formulas.

The real game began now. We captured many criteria like qualification, relevant skills (must haves & good to haves), how recently has he used these skills and for how long, salary drawn, notice period, current company, the distance between our office & his residence, etc. We started entering the values for these criteria gleaned from candidates’ profiles and multiplied them by their respective weightages. We named the sum of all these weighted values as “Suitability Score”. The candidate with highest suitability score was supposed to be the most suitable candidate. Obviously, all this analysis was based on whatever the hiring manager had mentioned in the JD and whatever the candidates had mentioned in their respective profiles. Our intention while computing this suitability score was to simulate how a recruiter assigns importance to various criteria while short-listing profiles.

techhiringblog_jan20173_piechart1We didn’t disclose the suitability score to the hiring managers to avoid influencing their judgment as our goal was to compare the ranking based on suitability score with hiring managers’ choice. Soon, we realized that we had to iterate. Our first shot at the scoring logic was way off. It’s very difficult to rationalize a hiring manager’s decision-making process while hiring a candidate or shortlisting a profile.

Subsequently, we also spoke to many hiring managers to understand how they evaluate a profile. Some of them were considering some unsaid criteria like whether the candidate had worked in a product company before or a service company. Others assigned higher importance to good academics or to overseas work experience or to the proximity of their place of residence to our office. All of them had valid reasons based on their past experiences and this is how they were hiring for years and getting desired results.

We kept on modifying the selection criteria in the tracker and weightage assigned based on different inputs that we got from the hiring managers. Eventually, we started coming closer to their way of shortlisting profiles. The whole team was delighted to see this exercise working. We came very close to achieving our goal of saving hiring managers time which was normally wasted in interviewing too many candidates to get one final SHORTLIST.

In this interesting journey of ours, we have come really long way. We did realize that just shortlisting the profiles based on the resumes(data) provided by the candidates is not enough. There was more to be done. So, we went ahead to address the next big challenge! We will talk about it in the next month’s article.

Hope you enjoyed reading.

Please share your views and suggestion:)

I hope you enjoyed this article. Don’t forget to share your views. You can also tweet your comments to @Rezoomex and @Dinesh_Gokhale with #TechHiringBlog


DineshSIr_ProfilePic
Dinesh Gokhale (Intrapreneur, Rezoomex)

Dinesh Gokhale is a seasoned professional with more than a decade of diversified experience into IT recruitment along with B2B sales, marketing, artificial intelligence and lean hiring technology. He is admired for his soft-skills.

Dinesh also is an enthusiastic sportsperson, reader, traveler and a die hard foodie. In his selfless desire of giving something back to mother nature & society, he has been involved in many social initiatives. Currently, he is working in the capacity of Sr. Principal Consultant with Rezoomex.


4 thoughts on “Artificial Intelligence, Hiring & You

  1. Pingback: Hey recruiters, now you too can do technical evaluation! – Tech Hiring Blog

  2. Pingback: Hey recruiters, now you too can do technical evaluation! – Tech Hiring Blog

  3. Pingback: Hire Better & Quicker using Machine Learning – Tech Hiring Blog

Leave a Reply