Shyamal Majumdar: Are you recruiting well?

THE HUMAN FACTOR

Intuitive hiring processes have a 40-50% failure rate. It's time to shift to a data-based approach.

A recruitment consultant, specialising in the IT sector, was looking for a candidate for the post of Vice-President in a top information technology firm. A couple of weeks went into researching the search assignment and another couple of weeks on identifying and holding preliminary talks with three of the 10 potential candidates. The consultant then organised separate dinner "interviews" between the company's CEO and the three candidates. At the end of it all, the CEO said he found none of them suitable.

Another consultant did an almost similar extensive exercise for a diversified Indian group. The result was the same, as the company changed the entire job brief mid-way. "They initially wanted a CFO, then changed the designation to President (Accounts) without realising that the jobs are inherently different and require different skill sets," the consultant says. Most of the candidates the consultant had approached cried off on the grounds that the company wasn't clear about what it wanted.

These are just two examples of bad recruitment policies in many Indian companies. They blow up a fortune on the best ads and hire expensive search consultants. Then it all goes down the drain when they get no one, or worse, hire someone who fails to deliver.

Recruitment consultants say most companies have a 40 to 50 per cent error rate in their existing hiring processes but choose to continue with them. This is surprising considering that no other processes in a company are permitted to be so random. In fact, the same companies would not hesitate to spend crores of rupees to re-engineer other flawed processes that have far less error rate. Not putting in the same efforts in faulty hiring processes can, however, be costly as studies have shown low-performing companies in this area have nearly twice as much turnover among top performing employees as high-performing companies.

Lack of a clear idea about the job profile and failure to brief the consultants properly is one major part of the problem. The other is that most companies still bank on typical assessment methodologies that rely too heavily on academic grades, degrees from top schools, prior industry experience and subjective interview results. Most managers still fall prey to the impulsive hiring of energetic, attractive and articulate candidates, as emotions, biases, chemistry and stereotypes play too big a role.

Yet another problem is very few companies are evaluating if their recruitment process is providing value for money. Surveys have shown (the Recruitment Confidence Index produced by the Cranfield School of Management is one such study) that less than half of the companies systematically evaluate the success of their recruitment processes. The studies suggest that when recruitment is unsuccessful, recruiters just spend more and more on the same processes, rather than systematically assessing the success (or failure) of the methods they are using and making changes accordingly.

The Cranfield research shows that around a third of organisations worldwide have invested in a recruitment management system that may be able to track the success of recruitment in terms of costs per hire, but only around half of these are using the system to produce statistics that could assess the success of individual recruitment methods.

The good news is some companies are doing things differently and achieving spectacular success with their recruitment practices. These companies are reducing the subjective element in recruitments by using basic statistical regression models.

Google is one such company. The search engine powerhouse has made a seamless transition from the subjective and intuitive approach to recruitment to a scientific, data-based approach to selection. Known as the Algorithm Candidate Screening Model, the new assessment tool helps Google to identify candidates that resemble existing top performers.

The model works as follows. First, all existing employees are assessed on a variety of parameters: teamwork, experience, academic background and so on. Next, the model determines which of these traits the existing top performers exhibit that differentiates them from the average employees.

Finally, each candidate's résumé is screened and given a score from 0 to 100 based on traits that match with Google's existing top performers. The candidates are called for an interview only after the initial shortlisting is done through the Algorithm model.

The process achieves three objectives: it reduces the subjective element in interviews; increases the success rate of choosing the right candidate; and saves costs as the recruitment department can spend time more productively rather than screening the résumés manually.

More importantly, Google is able to hire more innovative people who would have got rejected by the traditional résumé-screening methods because of their not-so-bright academic credentials.