So much of recruitment is like putting square pegs into round holes.
Roles require a unique combination of skills and abilities. Yet when candidates are selected, many of their attributes are unknown.
The CV, for example, shows qualifications and experience and lists references. But, does it really tell you if the person you are thinking of hiring is a good team player, or whether they are good in a crisis?
It Is Always a Tech Problem
As in so many areas of HR recruitment, there is – or claims to be – a technology solution for that, and it is all about predictive data modeling, machine learning, and artificial intelligence.
Pulsifi is a two-year-old company based in Kuala Lumpur, and Singapore with an offering co-founder and chief executive Jay Huang described as a "People Analytics Software Platform."
This not only helps to identify the right candidates for the job but also predicts how they will perform.
Huang used to work in customized marketing, where he took data and made predictions on consumer preferences to tailor marketing and advertising. Now, he is using a similar approach in the area of HR.
Pulsifi, he said, solves the recruiters’ dilemma of not being able to read every CV. It not only analyses all applications but also casts the wider net through partnerships with online job boards and third-party databases.
“If you look at what HR is like today, you see that so many hiring managers are flipping the CV’s of candidates,” said Huang.
"When you ask what they are looking for, they say ‘team players’ or ‘people who are motivated to learn’ or ‘communicators.’ But still, organizations know very little about their people and how they will perform.”
This also creates a big downside for employees. Many of them are in the wrong roles and find it hard to perform to their fullest potential.
“That’s why it is no real surprise that so few people look forward to coming into work on Mondays,” said Huang.
Searching Internal and External
Pulsifi’s software inputs not just CVs but other data, such as psychometric tests, video interviews, and a candidate's online footprint.
From there, the software identifies candidates and then makes predictions on how they might perform, giving them a “fit score” based on their suitability. The profiles are developed based on up to 2,000 identified competencies.
Pulsifi is also able to include not only external candidates but also existing employees and “passive candidates from databases.”
Potential candidates are also benchmarked against people in current roles to understand what is required for people to do well in particular roles.
"We do not have to invent new data sources," said Huang.
"Often, these sources are already there, like CVs, of course. And many organizations use personality tests and video interviews, so we put all this together and make sense of it.”
The software delivers scale that humans could never achieve. But Huang said it also provides a more holistic view of people – even one which can be more specialized.
"We are taking data points about people and optimizing them against work outcomes," he said.
“For example, you may have a sales role, and you want to optimize for people who can achieve higher sales. We [take] data points and put them into our model with inputs linked to higher sales, and the model finds people with these characteristics.”
Some of the insights gleaned so far have been counter-intuitive.
For example, Pulsifi found that people in sales in the financial industry don't like to be leaders and don't like multitasking.
“It just shows that so many times what organizations are trying to do with people just goes against their nature,” said Huang.
Pulsifi was engaged by Nestle Malaysia to assist with the company’s recruitment for its management trainee program and is using this as a case study.
Nestle receives thousands of applications each year for this program. It has historically reviewed the applications manually.
The Pulsifi software did this automatically, predicting and assessing a range of hard and soft characteristics.
The outcome was that the top 200 candidates were chosen from over 300 applicants in a process that saved the in-house recruiters 70% in terms of man-hours.
Beyond that, 97% of the hires performed as predicted, and there was a claimed 95% improvement in candidate selection.
The AI Factor
Overall, Pulsifi claims up to 90% accuracy on its predictions, “as validated by its users.”
That figure, said Huang, can get even better as the AI improves, and the methodology develops its ability to better predict employee suitability.
Perhaps round pegs can go into round holes after all.