If soft skills assessment is productivity nirvana, why isn’t every company doing it?

Every year, LinkedIn does an annual ‘Global Talent Trends Report’ that is widely read across the industry. The biggest talent acquisition problem that the 2019 report published earlier this year calls out is one that the industry has known for years — soft skills assessment. 92% of the polled 10000 recruiters say they matter as much or more than hard skills, but only 41% even claim to have a proper methodology in place to assess soft skills.

As LinkedIn puts it:

“This growing disconnect between the demand for soft skills and the inability to identify them is reaching a tipping point.”

The big questions that arises is : if psychometric assessments have been mainstream for at least 50 years and behavioural interviews are the norm, then why do we have this glaring gap?

Three words. Cost. Experience. Trust.

Traditional psychometric assessments are expensive, the ones that actually work cost anywhere between $50 -$1500 per assessment. Also, they are long, the good ones take at least 30–60 minutes. At this cost, there is just no practical way for many to consider the option. Organizations are anyway reluctant to do anything that can impact the candidate experience negatively.

Combined with the impractical costs of such an exercise, soft skill assessment at scale becomes a non-starter. Just as Linkedin has called it out in its report.

Gamified assessments are a step in the right direction as they reduce the assessment time to 15–30 minutes while also creating a positive candidate experience. They have started to see some success, especially for campus or junior roles. However, for managerial and other senior roles, organizations have concerns with asking senior candidates to play lightweight games.

Gamified assessments have also lacked behind on building scientific validation with the I/O psychology community for their methodology, further making it hard for organizations to build an acceptable level of trust. Video assessments also face almost an identical set of challenges.

To get over this roadblock, tech-savvy organizations are now actively adopting the third option — the one that Josh Bersin and Tomas Chamorro-Premuzic call ‘candidate data mining’ in their MIT Sloan Management Review article aptly titled ‘ Three new ways to gauge talent and potential ‘. It would be better described as instant assessment or perhaps as no-test assessment.

Our work at Humantic lies exactly in this area, we provide ‘no-test assessment’ by recycling existing data from the resume, Linkedin profile or any other information already provided by the candidate. The candidate doesn’t have to do anything extra, at all.

As a concept, the industry sees this as potentially revolutionary; for the first time they are able to imagine how they can finally assess soft skills and personality at scale. And at 1/10the cost of traditional assessments, it actually becomes practical to do it at scale.

The players in this segment also has higher focus on proving scientific validity — IBM Watson has voluntarily shared some of its studies and at Humantic, we have also supported studies by the I/O Psychology community, with very positive results. Our approach also addresses concerns around privacy and bias effectively and by sharing necessary details about our algorithms, we make sure that it does not become blackbox AI for its potential users.

It is quite possible that a potential solution to the biggest talent assessment problem might include a combination of all three approaches — gamified assessment for campus and other junior hiring; no-test assessment for all large scale hiring across managerial and leadership roles; and traditional psychometric assessment for re-validation or role specific assessment for a select few critical roles.

One thing that is certain is that winds of change are on the horizon and we might finally be entering the final stage where we can solve the biggest problem facing talent acquisition once and for all.

- Amarpreet Kalkat

Founder, Humantic AI

Originally published at https://www.linkedin.com.