Advertisement

The messy world of defining skills

Hoping that HR technology can simply list out all the skills a job requires is a long-shot at the moment, says David Creelman, meaning a more nuanced approach is needed:

Article main image
Dec 22, 2023

We’ve known for a long time that the world of work we once knew is no longer going to be the world of work we know in the future.

We’re often told that to prepare for workplace demands of the future, we need to create skills-based organizations. But I would add to this ‘skills-based technology’.

It’s exciting but like many new ideas, it also throws into question the extent to which it is helpful or just another example of HR hype?

Critically, if you are responsible for HR tech, then that’s a question you need to be able to answer. So how?

Understanding what skills-based technology means

The core idea underlying skills-based technology is that jobs and people can be understood as a bundle of skills.

The promise is that if we decompose each job/person into a list of their individual skills, then we can do everything from picking the right hires to identifying training needs to discovering career paths.

But as is usually the case in HR, this is partly true and partly false.

It is a sad thing to work in a world where all the key ideas are partly true and partly false, but it is what it is.

What’s true?

If you think critically about decomposing jobs/people into a bundle of skills the values and limitations are not hard to see.

For example, if you are trying to fill a job, then of course one of the questions you will ask is ‘what skills does the job need?’ In doing so, one might consider looking at other careers. For example, HRDs might look at a job in marketing and notice how some of the skills overlap with some finance roles. That opens up opportunities to consider people they might not normally have considered.

In this way, looking at skills can be great for HR.

What’s not so true?

On the other hand, however, when you think about what it takes to accurately define a skill, you soon see how impossibly messy it can get.

Take something simple like the question of whether a job needs skill in Excel.

It’s clear it’s not a “yes/no” question.

You need to define the level of the skill. People often decide there are three levels: beginner, intermediate, and advanced.

But deciding where an individual fits isn’t straightforward.

Worse someone with “advanced” skills in one aspect of Excel such as writing macros may have no skill at all in another area such as using statistical functions.

The bottom line is that it’s easy to say, “Let’s list the skills”, but when you try to do so rigorously it becomes impossibly complex.

Now try to do it across the organization, for all jobs (and all people), and the scale of the task may be unmanageable, and the results may not be especially useful.

There is also the case that when you take a person or job and reduce them to a list of skills something is lost.

It’s hard to describe exactly what that is, but in the case of a job we might simply say “context”. For example, an HR job in manufacturing may have the same list of skills as an HR job in investment banking, yet those jobs are different in important ways.

The skills paradigm, as is often presented by vendors, glosses over these complexities and limitations of building your HR system around lists of skills.

It’s also far from being a new idea. I remember doing similar work decomposing jobs into competencies over thirty years ago. There’s a reason why a skills approach to HR hasn’t become the universal solution to HR management.

What’s new?

What is new, however, is that with machine learning tools we can get much better at identifying skills from a job description or job posting.

Given there are millions of job postings on the web, AI can scale up our analysis of the skill sets in different jobs in a way we never could before.

This is a good thing. It does give us a reason to look at what new tools can do. We just shouldn’t buy the pitch that a job or person is nothing more than a collection of well-defined skills.

What to do

Perhaps the best tactic for dealing with the so-called skills-based revolution is to put aside the notion that we have a new paradigm.

Instead, see what specific applications these new tools are good for in your own business.

For example, if you are in a big software firm you might find AI’s ability to keep a skills taxonomy up-to-date is incredibly helpful.

But when you test it on identifying career paths, you might find it is useless.

In other words, it’s all case-by-case and you should move forward cautiously.

The worst thing you can do is invest in an expensive software suite, and then hope to accurately determine the skills in all your jobs/people based on the false premise that this will revolutionize everything you do HR.

Decomposing a job or person into skills has both real value and serious limitations.

So take your time in figuring where the approach is practical in your organization.