Incompetency Framework: The Truth About Skills Based Hiring
If you’ve been to an HR tech conference lately, sat through a VC pitch deck, or made the mistake of opening LinkedIn during business hours, then congratulations.
You’ve already been force-fed the gospel of “skills-based hiring.”
It’s 2025, and apparently we’ve all collectively decided that job titles, experience, and education don’t matter anymore. All that matters is “skills.”
Skills, we’re told, are the great equalizer. The future of work is a meritocracy where resumes don’t matter, degrees are obsolete, and anyone who can prove they’ve mastered the right blend of Power BI, emotional intelligence, and generative AI prompt engineering gets the job.
It sounds great – until you realize, like most other talent trends, it’s mostly theater (of the absurd).
Because here’s the thing: skills-based hiring, for all its promise, is fundamentally flawed. Not just flawed in practice, but flawed in premise, too.
The truth is, skills, as they’re currently defined and implemented in most HR systems and hiring processes, are too subjective, too fluid, and too dependent on context to serve as a stable foundation for hiring.
And the companies pushing it hardest? Yeah. They don’t know what skills they have internally.
They don’t know what they’ll need next year. And they sure as hell don’t know how to measure whether any of it actually works.
Let’s break this down.
Dynamic and Decaying: The Half-Life of a Skill
The first problem is epistemological: what even is a skill?
Is “communication” a skill? How about “problem solving”? “Data literacy”? “Strategic thinking”? Depending on which skills taxonomy you’re using – ESCO, O*NET, Burning Glass, Eightfold, Gloat, or whatever LLM someone just duct taped onto their HRIS this week – you’ll get wildly different definitions, examples, and categorizations.
Even for so-called “hard skills,” proficiency is a moving target. Just because someone knows Python doesn’t mean they can actually build scalable, secure software. And just because someone has “AI” on their LinkedIn doesn’t mean they aren’t prompting ChatGPT like a middle schooler trying to plagiarize a book report.
Skills are not binary. You’re not either “skilled” or “unskilled.” Proficiency exists on a spectrum, and that spectrum is usually defined by whoever’s shouting loudest in the calibration meeting.
Let’s say you could accurately define and measure skills today. Congratulations. You’ve now built a static system in a dynamic world.
According to the World Economic Forum, the average half-life of a skill is about 2.5 years. Which means by the time you finish building your skills matrix, 30% of it is already obsolete.
Take marketing as an example. In 2015, being good at SEO meant keyword stuffing and backlink farming. In 2020, it meant optimizing for Google’s EEAT framework. In 2025, it means understanding AI content detection algorithms and probably a sprinkle of TikTok virality for good measure.
Or look at cybersecurity. The skills needed to defend against ransomware today barely resemble those needed five years ago.
And next year? Entirely new threat vectors. Good luck building a competency model that keeps up with that. Hell, we still haven’t updated our good old fashioned job descriptions in a good decade and a half. This might be a good thing, since the sentiment is that traditional job related requirements are specious, biased and, most importantly, isn’t a “problem” for which VC backed software vendors have found a way to monetize a “solution.”
Unlike, say, skills based hiring.
No One Actually Knows What Skills They Have
Here’s a dirty little secret from inside the belly of the enterprise HR beast: most companies have no idea what skills their employees actually have.
Yes, they’ve all paid Gartner to run a “skills audit.” Yes, they’ve bought some shiny internal talent marketplace that gamifies self-reported skills endorsements. But most of those platforms rely on employees to self-report what they’re good at.
It’s like asking people to rate themselves on a dating app and expecting an objective inventory.
In fact, research from Deloitte shows that fewer than 1 in 5 companies have a robust, regularly updated skills database. Even fewer tie that data into workforce planning or hiring decisions in a meaningful way.
So what’s actually happening? Companies are creating complex ontologies of skills, mapping them to job architectures, and then… doing absolutely nothing with them.
Because when push comes to shove, hiring managers still default to “someone who’s done this exact job before at one of our competitors.” Proponents of skills-based hiring love to talk about how it will “future-proof” the workforce. If you know what skills you’ll need tomorrow, you can start developing or hiring for them today.
Cute. But totally delusional.
Let’s assume you’re a mid-size financial services firm. Can you tell me, right now, what technical and soft skills your business will need in Q3 of 2027?
Do you know how regulatory frameworks, AI and process automation, evolving customer preferences, geopolitical instability or global economics will shape your product roadmap and talent strategy?
Didn’t think so.
The Common Use Cases… and Why They Don’t Hold Up
Even when you can forecast a rough direction—say, “we’re moving to the cloud” or “we need more data governance”—you still have to translate that into roles, responsibilities, and skill proficiencies, none of which are standardized.
The idea that you can proactively align your workforce to future skill needs is appealing. It’s also a fantasy, unless you’re willing to update your skills taxonomy, job architecture, and training investments quarterly. Spoiler: you’re not.
1. Internal Mobility:
The theory: If you know what skills someone has, you can match them to roles or gigs internally.
The reality: Most companies don’t reward lateral moves. Most managers hoard talent. And most employees don’t trust that checking a box next to “machine learning” will result in career advancement.
You know what actually drives internal mobility? Politics, timing, and the manager’s gut instinct – not skill taxonomies (who, like their HRBPs., likely have no idea WTF this actually means). Trust me on this one.
2. Diversity & Inclusion:
The theory is simple: by removing degree and pedigree requirements and focusing on skills, you level the playing field. You know, in ways that systemic failures of the education system, legislation and class immobility haven’t. Enter HR, obviously.
The reality is a bit different: most “skills-first” systems are still trained on biased data (i.e., past hiring decisions), and most recruiters still prefer resumes that “look right.”
In fact, a 2022 MIT Sloan study found that algorithmic hiring tools worsened bias unless meticulously designed and audited. And, of course, there’s a growing body of longitudinal evidence that for the billions of dollars organizations have dedicated to improving DEIB outcomes, the impacts, if any, have been muted, at best.
3. Contingent Workforce Planning:
The theory: you can match freelancers and gig workers to precise project needs based on their skills.
The reality: Sure, if you’re Uber. But most organizations lack the infrastructure to integrate contingent skills data into core HR systems.
And good luck getting that independent contractor to fill out your Workday profile. Chances are, you don’t even know their rate card or what the actual hell they were hired on to do.
Maybe sweat the small stuff, first.
Can Skills Based Hiring Actually Work?
If “skills-based hiring” as it’s marketed today is largely smoke and mirrors, what should companies do instead?
Start simple. Focus on capabilities – broader clusters of behaviors, mindsets, and outputs – rather than trying to break everything into atomic units of skill.
Think “problem solving in ambiguous environments” vs. “proficient in Excel pivot tables.” Invest in talent intelligence platforms that actually ingest performance data, training history, and output, not just self-reported skills or third-party proxies.
And maybe, just maybe, stop pretending your workforce is a database. It’s a collection of messy, dynamic, fallible humans. Embrace it.
After all, that’s what the business of people is really all about.
Snark Attack 
Thank you for this article. We fully agree, that skills (as they are typically used today) are not a useful unit. And you also showed a better way forward: go for capabilities.
Now, however, the question is, what makes a capability? We think we have found the right approach – based on scientific research and especially suited for tech jobs.
If you are interested, have a look at our short article, which descibres the basic concept:
https://www.turn2talents.de/post/capabilities-vs-skills
Pingback: Swagger Like Us: Why the Future of TA Belongs To Talent Intelligence | Snark Attack