The Skills-Based Hiring Numbers Look Great. The Hiring Doesn't Match Them.

By NiceHire Team

Ask a talent-acquisition leader in September 2026 whether skills-based hiring has arrived, and the numbers say yes, decisively. The National Association of Colleges and Employers' Job Outlook 2026 survey found that 70% of employers now use skills-based hiring for entry-level roles, up from 65% the year before, and 71% of those employers say they apply it at least half the time. GPA screening — once a near-default filter — has fallen from 73% of employers in 2019 to just 42% today (NACE). Skills-based evaluation now shows up most heavily at the interviewing (87%) and screening (65%) stages of the process.

That is a genuinely large five-year shift in stated employer behavior, corroborated across NACE's own press coverage and third-party summaries of the same survey (Learn & Work Ecosystem Library; NACE press release). But "employers say they use skills-based hiring" and "skills-based hiring is what's actually deciding who gets hired" are two different claims, and the gap between them is the more important story — and one the adoption headlines tend to bury.

What "in name only" actually looked like, measured

The clearest evidence that a stated policy shift and an actual hiring shift can diverge comes from a 2024 study by Harvard Business School's Project on Managing the Future of Work and the Burning Glass Institute, which tracked roughly a decade of real hiring outcomes at companies that had publicly announced they were dropping degree requirements (Harvard Kennedy School / Managing the Future of Work; Burning Glass Institute). The study's headline finding is blunt: 45% of companies that announced dropping degree requirements made the change in name only — the policy existed, but actual hiring patterns never moved. A further 18% were "backsliders": they made the same public pledge and then went on to hire a higher share of degree holders than before. Only about 37% of the companies studied — the ones the researchers classified as genuine skills-based hiring leaders — drove essentially all of the real change, meaningfully increasing their share of non-degree hires. Across the full sample, the study found that fewer than 1 in 700 hires actually went to a candidate who benefited from the dropped requirement.

That study predates the 2026 NACE numbers by two years, which matters for how you read it: it isn't a rebuttal of this year's survey, it's the documented base rate for exactly the kind of policy-announcement gap that a "70% of employers" headline is vulnerable to reproducing. Nothing about NACE's methodology — a survey asking employers what they report doing — can distinguish a company in the 37% "leaders" group from a company in the 45% "in name only" group. Both would answer "yes, we use skills-based hiring" on the same survey. The self-report is real (employers really are saying this), but self-report was exactly the layer the Burning Glass Institute's work showed cannot be trusted to predict what the hiring data will show.

The disconnect compounds on the candidate side too

NACE's own 2026 release doesn't just report the adoption number — it pairs it with a second finding that's easy to miss: college students, by NACE's own account, largely don't understand what skills-based hiring means or how to prepare for it (NACE). That's a second, independent source of gap on top of the employer-side "in name only" problem. Even where an employer's shift to skills-based evaluation is real, candidates are still largely building resumes and application materials optimized for a degree-and-pedigree filter that, on the employer's side, may no longer be the actual gate. The two gaps compound: a genuine screening change on one side of the table meets an unadapted application on the other, and the mismatch reads to both parties as evidence that "skills-based hiring" is marketing language rather than a real mechanism — for different, and equally valid, reasons on each side.

Why the gap is worth closing, not just noting

It would be easy to read all of the above as a reason to shrug at skills-based hiring as another item on the pile of HR buzzwords that didn't survive contact with practice. The economic case against that shrug is specific. McKinsey's research on skills-based talent strategies finds that hiring for demonstrated skills is roughly five times more predictive of on-the-job performance than hiring for educational credentials, and more than twice as predictive as hiring for prior work experience (McKinsey & Company). The same body of research reports that non-degree workers hired into roles that formerly required a degree stay in those roles substantially longer — on the order of a third longer tenure — than degree-holding hires in equivalent roles. If those predictive-validity numbers are even directionally right, the 45%-in-name-only pattern isn't a harmless communications gap. It's employers leaving a measurable performance and retention advantage on the table while telling themselves, on the strength of a policy announcement, that they've already captured it.

Why the gap keeps happening: the mechanism doesn't change with the policy

The Burning Glass Institute's own framing of why the "leaders" group succeeded where the "in name only" group didn't points at something structural, not just cultural: the companies that actually changed outcomes typically paired the degree-requirement removal with a change in how candidates are actually evaluated — new assessments, restructured screening criteria, or retrained recruiters — rather than simply deleting a line from a job posting and leaving the rest of the funnel untouched. Removing a stated requirement without replacing the underlying evaluation mechanism leaves recruiters doing the same pattern-matching they always did, on the same signals (a familiar university name, a familiar-looking resume format) that a degree requirement was a proxy for in the first place. The policy changes; the actual filter, run by the same humans looking for the same signals, often doesn't.

This is the specific place we think a structured, criteria-based screening step earns its keep — not as a claim about the fairness of the outcome (a determination no vendor can credibly self-certify), but as a narrower, mechanical point about consistency. On NiceHire, every candidate who reaches AI screening is scored against the same fixed set of criteria — technical, communication and cultural fit, each on a 0–100 scale — evaluated against a configurable passing threshold that an employer sets once, in advance, for the role. That's a statement about mechanism, not outcome: it doesn't assert an accuracy rate, and it isn't a bias-elimination claim. What it does do is remove the specific failure mode the Burning Glass Institute's research describes — a recruiter reverting to informal, unstated pedigree signals once the formal requirement is gone — because the scoring criteria for the role are fixed before any candidate's resume or interview is seen, not decided case-by-case as materials come in. A policy that says "we no longer require a degree" changes nothing about how the next candidate gets evaluated unless the evaluation step itself is redefined at the same time; that's the redefinition NACE's stage-level data (skills-based evaluation concentrated in screening and interviewing) is measuring, and it's also the layer where "in name only" companies visibly stopped short.

What this means if you're the one deciding

Three things follow directly from the data above, without needing to take our word for any of it:

1. Don't trust your own adoption self-report. If your organization can answer "yes" to "do you use skills-based hiring," ask the harder follow-up question the Burning Glass Institute's methodology actually asked: pull your last twelve months of hires into roles where you dropped a degree requirement, and check what share of those hires were candidates who wouldn't have cleared the old requirement. If that number is close to zero, the policy hasn't reached the funnel yet, whatever the survey answer says.
2. Change the evaluation step, not just the posting. Removing a requirement from a job description is necessary but, per the 45%-in-name-only finding, nowhere near sufficient. The leaders group's actual lever was new or restructured assessment criteria applied consistently — the same shift NACE's data shows concentrated in screening and interviewing, not job-ad copy.
3. Tell candidates what actually changed. NACE's own students-don't-understand-it finding means employers who have genuinely shifted their evaluation are still losing well-matched non-degree candidates who self-select out because they're optimizing for a filter that no longer applies. Making the actual criteria visible — what's being assessed and how — closes a gap that exists independently of whether your hiring process itself needs fixing.

The 70% adoption figure is real and worth taking seriously as a direction of travel. The 45%-in-name-only figure is also real, and it's the more reliable predictor of what a given company's number actually means until it's checked against that company's own hiring data. Both numbers can be true about the industry at once; the only way to know which one describes your own hiring is to look at the hires, not the policy page.