GBAF Logo
Global Banking & Finance Awards® 2026 Nominations open, free to enter Nominate now →
Skills Based Hiring Faces an Evidence Test - Trends news and analysis from Global Banking & Finance Review
Trends

Skills Based Hiring Faces an Evidence Test

Published by Barnali Pal Sinha

Posted on October 16, 2026

9 min read
Add as preferred source on Google

Removing a degree requirement from a vacancy can widen the invitation to apply. Whether it changes who gets hired depends on what happens afterwards. Screening rules, interview habits and managers' confidence in alternative evidence can preserve an old selection pattern even when the advertisement looks different.

That gap between stated policy and operating practice is central to the debate about skills-based hiring. Employers want better ways to identify people who can perform a role. Candidates want experience and capability to count when their route into work does not match a conventional educational pathway.

The opportunity is meaningful, but the implementation is demanding. An organisation that relies less on familiar credentials needs a credible way to describe the job, assess relevant skills and support a new employee's development. Without that work, one imperfect signal may simply be replaced by another.

Define the work before changing the filter

The OECD's 2025 report on a skills-first approach examines skills signalling, hiring practices and the opportunities and barriers associated with broader adoption. It frames demonstrated capability as important to hiring and workforce development, while recognising that implementation depends on context.

For an employer, the starting point is a specific role. A job description that combines an entire department's responsibilities will be difficult to assess fairly, regardless of whether it mentions a degree. The business should identify what a new employee needs on arrival and what can reasonably be learned after joining.

An illustrative customer-operations role might require accurate written communication, sound escalation judgment and the ability to interpret an account record. Familiarity with one internal software package may be teachable. Treating all four as equally necessary before entry could exclude suitable candidates without improving the decision.

This does not make formal qualifications irrelevant. Some occupations have licensing or qualification requirements, while particular education can provide relevant evidence for other roles. A skills-based approach should examine the connection between the requirement and the work, rather than adopt a blanket rule for or against degrees.

Policy announcements do not establish outcomes

The 2024 research from the Burning Glass Institute and Harvard Business School examines the gap between removing degree requirements and changing actual hiring. Its central relevance is the need to measure outcomes, rather than assume an amended advertisement creates access on its own.

An employer may attract more applicants without changing interview selection. A recruiter may remove a degree field while continuing to prioritise familiar employers. A manager may formally accept equivalent experience but remain uncertain about how to evaluate it. Each creates a different point at which policy can lose effect.

The business should therefore examine the path from application to appointment. If candidates with alternative backgrounds leave the process at one stage, that stage deserves investigation. The explanation could involve a relevant skills gap, an inaccessible assessment or an informal preference that has little connection to performance.

These distinctions matter for public claims. A company should distinguish the number of vacancies with revised requirements from the number of people hired through the changed process. Neither alone demonstrates better retention, fairer opportunity or stronger performance after appointment.

Evidence should resemble the job

A work sample can give candidates an opportunity to demonstrate relevant ability. Its usefulness depends on what it measures and how consistently it is evaluated. An elaborate assessment can still be poorly matched to the role.

For an illustrative analysis position, a short exercise might ask candidates to interpret a small dataset and explain an uncertainty. If the actual job requires explaining decisions to colleagues, the assessment should consider that explanation as well as the numerical result. Testing an unrelated puzzle may reveal less about the work.

The US Equal Employment Opportunity Commission's guidance on employment tests and selection procedures explains that selection tools can raise discrimination concerns and that relevant US legal requirements apply. International employers need to consider the rules in each jurisdiction rather than treat US guidance as a universal legal framework.

The practical management issue is to justify the assessment. The employer should be able to explain why the task is relevant, how responses are scored and how accessibility needs are addressed. Assessors also need a shared understanding of what good evidence looks like, including more than one acceptable way to solve a problem.

Candidate effort is part of the design

A lengthy unpaid exercise can impose a cost on applicants, particularly those balancing employment, caregiving or limited access to equipment. A process intended to widen access should examine those burdens alongside the information it gains.

An organisation can consider shorter tasks, appropriate alternatives and clear instructions about the time expected. It should avoid using live commercial work in a way that leaves applicants uncertain about whether the exercise is an assessment or unpaid production.

Transparency can also improve the relevance of the evidence. Candidates should know which capabilities are being examined and whether external tools are permitted. A hidden expectation that applicants guess the employer's preferred method can reward familiarity with recruitment conventions rather than the skill itself.

This is not an argument for making every assessment easy. It is a reason to make the difficulty purposeful. If a demanding task is essential to the job, the employer should explain that connection and evaluate it consistently.

Skills signals differ in quality

The OECD's practical discussion of skills-first approaches considers barriers and implementation choices, including the challenge of recognising skills acquired through different routes. The existence of a badge or certificate does not by itself settle how much weight an employer should place on it.

A credential may demonstrate attendance, completion of a task or performance under independently assessed conditions. Those are different signals. Recruiters need enough information to understand what was examined, how recently and in what circumstances.

Experience also requires interpretation. A familiar job title may cover different responsibilities in different organisations. Candidates should have a reasonable opportunity to explain what they actually did, while employers should avoid assuming that a larger brand automatically implies stronger capability.

For applicants, the burden of proof should remain manageable. Requiring extensive portfolios, multiple certificates and a lengthy test can create a new form of credential accumulation. A better process seeks sufficient evidence for the decision and acknowledges what will still need to be learned at work.

AI changes both applications and assessment

Generative AI can help candidates prepare applications and employers organise information. It can also make a polished submission a less reliable indication of unaided writing or analysis. Employers need to decide which capabilities they are trying to observe before responding with more elaborate screening.

If the role allows staff to use AI tools, an assessment might examine whether the candidate checks an output, identifies unsupported claims and explains a decision. If a task requires independent capability, the employer needs an appropriate method for observing it. The policy should be clear to candidates and assessors.

The NIST AI Risk Management Framework offers a voluntary framework for managing AI risks across development and use. For recruitment systems, the relevant analytical implication is to examine data, context and outcomes rather than assume automated scoring is objective.

Human involvement alone does not establish fairness. A reviewer who routinely accepts an automated recommendation may add little independent judgment. Employers should understand what the system evaluates, what information is missing and how a candidate can raise a concern through an appropriate process.

Hiring must connect with development

A skills-based appointment can fail if the employer expects a new recruit to arrive with knowledge the selection process treated as teachable. Managers need a clear account of the capability demonstrated and the learning still required.

An onboarding plan should reflect that distinction. A recruit with strong customer judgment may need training in the firm's systems and procedures. Another may know the systems but need help handling difficult conversations. Assigning the same programme to both can miss the reason each was selected.

Internal mobility deserves similar attention. Employees may have relevant experience that is poorly represented by their current title. A fair opportunity to demonstrate it can help the organisation fill roles while retaining knowledge. However, internal assessments should have clear standards rather than rely exclusively on a manager's sponsorship.

Development opportunities also need to remain accessible after hiring. If advancement still depends on an educational signal the firm stopped requiring at entry, the policy may widen the first door while leaving later progression unchanged.

Measure performance with context

The test of a recruitment process extends beyond the offer. Employers can examine whether hires achieve relevant standards, how much support they require and whether they remain and progress. Comparisons should account for differences in roles, managers and training opportunities.

A weak outcome should lead to a specific investigation. It may reflect a poor assessment, inadequate onboarding or an unrealistic workload. Attributing every difficulty to a candidate's educational background would reproduce the assumption the revised process was intended to examine.

Likewise, an encouraging early result should not become a universal claim. A method that works for a defined operations role may require substantial adaptation for a specialist position. Employers can build confidence through bounded use, review and revision.

Skills-based hiring has a stronger foundation when it changes the evidence used in decisions and the support provided afterwards. The most useful question is whether an employer can show how the process identifies capability relevant to the work. That evidence, rather than the language of a recruitment campaign, determines whether the approach creates a meaningful change.

References

OECD — Empowering the Workforce in the Context of a Skills First Approach — 2025

https://www.oecd.org/en/publications/empowering-the-workforce-in-the-context-of-a-skills-first-approach_345b6528-en.html

Burning Glass Institute and Harvard Business School — Skills Based Hiring The Long Road from Pronouncements to Practice — 2024

https://www.burningglassinstitute.org/research/skills-based-hiring-the-long-road-from-pronouncements-to-practice

US Equal Employment Opportunity Commission — Employment Tests and Selection Procedures

https://www.eeoc.gov/laws/guidance/employment-tests-and-selection-procedures

OECD — Practical considerations for a skills first approach — 2025

https://www.oecd.org/en/publications/empowering-the-workforce-in-the-context-of-a-skills-first-approach_345b6528-en/full-report/practical-considerations-for-a-skills-first-approach_88eb89d7.html

NIST — AI Risk Management Framework

https://www.nist.gov/itl/ai-risk-management-framework

Related Articles

More from Trends

Explore more articles in the Trends category