Skills vs Degrees, The Ongoing Shift

The most capable VP of Engineering I have placed did not have a computer science degree. She had built two products from scratch, led engineering teams through two acquisitions, and could explain any technical decision in plain language to a non-technical board. The company that hired her did not ask about her degree. They asked about her track record. The degree question had already become irrelevant, they just hadn't updated their job description to reflect it.

The argument that skills matter more than degrees is not new. It surfaces in every generation's conversation about education, hiring, and economic mobility. What is new is the pace at which specific hiring markets are actually acting on it, and the infrastructure that is making it possible to verify skills without institutional intermediaries.

A degree signals that you completed something. A skill signals that you can do something. These are different claims.

The degree has not been replaced. Its dominance as the primary signal in hiring is eroding, unevenly, in specific domains. 52% of US job postings on Indeed did not include any formal education requirement as of January 2024, up from 48% in 2019. But a Harvard Business School study found that fewer than 1 in 700 hires in 2023 actually benefited from degree requirement removal, the gap between stated policy and practice is real. Understanding where credentials are eroding and where they are not is now a practical question for anyone building a career or making hiring decisions.

Why the degree became dominant

The degree wasn't always the dominant hiring signal. It became dominant gradually, as the labour market for knowledge work expanded and employers needed scalable ways to filter large candidate pools without evaluating each candidate individually.

The degree from a recognisable institution solves several problems at once for an employer. It signals a minimum capability threshold, the candidate could complete a substantial multi-year program. It signals a minimum persistence and reliability threshold. It provides some institutional endorsement of character and ability. And it provides a standardised way to rank candidates relative to each other.

None of these signals is a direct measure of the ability to do the job in question. They are proxies. They became dominant because direct measurement was too expensive to do at scale.

Where credentials are losing ground

Software development. The field where the skills-vs-degrees shift has gone furthest. GitHub, open source contributions, portfolio projects, technical assessments, and competitive programming results provide direct evidence of coding capability. The ability to demonstrate what you can build has created a path around the credential for employers who are willing to invest in direct evaluation. The shift is not complete, a degree from a strong institution still helps at the most selective employers, but it has gone far enough that significant numbers of people are building careers in software development without traditional credentials.

Data work. Kaggle competitions, portfolio analyses, GitHub repositories, and tool proficiency assessments have created verification mechanisms that partially replace the credential for data science and analytics roles. The signal is noisier than in software, the range of what "data work" means is wider, but the credential is less determinative than it was a decade ago.

Digital marketing and content. Performance data, portfolio work, and demonstrated results are straightforwardly verifiable in ways that replace the credential for many roles. A social media account with evidence of growth, a portfolio of content with documented performance metrics, or a demonstrated record of campaign results speaks for itself. The credential is almost irrelevant in this domain.

Early-stage and startup hiring. Companies in their first few years of operation don't have the hiring volume that justifies building credential-based screening systems. They often evaluate people more directly. The founder who interviews a candidate in a two-hour conversation and asks them to complete a relevant task is doing direct skills assessment, not credential screening. The startup world has always been more credential-agnostic than large enterprise hiring.

Where credentials still dominate

Professional licensing. Medicine, law, accountancy, engineering certifications, financial regulation, these are domains where the credential is legally required for practice. AI tools and portfolio evidence do not substitute for credentials that are enforced by regulatory frameworks. The path here is through the credential, not around it.

Large organisation entry-level hiring at scale. When a company is screening 1,000 applications for 10 graduate roles, the degree is functioning as a filter because evaluating 1,000 people directly is not viable. The credential isn't being used because it's a perfect signal, it isn't, but because it makes the process tractable. Until direct skills assessment scales to match application volume, credentials will continue to dominate this context.

Network-intensive domains. In consulting, investment banking, venture capital, and professional services, the credential from a specific set of institutions is not just a capability signal, it is membership in a network. The alumni network of a target institution is part of the product. The skills-vs-degrees shift does not apply here in the same way, because the credential is buying access to a social institution, not just certifying a capability.

Senior executive hiring. Most executive searches still treat the credential as a baseline expectation. Not the most important factor, but a default filter that is rarely explicitly examined. The shift toward skills-first evaluation has happened fastest at the entry level and has been slowest at the executive level, where the verification problem is harder and where the cost of being wrong is higher.

The practical implication

The skills-vs-degrees shift is real and is accelerating in the domains where skills can be directly verified. The implication for anyone early in their career is that the paths to economic opportunity that bypass the credential are more accessible than they were ten years ago, but they require the ability to demonstrate capability directly, which requires having actually built the capability.

The implication for hiring organisations is that screening systems that rely exclusively on credentials are excluding capable people unnecessarily in domains where direct skills verification is now feasible. The organisations that have updated their screening to include direct skills assessment have access to larger candidate pools at the same or higher quality bar.

Neither the credential-maximalism of the past nor the credential-is-irrelevant position of the most aggressive skills advocates is right. The honest picture is: in specific domains, the credential matters less than it used to. In others, it is as important as it has always been. Knowing which is which is increasingly valuable.

See also: AI and Education, AI and Hiring: What Actually Changes, AI and the Future of Talent Sourcing

Sources

Indeed Hiring Lab: Educational Requirements in US Job Postings (2024)

Harvard Business School: Skills-Based Hiring: The Long Road from Pronouncements to Practice

I've spent my career placing senior leaders and building businesses from places, Kalahandi, Odisha, where the credential gap is most acute. The people I grew up with who had the most capability did not always have the most access to the institutions that certify it. The shifts described above matter more in those contexts than in any other. They matter because they determine whether the next generation of capable people from those places have to leave to access opportunity, or whether opportunity can come to them.