The Resume Is a 500-Year-Old Technology

I. The logic

The logic is broken

The resume was never designed to find the best talent. It was designed to process volume. Those are not the same objective — and confusing the two is what's costing companies millions and costing talented people careers.

A corporate job posting attracts an average of 250 applications, according to Glassdoor. In tech, that number goes into the thousands. No team can evaluate that volume properly. So we handed it to software.

The software doesn't find the best candidate. It eliminates enough volume so a human can handle the rest.

And it's getting slower. Despite more automation than at any point in history, average global time-to-hire hit an all-time high of 44 days, according to The Josh Bersin Company and AMS. More software, more steps automated, slower outcome.

We are running a 500-year-old sorting mechanism at a scale it was never built to support.

II. Signal

What the system cannot read

What the system cannot read is the thing that actually matters.

The engineer who spent six months untangling a legacy architecture before it took down the entire platform. On a resume: "Maintained AWS cloud infrastructure."

The director who inherited a team running 40% attrition and brought it to 95% retention in a year. On a resume: "Cross-functional team leadership."

The architect who built something so structurally sound it ran seven years without a single production incident. On a resume: invisible — because preventing disasters leaves no dramatic bullet points.

Context. Recovery. The ability to work under real pressure. None of it fits in a text box.

Instead, the system ranks by keyword proximity. If your resume doesn't contain the exact strings from the job description, you don't get rejected — you just sink. And given that recruiters spend an average of 7.4 seconds on an initial CV scan (TheLadders eye-tracking study), sinking is the same as not existing.

III. The loop

It selects for the best resume writer

So candidates spend days optimising documents for machines. The loop looks like this:

Generic job description → keyword-stuffed resume → algorithmic ranking → 7.4-second human glance.

A person who knows how to game the parser outranks someone who simply wrote an honest account of their work. The system doesn't select for the best worker. It selects for the best resume writer.

IV. Leaders

88% of executives know this

A Harvard Business School and Accenture study found that 88% of company leaders acknowledge their ATS filters out highly qualified candidates because those people don't match the pre-set criteria. Nearly nine in ten. Most continue using the same systems anyway — because the alternative, processing raw volume manually, is untenable.

So they keep a broken system because fixing it feels worse. That's where we are.

V. Evidence

What a résumé actually predicts

Schmidt and Hunter's landmark meta-analysis of personnel selection data is worth sitting with. The two things a resume is built on — experience and credentials — are among the weakest predictors of how someone will actually perform.

• Work samples and structured interviewsr = 0.51
• Years of job experiencer = 0.18
• Education levelr = 0.10

When hiring goes wrong on that basis, it's expensive. The U.S. Department of Labor estimates a bad hire costs a minimum of 30% of first-year earnings. Topgrading research puts the all-in cost for executive mis-hires at five to 27 times annual salary.

We are routinely risking seven-figure decisions on a framework that correlates less than 20% with workplace success.

Adding more AI to this pipeline doesn't fix it. Summarising a flawed summary faster doesn't produce truth — it produces noise at scale.

The industry doesn't need a smarter scanner. It needs a different model.

VI. Direction

Why are we still scanning them at all?

The direction is already visible. Organisations moving toward skills-based hiring — demonstrating capability rather than declaring it — are 57% more likely to anticipate change effectively and 98% more likely to retain high performers, according to Deloitte. The World Economic Forum's Future of Jobs Report identifies skills-based strategies as one of the defining shifts in how competitive workforces will be built through 2030.

But a philosophy without infrastructure is just a philosophy. The missing piece is a way to represent professional identity that is dynamic, evidence-based, and owned by the person who built it — not scattered across a hundred ATS databases that monetise your activity while telling you nothing back.

Leonardo da Vinci's bulleted letter was a genuine innovation. In 1482.

The question confronting every talent organisation in 2026 is not "how do we scan resumes faster?"

It's: why are we still scanning them at all?

VII. Appendix

Sources