In 2021, Harvard Business School partnered with Accenture to study a question that the hiring industry had been avoiding: what happens to qualified candidates who never make it through automated screening?

The findings were stark. The research team estimated that more than 27 million people in the United States were hidden workers in 2020, a substantial portion of the working-age population (Joseph B. Fuller, Harvard Business School Working Knowledge, November 5, 2021). Harvard called them "hidden workers."

27 Million+
people in the United States estimated to be "hidden workers" in 2020 (Joseph B. Fuller, Harvard Business School Working Knowledge, November 5, 2021)

The study surveyed 8,720 hidden workers and 2,275 executives across the U.S., U.K., and Germany. The results pointed to hiring processes that screen out many people who may be able to do the work.

The Numbers That Should Concern Every Hiring Manager

The Harvard study did not describe a minor inefficiency. It described a structural problem in how companies identify talent.

88% of employers agreed that qualified high-skills candidates are vetted out of the process because they do not match the exact criteria established by the job description. The figure was 94% for middle-skills candidates (Hidden Workers report, as reported by The Register, September 8, 2021). In other words, the screening standard is a match to a written description, not a demonstrated inability to do the job.

Joseph B. Fuller, Harvard Business School: "In filtering out all but those candidates who match exacting and expanding requirements, companies reject many experienced, reliable, and creative individuals."

That observation deserves attention from any CHRO who reports a talent shortage. Rigid screening criteria can shrink the pool of candidates a company is willing to consider.

Who Gets Filtered Out

The hidden workers population is not random. Groups whose career histories do not follow the linear path that keyword-matching systems were built to evaluate are especially exposed.

Veterans. Military experience produces skills that translate directly to civilian roles, but the terminology does not. "Squad leader" maps to "team supervisor." "Maintained operational readiness" translates to "ensured equipment uptime." Many keyword-based systems will not make these connections. Fuller's article names veterans among the people behind the hidden-worker statistics, together with caregivers, older workers, and previously incarcerated people (Harvard Business School Working Knowledge, November 5, 2021).

Career changers. In a hypothetical example, a project manager from construction who wants to move into tech has transferable skills in budgeting, scheduling, stakeholder management, and team leadership. But their resume uses construction vocabulary, not tech vocabulary. The skills are there. The keywords are not. In a May 2020 LiveCareer survey of recently unemployed individuals, reported by HR Dive on June 1, 2020, 57% were unable to identify their transferable skills and 58% were unsure how to communicate transferable skills on their resume.

Candidates with employment gaps. The Hidden Workers report found that almost half of the companies surveyed weeded out resumes that show a work gap. Systems that penalize gaps automatically, or ignore them entirely, do not evaluate whether the candidate can do the job.

International candidates. Equivalent qualifications described differently, degree titles that do not match U.S. conventions, and professional certifications that use different naming. A "chartered engineer" and a "licensed professional engineer" may have identical qualifications. A keyword filter may not make that connection.

The Real Cost of Filtering Out Good Candidates

The cost is not abstract. It can hit the budget in at least two ways.

Bad hire costs. When companies settle for whoever makes it through the filter rather than the most qualified candidate, the downstream cost can be significant. 2016 CareerBuilder research put the average cost of hiring the wrong employee at $17,000, as cited by ASE.

Pipeline degradation. The more qualified candidates you filter out, the worse your remaining pool becomes. You end up interviewing a curated set of people who are good at writing resumes, not necessarily good at the job. Greenhouse reports that applications per job rose from 223 in 2024 to 244 in 2025. The flood of applications makes efficient screening essential, but efficient screening that removes qualified candidates is not efficient. It is just fast.

$17,000
average cost of hiring the wrong employee, according to 2016 CareerBuilder research cited by ASE (September 6, 2022)

Why This Is Not Just a Technology Problem

A November 2025 Enhancv interview study of 25 U.S.-based recruiters found that 23 (92%) said their systems do not auto-reject resumes for formatting, content, or design. The sample is small, but it is a reminder that the system is only as good as the criteria humans set.

That matters because the problem is not that ATS exists. The problem is that the criteria fed into ATS systems reward vocabulary over capability. When a job description lists 15 required skills and the ATS scores candidates by how many of those exact phrases appear in the resume, the result is a ranking based on language matching, not competency matching.

What the Hidden Workers Study Proved About the Solution

The Harvard study did not just identify the problem. It also looked at what happens when companies hire hidden workers.

According to Fuller, firms that hire formerly hidden workers rated their performance "better or significantly better" in terms of attitude and work ethic, productivity, quality of work, employee engagement, attendance, and innovation. They also saw lower rates of voluntary turnover (Harvard Business School Working Knowledge, November 5, 2021).

The evidence points to the screening criteria, not competency, as the barrier.

What Can Be Done

For job seekers: Understand that many systems rely heavily on keyword matching rather than evaluating competency. Before you submit a resume, compare your language to the job description. If you have the skill but used different words, the system will not make the connection for you. Scanning your resume against the job description before applying identifies the gaps the algorithm will see, so you can close them before you submit.

For employers: Audit your screening criteria. If your ATS is configured to auto-reject based on keyword density, you may be losing qualified candidates. Consider whether your job descriptions contain inflated requirements that filter out candidates who could do the job with minimal ramp-up. Fuller's article criticizes applicant tracking filters that exclude candidates without college degrees, a requirement worth checking against what the job really needs.

For both sides: The goal should be matching skills to requirements, not matching vocabulary to vocabulary. Tools that evaluate transferable skills, map experience across career paths, and assess genuine capability rather than keyword presence are the path forward. That is exactly what TrueScan HR was built to do.

Go Deeper: The Filters Inside Your ATS

Each of these guides takes one common applicant tracking system filter and shows what it screens out and how to fix it.

Free download: the one-page California AI Hiring Compliance Checklist, with the regulation section behind every item. Get the checklist

Sources:

Thabiti Adams is a CISSP and CCSP certified cybersecurity professional and founder of Adams Cloud & Cybersecurity.