The Knockout Line Standard, Each Requirement Graded Gate, Screen, or Wish

This standard is for anyone holding a job posting with a long list of requirements who wants to know which ones will actually stop the application before spending an hour tailoring. It gives every requirement line one of three grades (Gate, Screen, or Wish) using written tests two people would apply the same way. Then it turns those grades into an apply or skip verdict you can defend.

What do Gate, Screen, and Wish mean?

A Gate is a requirement that ends your application no matter who reads it. A Screen is one that a person or a soft filter will count against you. A Wish rarely decides anything. The grade comes from evidence outside the posting, not from the header the line sits under.

Some terms used throughout:

• Screening question (also called a screener or knockout question): a question the job platform asks during the application. The employer can use your answer to sort or reject you.
• "Or equivalent" clause: wording such as "bachelor's degree or equivalent experience". It tells you the employer will accept a substitute.
• Peer frequency: how often the same requirement appears in other postings for the same title.
• Incumbents: the people who already hold this title at this employer.
• Applicant tracking system (ATS): the employer's own hiring software. It can have filtering rules of its own, separate from the job board's.

Each grade proves something specific, and each can mislead you:

• Gate proves that someone outside the employer, usually a regulator, requires the item. It misleads you when you grade a line as a Gate just because the posting calls it "required".
• Screen proves that the line carries weight in this hiring process. It misleads you when a screening question you read as decisive was only set to flag you, not reject you.
• Wish proves that the line is ornamental for this title. It misleads you when the posting says nothing about something, like a degree, that the employer still hires on in practice.

Why is a match percentage the wrong test?

The common advice to apply once you meet 60% to 70% of the requirements describes when people choose to apply. It comes from an internal report that was never published, and it says nothing about which lines an employer enforces. A percentage also counts a state license and a nice-to-have tool as equal units, which no recruiter does.

The figure goes back to Tara Mohr's HBR piece on why women don't apply unless they are 100% qualified. She wrote that the finding comes from a Hewlett Packard internal report, and it has since been repeated in Lean In, The Confidence Code, and dozens of articles. That report has never been published. Mohr then ran her own survey of more than 1,000 people. The most common reason for not applying was: "I didn't think they would hire me since I didn't meet the qualifications and I didn't want to waste my time and energy." In the same survey, 15% of women said the main reason they didn't apply was "I was following the guidelines", compared with 8% of men.

So the people who hold back are guessing which lines will be enforced. A percentage doesn't fix that guess. It just puts a number on it.

The employer side points the same way. The Hidden Workers study by Fuller and Raman surveyed 8,720 hidden workers and 2,275 executives across the U.S., U.K., and Germany. It found:

• 88% of executives agreed that qualified high-skill candidates are filtered out because they don't match the exact criteria in the job description.
• 94% said the same about middle-skill roles.
• The study puts the number of hidden workers at as many as 27 million job candidates.

A Robert Half survey of more than 300 HR managers found that 42 percent of the resumes they receive, on average, come from candidates who don't meet the job requirements. Yet 84 percent said their company is open to hiring someone whose skills can be developed through training. No public source splits how recruiters treat a missing required item versus a missing preferred one, and none shows how many listed requirements successful hires actually meet. That gap is why this standard grades each line against outside evidence instead of guessing an average.

A match percentage counts a state license and a nice-to-have tool as equal units. No recruiter does.

Which requirements can auto-reject you?

Only a requirement the employer has turned into a required screening question, or written into its own ATS rules, can reject you before a person looks. Everything else reaches a human, who can choose to overlook it. Which items can be made required depends on the platform, and that limit decides how high a line can grade.

What the platforms' own help pages establish:

• Indeed: most pre-made screener questions can be marked required or preferred. Custom questions can't be marked required. If you fail a required question, you are moved to the employer's Rejected list. The employer can still review that list and move you back. If you fail a preferred question, you stay in the applicant pool with a note that you didn't meet that preference.
• LinkedIn: question categories include background check, education, location, language, certifications, and work experience, plus custom questions. Employers make a question required by ticking the Must-have qualification box. They can also choose to archive candidates who fail and send an automatic rejection email. LinkedIn's help page also says that screening questions are only available when collecting applications by email, and that LinkedIn does not prevent unqualified applicants from applying.
• Greenhouse and Oracle Taleo: a secondary source quotes their help pages describing rejection based on answers. I did not check those primary pages, so treat that row as unverified.

One structural consequence matters most. On Indeed, an unusual requirement the employer writes as a custom question can never auto-reject you. The most a custom question can do there is mark you as a Screen.

The limit you can't get around: as a candidate, you can't see whether a question was set to required or preferred. The standard's rule for that is in the ladder below.

What is the grading ladder?

Run each requirement down six rungs in order and stop at the first one it matches. That rung is its grade. Because the order is fixed and every rung is a yes or no test, two people grading the same posting should land on the same grades.

The six rungs, in order:

1. Legal source. A law or license requires the item and you can cite the source. Work authorization and a statutory minimum age count automatically. Grade: Gate.
2. Screener question. The line appears as a screening question in the application. Grade: Screen. (A yes/no question about a legal item has already stopped at rung 1.)
3. Incumbents hold it. In a sample of 10 people holding this title at this employer, 6 or more show the item. Grade: Screen, even if the posting doesn't mention it.
4. Softened wording. The posting says "preferred", "a plus", "nice to have", or "or equivalent". Grade: Wish.
5. Rare in peers. The line appears in fewer than 5 of 20 peer postings. Grade: Wish.
6. Otherwise. Grade: Screen.

The numbers at rungs 3 and 5 (10 incumbents, 6 of 10, 20 postings, 5 of 20) are this standard's own convention. No public source sets a minimum sample, a recency window, or a frequency cutoff for this kind of check. The closest published precedent gives a method but no threshold. Indeed's Hiring Lab counted postings that mentioned one or more degree or training requirements as of a fixed date, and limited its metro comparisons to areas with at least 10,000 postings. If your team adopts this standard, adopt the conventions as written so grades stay comparable, and change them only all at once.

How do you grade a posting, step by step?

Seven steps take a posting from raw text to a verdict. The first time you do this for a title it takes roughly one to two hours. Peer counts carry over to later postings for the same title in the same window, so later postings go faster.

Notes on the steps that hide the most judgement:

• Step 2. CareerOneStop's License Finder is sponsored by the US Department of Labor. You can search any state by occupation or job title, license name, or licensing agency, and a nationwide search lists 11,749 licenses. Its own help page warns that some states do not take part in the data collection. A missing result is not proof that no license is needed. Confirm with the state board.
• Step 3. Starting an application to read the questions costs nothing if you stop before submitting. If the Apply button sends you to the employer's own site, read the questions there. Rules on the employer's ATS replace whatever the job board would have done.
• Step 5. Deduplicate before you count. A single requisition reposted under three titles counts once.
• Step 6. This is the slowest step by hand: open profiles one at a time, filter by employer, and note each person's degree or credential. Refolk runs the same check against its index of professional profiles, so you can see what incumbents hold before you decide whether a silent or softened line is really a Screen.

Use this worksheet so the grade and its evidence stay together:

Posting: [title] at [employer], [country], read on [date]
Peer window: 20 postings, [title], [country], posted [start date] to [end date]

# | Requirement line | Posting label | Legal source (rung 1) | Screener Q? (rung 2) | Incumbents x/10 (rung 3) | Softened? (rung 4) | Peers x/20 (rung 5) | Grade | I meet it?
1 |  |  |  |  |  |  |  |  |
2 |  |  |  |  |  |  |  |  |
3 |  |  |  |  |  |  |  |  |

Gates missed: [count] -> Verdict: [Apply / Skip]
Screens missed: [list] -> Where each is addressed: [resume line / cover letter / screener answer]

How should you read degree and credential lines?

Degree and credential lines are where posting text and actual hiring drift furthest apart. They need the incumbent check (rung 3) and the peer count (rung 5) more than any other line. A posting that says nothing about a degree is weak evidence either way, and a credential line means different things in different markets.

Degrees: the posting changes faster than the hiring

The first three rows come from Indeed Hiring Lab. The last two come from the Burning Glass Institute and Harvard Business School study by Sigelman, Fuller, and Martin, which analyzed the hiring practices of 11,300 large firms. That study found the shift away from degrees amounted to just 97,000 new skills-based job opportunities out of 77 million annual hires. One published analysis names Bank of America, Amazon, and Oracle among firms that changed their postings without changing who they hired.

The study also gives a reason to apply anyway. At the firms it calls Skills-Based Hiring Leaders, workers without degrees have a retention rate 10 percentage points higher than their colleagues with degrees. Some employers mean it. The incumbent check tells you which kind of employer you're looking at.

So a degree line runs the ladder like this. If the posting says nothing and 6 or more of 10 incumbents hold a degree, it is a Screen (rung 3). If the posting says "or equivalent" and incumbents are mixed, it is a Wish (rung 4). It is only ever a Gate where a license that requires the degree puts it at rung 1.

Credentials: how common they are depends on the market

All columns come from Refolk's index of professional profiles. The shares and ratios are derived. PMP is the Project Management Professional certification.

The inference, and it is only an inference: a "PMP required" line in a UK posting is more likely to be one employer's outlier ask than the same line in a US posting. Refolk's index measures who displays the credential, not how often postings ask for it, so the peer count at rung 5 still decides the grade. Among UK project managers with PMP in their headline, the employers that came up most in a 25-profile sample were Shell (3 of 25) and Accenture (2). That's small enough to be a lead, not a pattern. If you're grading a UK posting from one of those employers, rung 3 is the check that matters.

How does this standard go wrong?

Most grading errors come from trusting the posting's labels, misreading what a platform can enforce, or counting the wrong thing. Each failure below has a false positive (the grade looks right but isn't) and a fix built into the procedure.

1. Grading the "Required" header as a Gate. This is the most common false positive. Recruiters often look past a skill listed as required, and 84% of HR managers in the Robert Half survey said they were open to training. Fix: only rung 1 produces a Gate. Check whether the line appears as a screening question before treating it as decisive.
2. Not being able to see the screener setting. A question set to preferred only flags you, and you can't tell it apart from a required one. Fix: treat yes/no questions on legal items as Gates and every other screening question as a Screen. The honest limit: a non-legal question the employer set to required can still move you to Indeed's Rejected list. That list is reversible by the employer, but you can't control that.
3. Off-site applications. LinkedIn screening questions only apply when applications are collected by email. When the Apply button redirects, the employer's own ATS rules apply instead. Fix: if the button redirects, re-run step 3 on the employer's site.
4. Degree removed in name only. The posting says nothing about degrees, so the line seems not to exist, yet as many as 45% of firms that removed it kept hiring graduates. Fix: step 6, the incumbent check, feeding rung 3.
5. Incomplete license data. CareerOneStop states that some states do not submit license data, so "no license found" can be wrong. Fix: confirm every license question with the state board before grading it anything other than a Gate.
6. Headline counts understate credentials. The 0.92% figure measures who shows PMP in a headline, not who holds it. Treat it as a lower bound. The title match is also loose: the 25-profile sample included CEOs and founders. Fix: use headline shares to compare markets, never as a holding rate.
7. Peer sample errors. Title variants, reposted duplicates, and stale postings can make a niche requirement look more common than it is. Fix: dedupe by requisition, fix the date window before counting, and keep the title exact.
8. Dishonest screener answers. Answering yes to a license you don't hold gets you past the filter and fails at verification. No public source measures how often this happens, but the grade doesn't change: a Gate you don't clear is a skip.

A note on where this standard disagrees with the advice that ranks today. The pages that rank for "should I apply if I don't meet all the requirements" start from a match percentage, and some suggest asking a recruiter first. None of them puts a legal or screener check first. Asking a recruiter is fine as a tiebreaker on a Screen. It isn't a substitute for rung 1, because a recruiter's reassurance doesn't change what a licensing board requires.

How do grades become an apply or skip verdict?

Apply if you clear every Gate. Skip if you miss even one. The Screens you miss decide how you tailor, not whether you apply, and Wishes take no tailoring time at all.

Why this rule and not a threshold on Screens: no public data shows how many missing Screens a recruiter will tolerate. This standard won't invent a cutoff. What it gives you instead is a list of the specific Screens you miss, which is a better input to a time decision than a percentage. If you miss many Screens, the application will take more tailoring. Whether that's worth it is your call, but now you're making it with the facts in front of you.

Addressing a missing Screen means putting the nearest true evidence where the reader will look for the item. That might be a resume line showing an adjacent tool, a screener answer that adds context where the form allows it, or one sentence in the cover letter. Matching each Screen to a line in your history is the tedious part, and it's the part Refolk handles when it tailors your resume and cover letter to the posting and scores how well you actually fit.

The posting lists [requirement]. I have not held [exact item], but I have [nearest true evidence: tool, project, or credential] at [employer or context], where I [specific result], and I can [close the gap: course, timeline, or transfer] if it matters for the role.

Before you call a posting graded, verify:

How do you keep the grades current?

Grades go stale at different speeds. Re-run the peer count for each new date window, re-read screening questions on every application, and re-confirm licenses with the board when you change states.

What changes, and how to re-check it:

• Posting language on degrees moves both ways. Indeed's share of postings requiring a bachelor's degree or higher fell from 20.4% to 16.6% and then rose back to 19.3%. Don't carry an old belief about whether a field still asks for degrees. Recount at rung 5 in the current window.
• Platform rules are set by the platform, not the employer. What can be made required on Indeed or LinkedIn is defined on their help pages, and those pages are not always consistent with each other. Indeed's own regional pages give different figures for how much screener questions lift hiring: 40% on one and 30% on another. When a grade depends on a platform rule, re-read the current help page instead of relying on memory.
• License coverage depends on which states participate. CareerOneStop's data depends on state submissions, so the state board stays the source of record.
• Incumbent samples change as people move. Re-run step 6 when you come back to an employer after a long gap, because this is the check most likely to catch a change in how they actually hire.