Open any job description and count the requirements. Then count how many of them the hiring manager would actually reject a strong candidate over.
The two numbers are rarely close. A senior accountant posting will list twelve requirements. Four of them decide the hire. The other eight are preferences, proxies, and things the last person in the role happened to have. But the posting doesn't say which are which, so the recruiter screens on all twelve, the candidate self-selects against all twelve, and the interview spends its first twenty minutes confirming things that never mattered.
That is the must-have trap. It is the quiet reason a lot of "skills-based hiring" efforts stall: the company changed its language and kept its gates.
What a must-have actually does
A must-have is a filter, not an emphasis. Every requirement labeled as required does three things, whether or not anyone intended them:
- It removes candidates from the pool before an interview happens.
- It tells the recruiter what to screen on, which means it decides what the first conversation is about.
- It tells candidates whether stretch is welcome. A list of ten hard requirements signals perfect match only.
So for each requirement, ask whether you would turn down someone who can do the job because they lack it. Almost everything on a job description is useful, so usefulness is not the test. If the honest answer is no, it isn't a must-have, and calling it one is costing you candidates and interview time.
What long requirement lists do to the pool
The research here does not isolate the label "must-have." It triangulates from job-ad length, requirement inflation, degree requirements, exact-match filters, and applicant behavior. The pattern across those studies is consistent.
Long lists shrink pools, and what is on them matters. LinkedIn's analysis of its own job posts found that qualification sections focused on hard skills lifted apply rates by 56%, while sections that led with pedigree (degrees, years, brand-name employers) lowered them. Posts that listed skills in the requirements section had an 11% higher view-to-apply rate than posts that did not. Shorter posts outperformed medium-length ones. Read that as a case for concrete, short requirements rather than a case for listing nothing.
Proxies screen out people who can do the work. In LinkedIn's Economic Graph research on skills-first hiring, 88% of hirers admitted they filter out highly skilled candidates for lacking traditional credentials. A degree requirement, a rigid years-of-experience number, or an exact title match are all standing in for something else. The degree stands in for "can learn technical material." The years stand in for "has handled complexity." Those are real things to assess. The proxy assesses them badly and excludes people who have the skill without the credential.
Requirement framing changes who applies. The most repeated version of this claim, that women apply only when they meet 100% of requirements and men at 60%, is folklore. The UK Behavioural Insights Team tested it in a randomized experiment with 10,468 participants and found apply thresholds of about 52% for men and 56% for women: a real gap, and a much smaller one, concentrated among less-qualified applicants. LinkedIn's own data show women apply to 20% fewer jobs than men and are 16% more likely to be hired once they do. So the problem is self-selection, not confidence, and the fix sits in how the employer frames the requirements rather than in coaching candidates to be bolder.
This is an argument for making the hard gate small, specific, and real, not for removing standards.
Must-have vs nice-to-have: the five tests a requirement has to pass
A requirement earns the must-have label if at least one of these is true:
- It is legally required. A license, a certification, a work authorization the law demands.
- It is safety-critical. Getting it wrong hurts someone.
- It is needed for an essential function on day one. Not week six. Day one.
- Normal onboarding cannot close the gap. If you would teach it to a new hire anyway, it is not a gate.
- No adjacent strength can compensate. If a candidate could be weak here and still excel because they are strong on the role-defining skill, it is a preference, not a gate.
And a requirement should not be a must-have when:
- it can be taught during onboarding
- it is a preference dressed as a requirement
- it is a proxy: degree, years, prestige employer, title match
- it is a vague trait ("self-starter," "strong communicator")
- it describes the mature version of a job you are hiring a trainee into
- it matters, but a weak score could be offset by strength on the skill that defines the role
Run a typical requirements list through those tests and most of it moves. The degree becomes "can learn the technical material," which is something to assess in the interview. The "five years" becomes "has run a month-end close under a deadline," which is a question. The "excellent communicator" becomes a specific behavior: has explained a technical adjustment to someone who did not want to hear it.
What this does to the interview
This is the part hiring teams underestimate. Requirements shape what gets asked, not only who applies.
When twelve things are required, the screen turns into a checklist. The recruiter confirms the degree, the years, the tool names, the certification. Every minute spent confirming a proxy is a minute not spent on evidence. And because the checklist items are easy to prepare for, a candidate can pass all twelve and still not be able to do the job.
When three things are required and the rest are scored, the interview changes shape. The must-haves get the hardest, most specific questions, because a miss there ends the conversation. The preferred skills get questions too, but a weak answer on one of them is information, not a verdict. The trainable skills get no interview time at all. That is where the hour goes to the things that predict.
There is a scoring consequence as well. A must-have should carry real weight in the final decision, which is exactly why there should be few of them. If eight skills are all non-negotiable, none of them is. The label only means something when it is scarce.
A working budget for requirements
The right number depends on what the list is for.
| What the list is for | Working default |
|---|---|
| The skills an interview plan is built around | 2 to 3 must-haves; 4 is the ceiling |
| A public job posting for a typical professional role | 4 to 6 core requirements; review anything above 7 |
| A regulated or safety-critical role | More is allowed, but each one needs a written reason |
| An entry-level or trainee role | 3 to 4, and never the craft you plan to teach |
The interview budget is the strictest on purpose. An interview plan is a plan for collecting evidence in a fixed amount of time, and the must-have flag decides where the hardest questions go. Two or three is enough to protect the hire. More than four and the plan stops making choices.
If a role appears to need more than four, that is a signal to make a decision the job description avoided. Which skills are essential on day one? Which are differentiators? Which are trainable? Which are credentials or proxies that should be verified outside the interview, if at all? The answer to those questions is the skills-based job description. The original list was the old one with better vocabulary.
Clean gates are not the whole job
One caveat. Cutting the must-have list fixes one kind of risk in a job description. A posting can have three clean, defensible gates and still carry language that narrows the pool or invites subjective screening: culture-fit phrasing, career-stage signals, physical demands no one checked against the actual work, blanket background language. That is a separate pass, and it deserves its own review rather than being folded into the requirements cleanup. Nothing in this article is legal advice; anything touching protected categories or essential-function rules needs a current check for the jurisdiction.
But the requirements cleanup comes first, because it is the one that changes what the interview tests.
How Ratio fits
The reason most job descriptions carry twelve requirements is that nobody had time to decide which four mattered. Ratio makes that decision explicit. Working from whatever you have on the job, whether that is a job description, intake notes, or a hiring-manager conversation, it picks the eight or so skills the role actually needs and marks two or three as must-haves with the reasoning shown. Proxies get named and set aside. Trainable skills stay out of the interview. Each must-have gets the hardest question in the plan, and the rest get scored rather than gated. The result is an interview built around what the job requires on day one, not around the last person's resume. Book a working session and bring the job description with the longest requirements list you have.
References
- Behavioural Insights Team (2022). Gender Differences in Response to Requirements in Job Adverts. Randomized online experiment, n = 10,468.
- LinkedIn Gender Insights Report (2019). Platform data on browsing, applying, and hiring behavior by gender.
- LinkedIn Talent Solutions (2023). Analysis of job-post performance: skills listed in requirements and post length versus view-to-apply rate.
- LinkedIn Talent Solutions. Qualifications-section text analysis: hard-skills focus versus pedigree focus and apply rates.
- LinkedIn Economic Graph. Skills-First: Reimagining the Labor Market and Breaking Down Barriers. Share of hirers filtering out skilled candidates for lacking credentials.
Ratio's internal source for these figures is its 2026 requirements research review; each LinkedIn figure should be checked against the original report before publication, and the year and exact title confirmed. The Behavioural Insights Team experiment is cited from the published report.
Frequently asked questions
What is the difference between a must-have and a nice-to-have requirement?
A must-have is a gate that removes candidates from the pool before anyone assesses whether they can do the work. A nice-to-have is something you score. A requirement earns the must-have label only if it is legally required, safety-critical, needed for an essential function on day one, cannot be closed by normal onboarding, or cannot be offset by strength on the skill that defines the role. Everything else is a preference, a proxy, or something to teach.
How many requirements should a job description have?
For a public posting on a typical professional role, four to six core requirements, and review anything above seven. For the skills an interview plan is built around, two or three must-haves, with four as the ceiling. Regulated or safety-critical roles can carry more, but each one needs a written reason. Entry-level roles should carry three or four, and never the craft you plan to teach.
Do long job descriptions reduce the number of applicants?
Yes, and what is on the list matters as much as the length. LinkedIn's analysis of its own job posts found that qualification sections focused on hard skills lifted apply rates by 56%, while sections leading with pedigree (degrees, years, brand-name employers) lowered them. Posts listing skills in the requirements section had an 11% higher view-to-apply rate, and shorter posts outperformed medium-length ones.
Is it true that women only apply when they meet 100% of job requirements?
No. That figure is folklore. The UK Behavioural Insights Team tested it in a randomized experiment with 10,468 participants and found apply thresholds of about 52% for men and 56% for women: a real gap, and a much smaller one, concentrated among less-qualified applicants. LinkedIn's data show women apply to 20% fewer jobs than men and are 16% more likely to be hired once they do. The effect is self-selection against the posted requirements, which makes it the employer's wording to fix.
How do job requirements affect the interview itself?
They decide what gets asked. When twelve things are required, the screen becomes a checklist that confirms degrees, years, and tool names, and a candidate can pass all twelve without being able to do the job. When three things are required and the rest are scored, the must-haves get the hardest and most specific questions, preferred skills get questions whose weak answers are information rather than a verdict, and trainable skills get no interview time at all.
Was this useful?