Why the usual answer is not an answer
Ask this question and you will be given a number. Apply if you meet sixty per cent. Seventy. Eighty. The figure varies by article, the advice is delivered with the same confidence each time, and a source is rarely attached to any of it.
Treat a precise threshold as a warning sign rather than as guidance. It would require data nobody publishes — the requirements of a large sample of postings, who applied, who was screened out and why — and it assumes every line on a requirements list carries the same weight, which is the one thing you can see with your own eyes is false.
A posting listing eight essentials and one preference is not comparable to one listing two essentials and seven preferences, and no proportion can tell them apart.
The distinction that actually decides it
Every requirement you are missing is one of two kinds, and they behave completely differently.
An evidence gap
The requirement is about capability, and your history speaks to it even if it does not name it. The same work in a different tool. The same responsibility at a smaller scale. A skill you have used without it ever appearing in a job title.
These are arguable, and arguing them is the entire purpose of a cover letter. A reader will accept a transfer you make explicit; they will rarely make it for you.
An eligibility gap
The requirement is binary and external. A right to work in a country. A professional registration. A security clearance. A degree that is a legal precondition. The language the role is conducted in.
No letter closes these, because nobody on the other side has the authority to waive them. An application against an eligibility gap is not a long shot; it is an evening spent on a decision that was already made.
Sorting the lines of a posting into the ones that can be checked and the ones that cannot is the skill underneath this: must-have versus nice-to-have job requirements is about the wording employers use to signal it.
Where the gap sits matters more than how big it is
Two applications can be missing the same number of lines and be worth completely different amounts of effort.
- Missing three soft preferences, with the accountabilities evidenced: apply, and do not mention the preferences.
- Missing one named tool, with the accountabilities evidenced: apply, and name the closest thing you have actually used.
- Missing the accountabilities themselves and meeting every listed tool: this is the weaker application, despite the better-looking count. You are claiming you can do a job your history does not show you doing.
- Missing anything binary and external: skip it and keep the evening.
The years-of-experience line
This is the requirement that stops the most people, and it deserves a straight answer: it is a proxy. Nobody can measure whether you can do the work from a document, so a number stands in for it and keeps the screening pile manageable.
That does not make it fake — a screener with two hundred applications will use it — but it makes it the kind of line that gives way when the accountabilities are answered unmistakably. If the rest of your application makes the judgement easy, the number stops carrying the decision.
Being a year or two under is routine. Being half the stated figure with nothing else compensating is a different bet, and worth making deliberately rather than hopefully.
What applying anyway should cost you
The honest caveat to all of this is that applying is not free, and the advice to apply anyway is usually given as though it were. Every stretch application is an hour not spent on one you can evidence.
So make the stretch a decision rather than a reflex: apply through evidence gaps, address them directly in the letter, keep the count small, and put them after the applications you can fully support — not instead of them. How to prioritise job applications when your time is short is about that ordering.