Résumé keywords, and what a checker can honestly tell you

Keywords matter to some readers and to some systems. Here is where they work, where they do not, and why this checker refuses to give you a keyword score.

ConxfolioPublished 2026-09-12

— no keyword match. That is on the report.

Almost every article about applicant tracking systems is about keywords, and most of them promise a match score. This page is about what can honestly be said, which is less than that and more useful.

What this checker does not do

The ATS résumé checker here takes no job description. It runs no keyword match. It predicts no outcome. Those three statements are printed on the report itself, above the score, because they are the limits of what a public tool can honestly claim. What it measures instead is whether a parser can read your document and whether the evidence is there to read.

A checker that offers a match percentage against a posting is measuring its own guess at how an employer's system weights terms. Nobody outside that company knows the weighting, and it differs between systems and between roles.

Where keywords genuinely matter

• A recruiter searching a database. They type a tool, a certification or a job title and read what comes back. The word has to be on your page.
• Systems that do match text. Some employer systems rank or filter on terms from the posting. You cannot see the rules, and having the right words for work you did costs nothing.
• The human reader. A hiring manager scanning for the thing they need is running the same search with their eyes.

How to choose them

Take the posting's own words for the work you have actually done. If they say "incident response" and you wrote "on-call rota", use their phrase. If they name a tool you have used, name it the way they spell it. Put each term inside a sentence that shows the work, because a term in context is credible and a term in a list is not. The tailoring guide is a twenty-minute pass that does exactly this.

What not to do

No keyword block at the bottom of the page. No white text, no hidden layers, no repeated terms in a font size nobody reads; the extraction takes everything, and what it takes is what gets seen. No terms for work you have not done, which is the same trade as an invented number and fails at the same moment.

What the check tells you instead

Whether the text comes out of your file in the right order. Whether your sections are recognised and carry content. Whether your contact details are readable. Whether your experience lines are written as achievements, carry figures, and go deep enough. Those are the things that are true of your document regardless of whose system reads it, which is why they are worth measuring.

The report prints the arithmetic behind the score, the counts of what it read, and the point cost of each fix. The guide to what an ATS reads covers the extraction step that all of this depends on.

Where the posting belongs

In the letter. The cover letter builder reads the posting and your résumé together, selects evidence from sentences your own document already contains, and returns the requirements your evidence cannot support as gaps rather than claiming them. That is the honest use of a job description in an application, and it is the one place in this product where a posting is read at all.

Terms that are genuinely worth checking

• Job titles. The industry's word for your role, not your employer's internal grade name.
• Tools and systems. Spelled as the posting spells them, including version numbers where they matter.
• Certifications. The issuer's own name for the certificate, which is what a search will use.
• Methods and standards. The named framework, regulation or process, written out once in full.

Each of these belongs inside a sentence that shows the work. A term in a list is a claim; the same term in a line about what you did with it is evidence, and the skills guide covers how a list is read.

What to do instead of chasing a match score

Spend the time on the three things that are true whoever reads the file: a layout that extracts in order, headings that are recognised, and experience lines that carry figures. Those improve the document for a parser, for a recruiter searching a database and for the person who decides. A match percentage against a posting improves none of them, and it is the one number in this subject nobody outside the employer can calculate.