What is a good ATS score? (No ATS gives you one)

No applicant tracking system gives your resume a score out of 100. Workday, Greenhouse, Lever, iCIMS and the rest do not compute one, do not store one, and do not show a recruiter one. Every “ATS score” you have ever seen came from a checker tool — including ours.

That makes the real question different from the one people ask. Not “what score do I need to pass”, but what is this number actually measuring, and is it measuring anything I can act on?

What an ATS really does

• Parses your file into fields: name, contact, employers, titles, dates, education, skills. This part is genuinely pass or fail, and it is the part a checker can honestly test.
• Stores you in a searchable database. Recruiters run searches against it — by title, tool, certification, location.
• Sometimes ranks or filters applicants against the posting, or applies knockout questions you answered in the form (work authorisation, notice period, minimum years). Ranking is decision support shown to a human; the knockout questions are the thing that actually auto-rejects, and they are not on your resume at all.

Notice what is missing: nothing in that list produces a grade for your document. The widely repeated claim that “75% of resumes are rejected by a robot before a human sees them” has no credible source behind it and has been traced back to marketing copy. Most rejections are human decisions made quickly.

So is a resume score useless?

No — as a checklist. A score is useful exactly to the degree that you can see what it is made of and fix the parts. It is useless as a verdict, and dangerous as a target, because optimising a number that no employer computes is a way to spend a week not applying to jobs.

A good rule: if a checker will not show you its categories and their weights, ignore the number. A score with no breakdown is a conversion device.

What our score is made of

Here is the whole rubric, weights included. There is nothing else in it:

• Parseability — 20 points. We render your resume, extract the text back out of the PDF, and compare it to what the document says. Missing text, scrambled order, no text layer at all.
• Bullet quality — 16. Action verbs, length, first person, duty-restating filler.
• Contact completeness — 12. Email, phone, location, links — present, and in the body where a parser reads them.
• Section structure — 12. Standard headings in the order screening tools expect.
• Quantification — 12. How many bullets carry a number, a scale, or a result.
• Dates — 10. Consistent format, no impossible ranges, nothing missing.
• Length — 10. Page count against your experience, and how full the last page is.
• Skills — 8. Present, grouped, not a wall of forty terms.

Eight categories, 100 points. The full breakdown, with what loses points in each, is published — you can read the rubric without running anything.

What counts as a good score

• Below 60: something structural is wrong. Usually parseability — a scanned file, a text-box layout, contact details in a header.
• 60–79: the document is readable and the writing is doing less than it could. Almost always quantification and bullets.
• 80–94: good. Remaining points are usually taste.
• 95+: chasing the last points is not worth an evening. Tailor to a posting instead.

And one line worth being blunt about: a 98 does not mean you will get interviews. It means the document is not the thing standing in the way. Fit, targeting, timing and volume are still out there, and this is how to work out which of them is the problem.

What about the Naukri or LinkedIn score?

Those measure profile completeness on their own platform — filled fields, a photo, verified contact details, recency. That is a real thing with a real effect on that platform’s search, and it is unrelated to how a PDF parses at an employer’s ATS. Both are worth filling in. Neither is a resume score.

Where this site fits

The free checker takes a PDF, needs no account, stores nothing, and prints the eight categories with the specific findings under each. It also tells you when it cannot help — a scanned image gets a refusal and an explanation, not a low score with vague advice.

Inside the builder, the same rubric runs live on the resume you are editing, against the text extracted from a real render of your actual PDF, not against the form fields.