What is an ATS score? What the number actually measures
Two different things are called an ATS score. The one you can buy is calculated by the company selling it, against its own rubric, and the employer never sees it.
An ATS score is not one thing. Two different numbers go by that name, and telling them apart answers most of the question.
The first is the score a resume checker sells you. It is calculated by the company selling it, on its own servers, against its own rubric, and the employer never receives it. The second is a ranking signal inside the employer's hiring system. That one is real, recruiters do see it, and no vendor publishes its scale or the threshold at which it matters.
Neither is a grade your resume gets.
| The score you buy | The score the employer sees | |
|---|---|---|
| Who calculates it | The resume tool | The hiring system vendor |
| What it compares | Your file against a posting you pasted in | You against the other applicants |
| Who sees it | You | The recruiter |
| Published method | None found | None found |
| Reaches the employer | No | It is already there |
| Rejects you | No | Not on any documented path |
The score you can buy
A resume checker takes two inputs, your file and a job description, and returns a percentage. Jobscan calls it a match rate. Others call it a resume score, a compatibility score or an ATS score. The mechanism is the same in all of them: extract terms from the posting, look for them in your resume, weight the result, print a number.
The extraction half of that is real work, and Jobscan does it better than most. Its product simulates parsing against named vendor systems rather than checking generically, which is a harder thing to build than it looks.
The number layered on top is where it stops being a measurement.
Jobscan's own page on the subject says: "We generally recommend a match rate score of 80%. However, many career counselors and Jobscan users see success even with just a 75% match rate." No study is cited for either figure. The same page describes the mechanism plainly: "our matching algorithm is based off of top applicant tracking systems." Based off. Not connected to.
The threshold is the part of a score that has to be validated, because the threshold is the part that tells you to act. A recommendation that moves five points inside two sentences, with no dataset behind either position, is a suggestion wearing a percentage sign.
This is a vendor page about the vendor's own product, so read it as a commercial claim. It is also the clearest public statement of how these numbers work, which is why it is worth quoting rather than paraphrasing.
The score the employer sees
There genuinely is scoring inside hiring software. Workday markets HiredScore with the line "Identify the most relevant candidates quickly with unbiased, AI-driven candidate grading." Its product documentation describes presenting parsed resumes and calculating derived signals: "HiredScore calculates employment gaps and time in position."
What the documentation does not contain is a grade scale, a rejection threshold, or a published bias audit for the word "unbiased". A recruiter opens a profile, reads a grade, and decides. The decision is the human's.
Rejection, where it is automated, runs on a different track entirely. Greenhouse documents auto-reject as firing "based on an applicant's answer to a question", and supports only Yes/No, single-select and multi-select question types. Ashby documents the same design, evaluating conditions such as is empty, equals, contains and similar to against application form fields. Neither reads your resume to do it.
A resume cannot fail a filter that does not read it. The four-minute rejection is almost always a knockout question you answered, not a document that scored badly.
One small study is worth naming here with its limits attached. Enhancv, a resume company, interviewed 25 US recruiters across more than ten platforms and reported that 92% said their systems do not automatically reject on formatting, design, missing keywords or low AI match scores, and that 44% of the platforms in use offered automated fit scores which most recruiters disabled or treated as preliminary. That is a sample of 25, it is qualitative, and the publisher sells resumes. Treat it as directional. It agrees with the vendor documentation, which is the stronger source.
Why nobody outside the company can calculate your score
A score needs three things to mean anything: a rubric, a population, and a threshold. From outside the employer's system you have none of them.
You do not know which of the posting's requirements the recruiter marked as hard and which are aspirational. You do not know who else applied, and a ranking without a population is arithmetic without a denominator. You do not know at what point the recruiter stops reading the list, which is the only threshold that decides anything.
A tool that has none of that can still tell you something useful. It can tell you which words in the posting do not appear in your resume. That is a real, checkable finding, and it is what the percentage is standing in front of.
The filter that does exclude qualified people is documented, and it is not a score. Harvard Business School and Accenture surveyed more than 8,000 workers and over 2,250 executives for Hidden Workers: Untapped Talent. They found that "A significant majority, 88%, of employers believed that qualified high-skills candidates were vetted out of the process because they did not match the exact criteria established by the job description." The number rose to 94% for middle-skills roles.
Criteria written by a person, applied by software. A percentage does not tell you which criterion you missed. A list of the requirements your resume never mentions does.
The scoring that is actually regulated
Worth knowing if you apply in Europe. The EU AI Act lists as high-risk "AI systems intended to be used for the recruitment or selection of natural persons, in particular to place targeted job advertisements, to analyse and filter job applications, and to evaluate candidates."
That classification lands on the employer's system, not on the checker you run on yourself. Its practical effect is the direction of travel: scoring that influences a hiring decision is becoming more accountable and more advisory, not less. A vendor grade nobody can explain is a harder thing to defend in 2026 than it was in 2019.
What to check instead
Three things, in order, and none of them produce a number.
- Read the parse. Upload your resume somewhere that shows you the extracted fields instead of a score. Is your current employer in the employer field? Did your dates survive? Did a two-column layout run your job titles together? This is the only part of the process you fully control, and it is checkable in about twenty seconds. The full method is in is my resume ATS friendly.
- List the requirements your resume never mentions. Not a match rate. Coverage: which stated requirements you answer, which you answer weakly, and which are absent. Then decide what to change, using only work you actually did. The twenty-minute version is in how to tailor your resume to a job description.
- Stop trying to move the number. Every tactic aimed at a score rather than a reader ages badly, and the current crop ages fastest. Hidden white keywords are the clearest example, and what happens when you try them is not what the guides promise.
The same logic applies to the newer worry about whether a machine can tell you used AI. The answer there is also no, for the same structural reason, and it is covered in can recruiters tell if you used AI.
The short version
An ATS score you can buy measures your agreement with the seller's rubric. An ATS score inside a hiring system is an advisory signal with no published scale. Neither one is a pass mark, and there is no threshold in either that anyone has validated against getting hired.
Rzume does not calculate one. It shows you the parse first, field by field, including what it could not place, and then names the requirements your resume does not answer. The parse and the full review are free and need no account.
Upload your resume, or read what each plan includes first.
Sources
- Greenhouse, Auto-reject support documentation.
- Ashby, Auto-reject applications product documentation.
- Workday, HiredScore candidate profiles documentation and the AI recruiting product page.
- Jobscan, What Jobscan match rate should I aim for?. Vendor page about the vendor's own score, quoted as a commercial claim.
- Fuller, J., Raman, M., Sage-Gavin, E., Hines, K., et al., Hidden Workers: Untapped Talent, Harvard Business School Project on Managing the Future of Work and Accenture, September 2021. More than 8,000 hidden workers and more than 2,250 executives across the US, UK and Germany.
- EU AI Act, Annex III, point 4 on employment and worker management.
- Enhancv study of 25 US recruiters, as reported by ITBrief. Vendor-published, qualitative, n=25.
All accessed 3 September 2026.
