What is an ATS score?
It is a rating of how well a resume fits a job, and nobody owns the term. A resume checker gives you its own number. Inside an employer's applicant tracking system (ATS), the rating is usually a category or a letter grade. Each one is built from different criteria, so two scores for the same resume can disagree.
An applicant tracking system is the software an employer uses to collect applications and move candidates through hiring. Greenhouse and Workday are two of them, and both publish help pages for the recruiters who use their ranking features. Those pages are the closest thing to a primary source on what an employer's score means, and everything below about employer systems comes from them, checked October 26, 2026.
The number a resume checker gives you is a different thing. It is that tool's own estimate, run before you apply, and the employer doesn't see it. Keep the two apart when you read any advice about scores, including this page.
What does an ATS score measure?
Most scores measure two things: whether software can read your file, and whether your resume shows what the job asks for. The first is about format, such as columns, headers and images. The second is about content, and it is usually judged by matching the words and skills on your resume against criteria taken from the posting.
Reading comes first, because a file the software can't read can't be matched. Greenhouse's help page on unsuccessful resume parsing, last updated March 2, 2026, lists what breaks it. The list includes "Resumes that have a columned layout," a resume uploaded as an image, contact details in a header, footer or text box, and "Resumes with incomplete job titles."
Matching comes second. Greenhouse's Talent Matching page, last updated October 5, 2026, says the recruiter first sets a "calibration," a set of "defined and weighted criteria" such as skills, experience and job titles. The resume is then compared with those criteria. Workday's August 2026 release note for HiredScore Spotlight describes the same idea, assessing each qualification "against the candidate's parsed resume."
What can't an ATS score see?
It can't see whether a claim is true. A score compares the words and figures on your resume with a list of criteria, and a made-up figure reads the same as a real one. Neither Greenhouse's nor Workday's help pages, checked October 26, 2026, describes any step that verifies what a resume claims.
Greenhouse's own parsing page shows the limit in a small way. Given a resume with a placeholder name and a real-looking phone number, it says the phone number "will parse correctly, since that information can't be definitively proven as real or fake." The software skips names like "First Last" and "Company 1" because they look fake. A plausible number gets through.
Here is the same point on a resume line, as an example. The person and every figure are made up for this page.
- Before: "Spearheaded customer success strategy, boosting retention 40% through stakeholder engagement and CRM optimization."
- After: "Ran quarterly check-ins for 35 accounts in Salesforce, and 31 of them renewed in 2025."
The number that changed is the 40%. Nobody measured it, and it came from wanting the line to sound bigger. The 31 of 35 comes from the CRM's renewal report. A matching system sees "customer success" and "CRM" in the first line and "Salesforce" in the second, and it has no way to tell which figure would survive a question.
How do recruiters actually use an ATS score?
By the vendors' own accounts, mostly to sort a long list so a person reads the likely fits first. Greenhouse groups candidates into five match categories and says its tool "does not auto-reject or auto-advance any candidate." Workday ranks candidates from A to D. In both systems, the employer decides what the job requires.
Greenhouse's categories are Strong match, Good match, Partial match, Limited match and Needs manual review. Its Talent Matching FAQ, last updated October 5, 2026, says "humans decide how to advance, reject, or follow up with each candidate." A recruiter can override a match score after reading the resume, and a resume with "rare formatting or technical issues" goes to manual review instead of being scored.
Workday's candidate grades reference, last updated June 23, 2023, defines a B as a candidate who "meets all basic qualifications" and an A as one who also meets most preferred ones. Under the August 2026 update, which customers opt into, recruiters see "which requirements are met and not met, and how they influenced the candidate's grade."
One feature goes further. Workday's HiredScore feature list, last updated October 18, 2024, includes Candidate Fast Lane, which can "automatically advance top-rated applicants" to the hiring manager review stage. So at an employer that turns it on, a top grade can move you forward without a recruiter clicking first. The page doesn't say which employers use it.
Recruiters also search. Greenhouse explains how to search resumes for keywords and says "a small snippet under their name will show how the search terms are used in their profile." That help page was last updated June 6, 2022. The recruiter sees your word inside its sentence, so a skill with real work around it reads better than one in a list.
How do the scores compare side by side?
The table below sets the two employer systems next to the score you get from a resume checker before you apply, using Prept's as the checker example. Read the rows on who sets the criteria and what each score decides. The employer's criteria are the ones that count, and you never see them.
| Question | Greenhouse Talent Matching | Workday HiredScore Spotlight | Prept's ATS score (91 is the sample on its page) |
|---|---|---|---|
| What you get | One of 5 categories, Strong match to Needs manual review | A letter grade, A to D | A number out of 100; 70 is a pass |
| What it compares | Your resume against the recruiter's weighted criteria | Your parsed resume against the job's basic and preferred qualifications | Your file, read the way screening software reads it |
| Who sets the criteria | The recruiter, for each job | The employer's job requirements | Fixed checks: headings, email in the body, one column, standard fonts, selectable text |
| Does it decide anything | No; people advance and reject | Ranks candidates; Fast Lane can auto-advance top-rated ones if turned on | No; it runs before you apply |
| Does it check that a claim is true | Not described in its help pages | Not described in its help pages | No; confirming facts is a separate step |
The last row is the reason for this page. Every score in the table looks at the page, and none of them asks where a number came from. An interviewer asks exactly that.
Why isn't 91 the goal?
Because a high score tells you the file reads cleanly and little else. The 91 is the sample score on Prept's resume optimizer page, where 70 is the pass line and a critical problem fails a resume outright. Past the line, the points don't predict how an employer's system will rank you, because that system uses criteria you can't see.
Greenhouse adds a detail that undercuts chasing points. Its FAQ says several terms on a resume can map to the same skill, "so a longer list of matched terms doesn't always mean a higher match score." Its setup page tells recruiters to focus on 4 to 6 key skills for clearer matches. Repeating a posting's words doesn't add up the way a checker's percentage suggests.
Chasing the number also has a cost later. The easiest way to raise a match is to add terms and figures, and every one you add becomes an interview question. The page on a fact-locked resume explains why keyword stuffing backfires, and Prept vs Jobscan covers how a keyword match rate differs from a score with reasons.
What happens to a claim the score couldn't check?
It gets checked in the interview. A score can pass a line about 40% retention growth, and then a person reads it and asks how you measured it. Interviewers are trained to push on vague answers and on "we" answers until your own part and the result are clear. A number with no record behind it rarely survives that.
The US Office of Personnel Management's structured interview training tells interviewers to "seek clarification when candidates are vague." Its sample probes include "What was your role?" and "What was the outcome?" Try both on every figure on your resume before anyone else does.
The figure also has to stay the same across rounds. If 40% becomes 45% by the third interview, the notes disagree, and the page on whether your story changed between rounds shows how that gets noticed.
What should you do with your ATS score?
Use it as a format check, fix what it flags, and then stop. Once the file reads cleanly, the time is better spent on the claims. Check each figure with three questions: what the figure is, what it measured and where it came from. A line that fails one gets fixed or cut.
- Run your file through one checker and fix anything about reading: columns, contact details in a header, text inside images, abbreviated job titles.
- Read the posting and list what it requires, in its own words.
- Use the posting's word for a skill only where one of your facts shows you did that work.
- Ask the three questions of every number on the page.
- Cut or rewrite any line that fails, and leave a blank where you know a figure exists but haven't found it yet.
The page on how to turn your resume into facts walks through the three questions with a filled example. If step 4 leaves you with almost no numbers, the page on what to do when your resume has no numbers shows where to look. Step 3 is covered in full in how to tailor a resume to a job description without lying.
Prept's position is that a resume should read cleanly to software, and a number you can't stand behind is left off rather than invented. After each interview, note which figures came up and how you answered, using the free 10-minute debrief template. If a tool wrote numbers into your resume, the page on why AI resume tools invent numbers explains how they got there.