We took one clean resume that scores 100/100, broke it one way at a time, and measured what each formatting mistake actually costs. Real numbers, one variable at a time, fully reproducible.
How much does formatting affect your ATS score? In this controlled benchmark, single formatting mistakes each cost up to 4 points of ATS parse score, and a resume with every issue at once dropped from 100 to 85/100. The most damaging factors were non-standard section headings and inconsistent date formats. Measured with our deterministic engine — not a vendor pass/fail.
What we found
The single most damaging formatting choice was non-standard section heading (4 points), because the parser can no longer tell where that content belongs.
Issues don’t stack linearly. The 11 factors sum to 23 points in isolation, but a resume carrying all of them scored 85/100 — a 15-point drop, not 23.
A few flagged issues (like a sloppy file name) are worth fixing for a human reader but barely move the machine-parse score — useful to know where the points actually are.
Parse-score impact, ranked
Each row changes exactly one thing on the baseline resume. “Points lost” is the drop from 100/100.
Non-standard section headingMajor
“Work Experience” → “Where I’ve Made an Impact”
−4pts
Inconsistent date formatsMajor
“Jun 2018” → “06/2018” (now two formats in use)
−4pts
Unparseable dateMajor
“Jun 2018” → “Spring 2018”
−3pts
Present tense in a past roleMajor
“Launched a merchant…” → “Manage a merchant…” (in a 2018–2020 role)
“Jordan Lee Product Manager Resume” → “resume_FINAL_v3”
0pts
A resume with all 11 issues at once
12 findings, dropping it from 100 to 85/100.
85/100
Methodology
We start from one clean baseline resume — a realistic mid-career profile that scores 100/100 with zero findings. For each factor, we clone that baseline, change exactly one thing, and re-score it with the same engine. The difference is that factor’s isolated impact.
The scorer is the ResumesTailor ATS engine (ats@1.0.0): a deterministic, rule-based model of the checks applicant tracking systems run — contact-block parsing, date consistency, standard section headings, clean bullets, verb tense, spelling, and document metadata. Because it’s deterministic, the same resume always produces the same score, so this is fully reproducible from the published CSV.
What this is not: these are parse-readability scores from our engine, not accept/reject decisions from Workday, Greenhouse, Lever, or any specific vendor. We don’t have access to proprietary ATS parsers, so we don’t publish a number we can’t defend. See our methodology page for how the engine works.
How common are these issues in the wild?
This is the part that grows. As job seekers run resumes through our free ATS checker, we’re building an anonymized corpus to publish how often each issue shows up in real resumes. We’ll release those real-world frequencies here once the sample is large enough to be meaningful — never before. (No resume text is ever stored; only aggregate counts.)
Frequently asked questions
How much does formatting affect your ATS score?
In this controlled benchmark, individual formatting mistakes each cost between 0 and 4 points of ATS parse score, and a resume carrying every issue at once scored 85 out of 100. The single most damaging factor we measured was "non-standard section heading" at 4 points.
What is the worst resume formatting mistake for ATS?
Of the factors we tested, "non-standard section heading" had the largest single impact (4 points), tied with inconsistent date formats. Both confuse the parser about where your content belongs or how to order your history.
Is this a test of real ATS systems like Workday or Greenhouse?
No, and we're explicit about that. These are parse-readability scores from the ResumesTailor ATS engine — a deterministic, rule-based model of the checks applicant tracking systems run (contact parsing, date consistency, standard headings, clean bullets, spelling, and document metadata). They are not accept/reject decisions from any specific vendor. We don't have access to proprietary ATS parsers, so we don't claim numbers we can't defend.
Can I download the data?
Yes. The full results are available as a CSV under a Creative Commons CC BY 4.0 license — free to cite or republish with attribution. Each row lists the factor, the rule it triggered, and the measured point impact.
How was this measured?
We start from one clean baseline resume that scores 100 out of 100 with zero findings. We then change exactly one thing at a time — one formatting mistake per variant — and re-score it with the same deterministic engine. The difference is that factor's isolated impact. Because the engine is deterministic, the same input always produces the same score, so anyone can reproduce it.
Where does your resume land?
Run the same engine on your own resume — free, no signup. See which of these issues you have and exactly what they cost you.