Most CV-vs-ATS advice operates in the abstract. "Use a simple layout." "Avoid tables." "Stick to standard fonts." Useful as general rules, but you can run a 90-second test on your own CV today that turns the abstract into evidence.
This post is the test, the tools, and what the output actually means.
What an ATS parser actually does
When you upload a CV to a job posting, the ATS extracts the text into a structured profile: name, contact details, work history (one entry per role with company, title, dates, bullets), education, skills. Some of this comes from explicit fields, most from inference based on visual cues — headings, font sizes, indentation, column structure.
The parser does two things you can test for:
- Text extraction — does it pull every word from the document, or does some get dropped (text in images, text in headers/footers, text in non-standard columns)?
- Field mapping — does it correctly identify what is a job title, what is a company, what is a date, what is a bullet under which role?
A CV can pass step 1 (every word is extracted) and fail step 2 (the parser thinks your job at Google was at "Sept 2021"). Both matter, and the failure modes are different.
The free tools worth using
Three categories of free ATS-check tools, with different strengths:
1. PDF-text extraction tools (testing step 1)
The lowest-tech and most reliable approach: open your CV as a PDF, save it as plain text (most PDF readers have a "Save As → Text" option, or use any free online PDF-to-text converter). Look at the raw text output.
This tells you what an ATS sees before any parsing logic is applied. If a word is in the visual document but missing from the text dump, it will be missing from the ATS profile too. Most failures here come from:
- Text inside images (a logo, a graph, an icon with text)
- Text in headers or footers (some ATS parsers ignore these entirely)
- Decorative characters used as bullets (•, ★, ▶) that show as
?or get dropped - Two-column layouts where the parser reads across rows instead of down columns
If the text dump looks coherent and complete, your CV passes step 1.
2. ATS-simulator tools (testing step 2)
A second category of tool uploads your CV and shows you the extracted structured profile — work entries, skills, education — the way an ATS would build it. Useful free options include Jobscan's free tier, Resume Worded's scan, CV Compiler, and EnhanCV's checker. None are perfect (each ATS in production parses slightly differently), but they catch the big structural failures.
What to look for in the output:
- Each job appears as one entry with the right company, title, and dates. If two jobs got merged into one, or one job got split across two entries, your dates or section breaks are confusing the parser.
- Skills appear in the skills bucket rather than scattered into the experience section. If the parser cannot find your skills list, recruiter search filters will not find you either.
- Education shows the institution, degree, and dates correctly.
- Contact details map cleanly — name, email, phone, location all in the right slots.
3. Job-match scoring tools
A third category compares your CV against a specific job posting and gives you a percentage match. These are useful for one thing: spotting required keywords from the job ad that are missing from your CV. Ignore the overall score ("68% match") — the methodology varies wildly between tools and the number is not calibrated. Use the keyword diff: "You don't mention Terraform, the job asks for Terraform."
A reasonable workflow: run one ATS-simulator scan (step 2) to check structure, then run one job-match scan against the target posting to spot missing keywords. Do not run six and average them.
Check resume ATS compatibility — what the output is telling you
The most common parser failures, and the fix for each:
Your name is missing or in the wrong field. Usually caused by a name in a non-standard font, or your name placed in the document's header rather than the body. Fix: name as the first line of the document body, in plain text, in the same font as the rest.
Job dates look scrambled. Usually caused by inconsistent date formats ("Mar 2024", "03/2024", "March 2024" all in the same CV). Fix: pick one format and use it across every role.
One job entry got split into two. Usually caused by a horizontal line, a page break, or an unusually large gap between bullets. Fix: keep each role's content visually grouped on the same page where possible, and remove decorative dividers.
Two job entries got merged into one. Usually caused by missing or inconsistent section breaks. Fix: leave clear vertical space between roles, and make sure the company name is on its own line.
Skills are missing entirely. Usually caused by skills shown as icons or progress bars instead of plain text. Fix: plain text under a clearly-labelled Skills heading.
Bullets show as question marks or empty squares. Usually caused by exotic Unicode bullets. Fix: use the standard • (which most parsers handle), or a simple - dash.
Text from a two-column CV is reading left-to-right across rows. Usually caused by the parser not recognising the column structure. Fix: switch to a single-column layout. Two-column CVs look elegant and fail parsing more often than any other layout pattern. This is the single highest-ROI fix for ATS-incompatible CVs.
A 90-second self-test you can do today
No tool required:
- Open your CV PDF in any PDF reader.
- Select all the text (
Cmd/Ctrl+A). - Copy and paste it into a plain text file or empty email draft.
- Read what you see.
If the output reads like a coherent CV — name at the top, jobs in order with dates and bullets, skills as a list — you are in good shape. If it reads like a word salad with random line breaks, missing sections, or words run together, the ATS is seeing the same thing.
This test catches roughly 80% of ATS-formatting problems for the price of 90 seconds. The other 20% — field-mapping issues — need an actual ATS-simulator tool.
What not to bother with
- "AI-powered" tools that promise a score with no breakdown. Use the diagnostic, not the verdict.
- Tools that demand a paid upgrade to show the parsed output. The free tier should at least show you what was extracted.
- Running the same CV through ten tools and averaging. Each tool has its own parser. Pick one good one and trust the structural feedback.
- Pretending Word vs PDF matters. Both parse well when the underlying structure is clean. The format is not the problem; the layout is.
The Postulit team works on a related problem: starting from a LinkedIn profile, generating a CV that parses cleanly on the first try. The same principles apply whether you build the CV from scratch or generate it from LinkedIn — plain text, single column, clear section headings, consistent dates, defensible skills.
Spend 90 seconds on the copy-paste test today. It catches more problems than a week of layout tweaking based on advice articles.
If the test surfaces something broken, fix that one thing. Then re-test. Three iterations on a real test beats fifty hours of speculative formatting changes.