Loading account…
ATS and parsing

How Resume Parsing Works: Extraction, Field Recognition and Common Failures

Luc Resume · Editorial team

Resume parsing is the process of using software to interpret a document and populate structured candidate information. It may involve obtaining readable text, recognizing names and organizations, associating job titles with dates, and placing data into profile fields. Parsing is not equivalent to deciding who should be hired.

A conceptual parsing pipeline

  1. File acceptance: The upload channel checks its permitted formats and size restrictions.
  2. Document interpretation: Software reads or extracts content from the document.
  3. Field recognition: It attempts to identify name, contact information, employers, job titles, education and dates.
  4. Association: It tries to decide which title and date belong to each employment or education entry.
  5. Candidate profile: Extracted values may populate fields, sometimes requiring manual confirmation.

This is an explanatory model; vendors do not all expose the same implementation or process the same formats.

Why a visually correct resume can extract badly

Human eyes naturally connect a right-aligned date with a role to the left. A parser may instead read two columns in an unexpected sequence. It may mistake a header for decoration, or treat characters with large letter spacing as disconnected fragments.

The Greenhouse Recruiting support article specifically identifies possible issues with graphics, images, text boxes, headers/footers, tables, columns, spaced-out letters, unclear sections and some incomplete organization or job names. Its article was updated on March 2, 2026. Greenhouse notes that a resume may remain attached when its fields are not parsed.

These are vendor-documented possibilities, not an exhaustive list of behaviors in every ATS.

Distinguish the three failure types

FailureWhat you might observeWhat you can do
Upload failureFile type or size rejectedCheck allowed extensions and current limit
Extraction failureText missing or scrambledTry selectable text and simpler reading order
Field association failureDates, employers or titles put in wrong fieldsCorrect portal fields where possible and simplify grouping

A fourth issue is not technical at all: the resume may lack the actual qualification the employer requires. Better extraction cannot fix missing qualifications.

The importance of file limits

Workable's official application-form help lists 5 MB as a resume upload limit for its application flow and accepts PDF, DOC, DOCX, RTF, HTML and ODT.

Greenhouse's parsing troubleshooting article says its named parser cannot process resumes above 2.5 MB. A 5 MB upload rule and a 2.5 MB parsing limit are not the same operation. Neither number should be stated as the universal ATS maximum.

Fictional troubleshooting case

Fictional applicant: Amir has a two-column resume designed with text boxes.

In a text editor, the document appears clear: role titles left, dates right, skills in a narrow sidebar. In the exported PDF, the visible layout still looks correct. But copying the text places “SQL, Excel, Reporting” in the middle of a former employer's name, and a date appears before the corresponding role.

Diagnosis: The content order is not clean enough for a predictable text reading sequence. This is a document risk, not proof that Amir has been rejected.

Repair: Move the skills block after Experience, make each role and date part of the main text flow, remove unnecessary boxes, export again, and review copied text. If an application platform has editable extracted fields, check them before submitting.

What is a useful local test—and what is not?

A selectability test checks whether text can be highlighted. A copy-order test checks whether common readers extract information coherently. A visual check confirms dates, headings and line breaks. All three are useful document QA procedures.

They are not ATS certification, cannot emulate every vendor's underlying parser and cannot predict a hiring outcome. Also, be careful with online “ATS scanners”: understand where the resume data goes before uploading private information to a third-party tool.

More causes of confusion

  • A scanned PDF may contain only images, even though the extension is PDF.
  • A resume may use abbreviations a parser does not associate correctly with employers or jobs.
  • Two roles at one company may be grouped in a way that makes dates ambiguous.
  • Contact fields placed solely in a header or text box may not be reliably extracted in some systems.
  • A recruiter may manually read an attached resume even when extracted fields are incomplete.

Do not infer automatic rejection from any one of these observations without evidence about that employer's actual process.

Practical preparation checklist

  • Document is within the target portal's current file limits.
  • Text is genuinely extractable where the format supports it.
  • Section headings identify common information clearly.
  • Dates are near the corresponding jobs and qualifications.
  • Names and titles have recognizable complete wording.
  • Copy/paste order is reasonable.
  • Auto-filled fields are inspected and corrected where possible.
  • I have not mistaken a formatting test for a hiring prediction.

Parsing is a technical interpretation step; accuracy depends on the document, software and workflow.

Sources and further reading

Verify named vendor behavior and employer requirements again before publication if material policy or file limits change.

Put this guidance into your resume

Organize your actual experience, preview templates, and inspect the final PDF in the resume workspace.

Open the free resume workspace

Related articles

Knowledge library articles provide general job-application guidance. They do not provide legal, immigration, medical, or hiring-outcome guarantees. Follow the latest instructions from employers and official bodies.