How ATS resume parsing works

1. Parsing fills structured fields, it does not read a page

Every vendor source below describes the same mechanism: a resume is uploaded, software extracts the data, and the extracted data lands in the fields of a candidate profile. SAP's documentation describes parsing that "reads information from candidates' resumes and automatically enters it in the corresponding candidate profile fields", with targets like name, contact email, phone, work experience, education and certificates. Greenhouse describes uploads being "automatically processed, or 'parsed'". What a recruiter first sees, in other words, is the parser's output: the fields, not your layout.

2. What the vendors say breaks parsing

Greenhouse's documentation on unsuccessful parses is the most concrete. It names tables, headers and footers, multi-column layouts, contact information placed in a header, footer or text box, graphics, image-only files, and large files as causes of failed or partial parses. Workday's admin guide says results "can vary based on resume format and order of words" and warns against image-based styling. Two practical conclusions follow, both grounded in vendor text rather than folklore: keep contact details in the body of the document, and expect a two-column or table-heavy layout to be read out of order.

3. What is and is not filled automatically

Workday's documentation notes that Skills and Languages are not auto-filled, and that parsing populates fields with varying reliability. SAP's configuration maps experience, education and certificates. The honest summary: contact and history fields are routinely extracted; skills are the field most likely to come out empty or wrong, which is why a skills section written in the posting's own terms matters.

4. Keyword screening: what documentation does and does not say

Here we hedge deliberately. No primary product documentation we could find describes automatic keyword-based rejection. Vendor help docs frame parsed data as something recruiters then review. Vendor blog content (Jobvite's candidate guide) does advise that keywords influence how a resume ranks for a role, and recruiters do search their own databases with keywords. So the defensible claim is narrower than the folklore: matching the posting's genuine terms helps a human find and understand your parsed profile; the popular claim that software auto-rejects every resume below a keyword threshold is not something any vendor's product documentation states.

5. What CareerForge does with this

The free ATS resume checker looks for the formatting habits this documentation describes: contact details in the body, consistent dates, standard headings, single-column text that survives extraction. It is CareerForge's own analysis, not a reproduction of any vendor's parser. The methodology page states exactly how the score is computed.

Sources

• Workday (admin guide, last updated 2023-06-23)

  "Resume parsing populates fields from a resume." Parsing results "can vary based on resume format and order of words." The guide tells administrators to expect better results without image-based styling, and notes that Skills and Languages are not auto-filled.

• Greenhouse support documentation

  Lists the layouts that cause failed or partial parses: tables, headers and footers, multi-column layouts, contact information inside a header, footer or text box, graphics, image-only uploads, and files over 2.5 MB.

• Greenhouse support documentation (candidate uploads)

  An uploaded resume is "automatically processed, or 'parsed'", and "embedded images or other resume layouts may cause the candidate's information to be uploaded incorrectly."

• SAP SuccessFactors Recruiting (help portal, 2H 2026 version)

  Parsing "reads information from candidates' resumes and automatically enters it in the corresponding candidate profile fields." The configuration documentation maps parsed targets including name, contact email, cell phone, address, work experience, education and certificates.

• iCIMS (vendor blog)

  Parsing "extracts data" from a resume and "converts it in a structured way, often self-completing a form", identifying contact details, language levels and years of experience.

Vendors change their documentation; if a link has moved, the substance above is what each page said when last checked (October 2026).