Workday resume parsing, quoted from Workday

Workday is the only major vendor on this site that publishes parsing guidelines in full. Its developer documentation tells you, in writing, how to write a resume its parser reads cleanly. Below is that text, in full.

What Workday's parser returns

Workday's Resume REST API scans a file and returns a fixed set of sections. Workday lists them as contact, education, experience, references and summary.

That list is the whole surface. If your experience is in a section called "Where I've Made an Impact", Workday has no field for it. If your languages are in a section called "Sprachen", there is no field for that either — see the next section, which says so explicitly.

The two fields Workday will not fill for you

"Resume parsing populates fields from a resume, and Workday enables you to review the resume data. Workday doesn't auto-fill fields you configure as hidden or these fields: Languages, Skills."

Workday Documentation, Concept: Resume Parsing

Read that again, because it is unusual and it is frequently misunderstood. In Workday, your skills are not extracted from your resume. Neither are your languages. Whatever a recruiter sees in those fields, they typed, imported, or selected. It also means any field the employer has configured as hidden never gets filled, silently.

What Workday says about quality

The admin guide is blunt about the limits:

• "Resume parsing results can vary based on resume format and order of words."
• "For best results, use resumes that don't have images or image-based styles."
• Employers can "configure job application templates to hide resume parsing for external candidates", in which case nothing is parsed at all.

That last point deserves emphasis. On some Workday careers sites there is no parse, and a well-formatted resume will make no measurable difference. You cannot tell from the outside which configuration you are in.

The published parsing guidelines, in full

Workday's Resume REST API documentation includes a section titled "Resume Guidelines for Improved Parsing Results". It opens with the reason the guidelines exist, then lists the rules. Both parts are below.

"Unstructured resumes with inconsistent layout and format can lead to inconsistent results from the resume parser. Consider these guidelines when using the Resume REST API:"

Layout

• "Avoid personal information in the header and footer sections, such as name, phone and address."
• "Avoid visual objects such as images, text boxes, clip arts, logos, and annotations."
• "Avoid tabular and column formats."

Structure

• "Provide proper headings for resume sections, such as EDUCATION and EXPERIENCE."
• "Don't combine multiple sections under a single heading such as EDUCATION and EXPERIENCE."
• "Insert additional new lines between work history job profiles or education."
• "To clearly separate job profiles and school sections, add bullets or numbers at the start of each job profile or education."
• "Write the description for each job in a single paragraph."
• "Avoid combining positions based on employers. Separate each job for better parsing."

Name

• "For better name parsing, always start writing names with the hint 'Name: Xxxx Yyyy'."
• "Always mention names in separate lines."
• "Place the candidate's name at the top of the document, which enables the resume parser to quickly identify the candidate."
• "Multi-word names should always start with either Initial Capital letters or FULL CAPITAL letters."
• "Avoid writing names in all lower case letters."
• "Writing contact information at the starting of resumes results in better identification."

Labelled hints

Workday gives literal examples of the labels it looks for:

• Name: Xxxx Yyyy
• Address: xxxx yyyy zzzz, abcd, efgh, 00000.
• Mobile Phone: 123 456 7890
• Email: xxxx@yyyy.com
• Date: Jan 2015 – Present

And for the last section: "Write the reference section at the end of the document with the proper hint REFERENCE:."

Accepted file formats

Source: Workday Developers — Resume REST API, "Scanning Resumes".

What Workday does not document

• No size limit. The Resume REST API documentation states formats and conditions but no maximum file size. The 2.5 MB number people attribute to Workday belongs to Greenhouse.
• No accuracy figures, and no guarantee. "Results can vary" is the whole of it.
• No ordering rule. "Order of words" affects results, but Workday does not say which order.
• Nothing about the careers site you are actually on. Each employer configures which fields are hidden and whether external candidates see the parse at all.

The version of this that fits on a checklist

1. Export a DOCX or a text-based PDF. Never an image-only PDF.
2. No images, logos, text boxes, clip art, annotations, columns or tables — anywhere, including header and footer.
3. Your name on the first line, in initial capitals, on its own line, followed by your contact details on the lines below it.
4. Standard headings. EXPERIENCE on one line, EDUCATION on another. Never EDUCATION AND EXPERIENCE.
5. Each job a separate block, separated by a blank line, and each block written as one paragraph rather than a series of fragments.
6. One position per block, even if you held two jobs at the same employer. Workday explicitly says not to combine them.
7. Do not expect your skills to be read out of the document. On Workday that field is filled by a person.

Every template in the gallery satisfies items 1 to 6. The classic template and the print-safe template come closest to the literal Workday guidelines, because both use plain upper-case section headings and a name at the top of the body.

• Workday sources Resume REST API — accepted formats, returned sections, and the full "Resume Guidelines for Improved Parsing Results" list
  Concept: Resume Parsing — the Languages and Skills exclusion, and the image guidance