How Workday reads your resume vs Greenhouse and Taleo

How does Workday parse your resume when you upload it?

Workday parses your resume once, at the moment you upload it, and uses the result to autofill a structured application form — and it's that form, not your file, that recruiters search and filter on later. That single design choice explains most of the frustration people feel with Workday applications. Each employer runs its own separate Workday tenant, which is why you create yet another account every time. The parser then tries to split your document into discrete fields: name, email, phone, location, a repeating block of jobs with employer, title, start date and end date, education with school, degree and field of study, and a flat list of skills.

What it can't confidently map, it quietly discards. Multi-column layouts are the worst offender, because the parser reads across the page and interleaves your left-hand skills sidebar into the middle of a job description. Tables scramble date columns. Contact details placed in the document header or footer frequently vanish, since that region often isn't part of the main text flow the parser walks. Graphics, icons, skill rating bars and text boxes contribute nothing at all. Creative date formats like "Spring '23 – present" or "03.2021–06.2023" regularly fail to resolve into the start and end dates Workday expects, leaving gaps in your work history block.

The single fix that recovers the most ground in Workday is embarrassingly simple: after the autofill runs, read every field it produced and correct it by hand. Most candidates click straight through that review screen. Don't. Retype the job titles Workday truncated, restore the employer names it merged, fix the dates it inverted, and paste your real bullet points into the description boxes if they came through empty. Pair that with a source file that gives the parser a fair chance — one column, conventional headings like Experience and Education, and dates written as MM/YYYY — and the form you end up with actually resembles your career.

What does Greenhouse actually do with your resume file?

Greenhouse treats your resume as a document first and a data source second: it extracts your name, email and phone to create the candidate record, converts the file's text for full-text search, and then puts the original attachment in front of the recruiter. There's no lengthy form to re-key, which is why Greenhouse applications feel so much lighter than Workday ones — often just a file upload, a few questions and a LinkedIn field. Startups and scale-ups favor it for exactly that reason. The trade-off is that your formatting and your wording are doing far more of the work, because a human really will open the file you sent.

That makes one failure mode dominant in Greenhouse: a PDF with no text layer. Resumes exported from design tools as flattened images, scans of printed copies, and files where the text is drawn as vector outlines all look perfect to you and contain zero searchable words to the system. When a recruiter runs a keyword search across a thousand applicants for "Kubernetes" or "revenue operations", you simply don't appear. Headers and footers cause the second-most trouble, since contact details tucked up there can miss the extraction pass and leave a record with a blank phone number that nobody chases.

The fix in Greenhouse is a text-layer check plus keyword honesty. Open your PDF, try to select a sentence with your cursor, and copy it into a blank document. If nothing selects, or the paste comes out as gibberish, re-export from Word or Google Docs rather than from a design tool. Then make sure the exact terms from the job ad appear in the body of your experience bullets, not only in a skills strip — recruiter searches in Greenhouse are literal, and a sidebar tag that says "SQL" carries no context about where or how you used it.

Why does Taleo mangle your resume more than newer systems?

Taleo mangles resumes more than Workday or Greenhouse because it's the oldest architecture of the three and was built around a typed candidate profile, with resume parsing bolted on afterwards as a convenience. Oracle's Taleo still runs behind a large share of banks, airlines, hospital groups, government agencies and legacy enterprises, and many of those tenants are configured much as they were a decade ago. The tell is the application flow itself: a long multi-step wizard, a separate account per employer, and screens that ask you to enter each job manually even after you've already uploaded a file.

Its parser is genuinely less forgiving. Text boxes, shapes and columns are commonly ignored outright rather than merged awkwardly. Tables lose their structure, so a two-column table holding dates on the left and roles on the right can come back as a run-on string. Bullet characters that aren't standard list bullets — arrows, checkmarks, emoji, custom wingding glyphs — sometimes survive as question marks or boxes. Because many Taleo screens are keyword-matched against the profile fields and the extracted text, anything lost in translation is effectively invisible to the recruiter running a search.

The fix for Taleo is to stop treating the upload as the application. Submit a plain, conservative file — a .docx or a text-based PDF with one column, no tables, no text boxes and standard round bullets — then complete every profile field the wizard offers, including the optional ones. Fill in the skills picklist if there is one. Paste your summary into the free-text box rather than leaving it blank. Yes, it takes twenty minutes instead of two. For the kind of employers still running Taleo, that profile is the record a recruiter actually queries, and an empty field is a missed match.

Which ATS do companies use, and how can you tell?

You can identify the ATS in about three seconds by looking at the URL of the application page. Greenhouse applications sit on job-boards or greenhouse.io addresses, Workday ones contain "myworkdayjobs" and a company-specific tenant name, and Taleo pages show "taleo" in the address along with a long numeric requisition ID. Lever, Ashby, SmartRecruiters, iCIMS and SAP SuccessFactors are all equally obvious once you start reading the address bar. This matters because it tells you, before you spend twenty minutes applying, whether your file is about to be turned into a form or handed straight to a human.

The pattern by employer type is fairly reliable. Large multinationals, banks and healthcare systems tend toward Workday or SuccessFactors. Venture-backed tech companies and scale-ups cluster around Greenhouse, Lever and Ashby. Public sector bodies, airlines and older industrials still run Taleo or iCIMS. Recruitment agencies use their own tools entirely, such as Bullhorn. None of these systems auto-rejects you for having the wrong font, whatever you've read — rejection comes from knockout questions about work authorization, notice period or salary, and from a human skimming a list. Parsing failures do something subtler and worse: they make you unfindable.

Once you know which system you're facing, you can adjust your effort sensibly. A Workday application deserves extra minutes on the autofill review screen. A Greenhouse application deserves extra care in the file itself and in the optional cover letter box, because both get read. A Taleo application deserves patience with the profile wizard. What none of them deserve is the same decorative template fired off twenty times. Tailoring here isn't about rewriting your whole resume; it's about knowing where the system will look and making sure the evidence is sitting in that exact spot.

What formatting changes fix the most parsing failures across all three?

Four changes fix more parsing failures than everything else combined: move contact details out of the header and footer into the body of page one, collapse the layout to a single column, replace every table and text box with plain paragraphs and bullets, and write dates as MM/YYYY. Do those four and your file will parse acceptably in Workday, Greenhouse and Taleo alike. Everything else — font choice, margin width, whether you use a horizontal rule — is far down the list and mostly folklore repeated by people who've never watched a parser output side by side with the original.

The column question is the one people argue about hardest, and the argument is settled. A designer-style resume with a shaded sidebar looks sharper on your screen and reads as interleaved nonsense in a form-filling parser, because the extraction follows the text flow rather than the visual grid. If you're attached to a sidebar design, at minimum keep a plain single-column version for applications and save the pretty one for your own website. There's a full walkthrough of what a parser actually sees in a two-column layout if you want to see the before and after for yourself.

Then stop guessing and test. Export your file, paste its extracted text into a blank document and read what comes out: if the job titles have merged, the dates have gone missing or your skills have landed inside a bullet about your 2019 internship, that's precisely what the ATS sees. Faster still, run the file through HRLens's free CV analysis, which checks parsing and ATS compatibility and flags the structural problems — buried contact details, unreadable sections, missing keywords from the role you're targeting — before an employer's system does it silently, and without telling you.

Frequently asked questions