11 min read
Your resume looks perfect in Preview. Workday shows Skills filled in and Experience blank. That gap is usually layout, not missing keywords. Two-column templates and table grids scramble what Greenhouse and Workday extract. The parser doesn't read left-to-right like you do. It reads cells in file order.
You didn't fail to tailor. You're exporting a file where columns and tables resume accuracy dies in the first parse pass. If bullets look fine on screen but the portal preview doesn't, don't rewrite keywords yet. check your resume for free and confirm employer names, titles, and Month Year dates imported in the right blocks. That's the ten-minute test before you touch bullet wording again.
Job searching when portals ghost qualified files is draining. This page maps symptoms after upload, three layout causes behind scrambled text, how to tell which one's yours, and the fix for each. You need a version that parses before keyword work pays off.
When the posting asks for a cover letter after upload, don't paste text from a sidebar the parser skipped. Generate a cover letter from bullets you verified landed under dated Experience lines.
Quick Wins
- Paste your PDF into Notepad. If lines read out of order, flatten layout first.
- Move Skills below Experience in one column. Sidebars parse first and steal field mapping.
- Replace Experience tables with employer lines plus Month Year on the title row.
- Re-upload once and read the portal preview before screening questions.
Symptoms: when the portal sees a different resume than you do
The file on your desk lists four employers in order. The Workday preview shows Python and Excel under Skills, a garbled employer name, and two jobs with dates that do not match titles. Greenhouse keyword match fires on sidebar tools while Experience bullets never attached to the right company. You feel keyword-starved. The parser never got your chronology straight.
Paste-test scramble is the fastest symptom check. Open the exported PDF, select all, copy into Notepad. If Skills lines appear before your name, or job titles sit three lines away from their dates, columns or tables are reordering text. Humans read the visual column. Parsers read extraction order.
Other signals: screening questions ask for years of experience and the number feels wrong because early roles dropped off page two of a merged column. Recruiters open your PDF and it looks fine while their system profile shows half-empty Experience. I've screened stacks of these in Workday and Greenhouse. The pretty template is often the reason the qualified file never matched the req.
This is not a moral failure about choosing Canva. Corporate portals expect a linear spine: contact block, Summary optional, Experience with employer plus Month Year on each role, Education, Skills last. Anything that splits that spine across columns or cells breaks columns and tables resume accuracy before a human opens the queue.
For the paste-test walkthrough, read how to run a resume parsing test step by step . This page is the layout layer that test usually exposes.
Columns and tables resume accuracy: three parser failures in Workday
Scrambled imports usually trace to one of three layout patterns. Most files show a primary cause plus a secondary one. Fix structure before you tune keywords. The pairs below use composite roles only. Swap in your employers and tools.
Cause 1: Two-column sidebar (Skills left, Experience right)
Symptom: Skills and contact blocks populate in the portal. Experience is empty, duplicated, or missing your most recent employer. Paste-test shows twelve skill tokens before your first job title.
Why it breaks: many parsers read down the left column completely, then the right. Your sidebar Skills and certifications export ahead of employer names. Workday maps that first text block to Skills and never pairs bullets with dates on the right.
Before: two-column template with Skills, languages, and tools in a narrow left rail; Experience in the wide right column.
After: single column: name and contact, three-line Summary, Experience with each employer line, Education, Skills as a vertical list at the bottom.
Before: marketing coordinator file where Greenhouse preview lists Canva, HubSpot, and Mailchimp but shows one employer from 2019.
After: same tools listed once under Skills after four dated employer blocks; HubSpot named inside a 2024 bullet about email campaigns.
Fix: rebuild in Word or Google Docs with one column width. Move sidebar content below Experience or weave tools into bullets. Export PDF from that source, not from the design template.
Cause 2: Experience tables (dates in column A, titles in column B)
Symptom: dates attach to the wrong company, two jobs merge into one line, or tenure math looks short because end dates paired with the next employer's title.
Why it breaks: table parsers read row by row but still treat each cell as a separate text island. Greenhouse often concatenates cells left to right within a row, then stacks rows. A date cell plus empty spacer plus title cell becomes a broken string. Multi-page tables drop rows when a page break sits inside the grid.
Before: table with Jan 2021 to Mar 2024 in column one and Senior Analyst, Contoso in column two.
After: Senior Analyst, Contoso · Jan 2021 to Mar 2024 on one line, bullets below, no grid.
Before: warehouse supervisor résumé where Workday shows Mar 2020 beside the wrong logistics company name.
After: each employer as a heading line with Month Year range on the same line, three bullets underneath, blank line, next employer.
Fix: remove tables from Experience entirely. Use bold employer names or plain text headings. Keep Month Year ranges on the same line as title and company. Read how to stop ATS from merging company and title lines when lines still collapse after flattening.
Cause 3: Header, footer, and nested tables for contact blocks
Symptom: email and phone appear twice, city lands in Skills, or the first page header repeats across every extracted page so employer names disappear mid-file.
Why it breaks: contact details trapped in a header table often parse as body text. Nested tables inside a single Experience row split bullets from employer names. Footer skill grids on page two inject keyword tokens without dates, which confuses keyword match in Greenhouse.
Before: name and phone inside a header table; LinkedIn URL in a footer table on each page.
After: contact lines typed once at the top of page one as plain text, no header/footer objects, standard 11-point Calibri or Arial.
Before: RN résumé with licenses in a nested table beside clinical rotations.
After: RN licensure on one line under the name; rotations as dated bullets under Education and clinical roles under Experience.
Fix: strip Word header and footer content into the body once. Delete nested tables. Unmerge cells before export. If paste-test still duplicates contact lines, save as DOCX from a blank template and paste plain text only.
How to tell which cause is yours
Run the Notepad paste test. If Skills lines appear before your first employer, cause 1 is primary. If dates and titles sit on separate lines that do not pair correctly, cause 2 is primary. If contact info repeats or splits across sections, cause 3 is primary.
Then open the portal preview on re-upload. Experience should list employers in reverse chronological order with bullets nested under each block. If preview still fails after a single-column rebuild, check for text boxes and icons layered on top of the grid. Those are a fourth layer, but columns and tables are the usual root.
Copy-paste block: single-column Experience spine
[Name]
[City, ST] · [phone] · [email] · [LinkedIn URL plain text]
SUMMARY (optional, 3 lines max)
EXPERIENCE
Senior Analyst, Contoso · Jan 2021 to Mar 2024
• Built SQL dashboards on 2.4M-row sales table; cut manual reporting 6 hours per week
• Documented ETL specs for finance stakeholders in Confluence
Operations Coordinator, Fabrikam · Jun 2018 to Dec 2020
• Scheduled 18-person warehouse crew across three shifts in UKG
EDUCATION
B.S. Economics, State University · May 2018
SKILLS
SQL, Excel, Tableau, Confluence (list once; tools also named in bullets)
Before and after: graphic designer template vs parser view
Before: colorful two-column résumé with portfolio links in a left icon column; client list in a table on page two.
After: plain single-column PDF for applications; portfolio URL once under contact; client work as dated contract lines with bullets, no table grid.
Before and after: IT résumé skill grid
Before: three-column Skills table listing AWS, Terraform, and Python in separate cells above Experience.
After: Skills line after Education; AWS and Terraform named inside 2023 platform migration bullets with scope metrics.
Edge case: academic CV with publication tables. US corporate Workday reqs still need a one or two-page single-column upload. Move publications to a separate PDF only when the posting allows attachments beyond the résumé.
Edge case: European CV with photo column. Strip photo for US uploads; photos sit inside table cells and can push Experience text out of order. Keep photo version for human networking if local custom expects it.
Edge case: template says ATS-friendly but uses invisible table borders. Turn on paragraph marks in Word. If gridlines appear around Experience, rebuild without them. Marketing labels on templates do not control parser behavior.
Layout patterns that compound column and table damage
Flattening columns fixes most files. These add-ons keep accuracy low even after you remove the obvious grid.
Text boxes floating over a column break. Parsers skip shapes. Your summary inside a box may never export while sidebar skills flood the parse buffer.
Section columns inside one Word section break. Two columns on page one only still reorder page two when the break resets. Use single column for the whole document.
Skill bars and star ratings in table cells. Icons become empty cells or garbage characters in Greenhouse. Replace with a comma list or plain bullets.
Multiple tiny tables for each job. Four two-row tables read as four disconnected fragments. One employer block per role parses cleaner than four micro-grids.
Exporting from design PDF without editing source. Flattening in Acrobat does not restore reading order. Rebuild text in Word or Docs, then export a fresh PDF.
Tweaking keywords before layout. Adding SQL twelve times in a sidebar table does not help if Experience never imported. Structure first, tailoring second.
This will not fix applying to roles you do not qualify for. It stops qualified files from dying because Workday paired your 2022 title with a 2019 end date.
Test the flattened file
After you rebuild single column, upload the same posting to HireFlow's free ATS resume checker and confirm Experience blocks show employers with Month Year ranges before Skills. You are validating parse order, not chasing a keyword score on a broken spine.
Then run build your resume if you want a clean template without hidden tables. Export DOCX, paste-test, then save PDF for portals that require it.
Do this now: Paste-test your current PDF, flatten one column, re-upload to preview, then tailor bullets on the version that parsed.
Flatten before you tailor again
Columns and tables kill resume accuracy because parsers read file order, not design intent. Fix the spine: one column, employer lines with Month Year on the title row, Skills after proof, no Experience grids. Tailoring scrambled text only rearranges words the system already misfiled.
Paste-test tonight. Match your symptom to cause one, two, or three. Re-upload once and read the portal preview. When Experience looks right, then mirror posting language inside dated bullets.
Save a master single-column file and a tailored copy per req. Candidates who stop fighting sidebar templates move faster than those who keep adding keywords to a layout Workday never mapped correctly.
Open the posting you are targeting. Run a free ATS check on the flattened export and confirm columns and tables resume accuracy recovers before you submit again.
Read more
Frequently asked questions
Not for Workday or Greenhouse uploads if you need accurate parsing. Parsers read plain text in extraction order, not your visual layout. A sidebar that looks clean on PDF often exports as Skills lines before employer names, which breaks date-to-title pairing. Keep a designed version for networking PDFs if you want. Submit a single-column export for applications.
Simple one-row contact lines sometimes survive. Multi-row Experience tables, skill grids, and nested tables inside headers fail often. If paste-test into Notepad shows dates on one line and job titles three lines away, the table is the problem. Replace tables with employer lines, Month Year ranges on the same line as the title, and vertical bullets.
Export does not change reading order. A two-column Canva template still extracts left column first in most parsers. Word table borders survive as cell boundaries Greenhouse reads cell by cell. Rebuild in a single column in Word or Google Docs, then export PDF. The fix is structure, not file type alone.
The parser often grabbed sidebar or header table text first and never mapped body rows to Experience fields. Empty Experience with a full Skills block is a classic column-layout symptom. Flatten layout, re-upload, and confirm employer names and Month Year ranges appear in the preview before you answer screening questions.
Often yes when the PDF came from a design tool. A plain single-column DOCX with standard headings usually maps cleaner in Workday. Some postings require PDF only. In that case fix layout first, then export text-based PDF from Word. Scanned image PDFs and design exports fail regardless of column count.