11 min read
Your PDF looks perfect. You open the Greenhouse preview and Skills sits between your second and third jobs. Education jumped above Experience. A bullet from 2019 now hangs under your current title. You're not crazy, and you're not alone on US corporate reqs. The importer doesn't read your layout. It reads document order, not the path your eyes followed on screen. That's why drag-and-drop fixes in Canva don't stick.
Quick wins
- Paste your PDF into Notepad and read employers top to bottom before re-uploading.
- Rebuild as single-column body text with plain Experience, Education, and Skills headings.
- Put title, company, and Month Year dates on one strip per role.
- Open Greenhouse preview and confirm order before knockout questions lock.
Check your resume for free with the posting pasted in after you fix layout. Most candidates try dragging boxes in Canva or Figma for an hour. Greenhouse doesn't see boxes. It sees a stream of text. When a sidebar loads before a job block in that stream, the portal shows exactly that scramble.
Below: the symptom you'll recognize, three causes behind scrambled imports, how to tell which one hit your file, and the fix for each with before/after pairs you can paste against your export. Job searching's already draining. At least you'll know why the preview betrayed a pretty PDF, and what to rebuild tonight. You won't fix this with a new template subscription. You need single-column body text recruiters can search.
Pull up your last Greenhouse preview screenshot if you have one. We're going to match it to a cause, not guess from generic ATS advice.
The preview scramble you'll recognize after upload
The symptom isn't a slow server. Upload succeeds. The attachment icon looks fine. Then the parsed view shows sections in an order you never designed. Skills might appear after your first employer but before your second. A summary paragraph might land below Education. Dates from one role might stick to the wrong company name.
What candidates expect: the portal respects visual hierarchy. Contact, summary, experience top to bottom, education, skills at the foot. That's how humans read a one-page PDF on a desk.
What the import often does: walk the file as a single text stream, left to right, top to bottom, block by block in DOM order. Sidebars, floating boxes, and table cells enter that stream where they sit in the file structure, not where they sit on the printed page.
A composite digital marketing manager opens preview and sees HubSpot, Google Analytics, and Meta Ads listed between two employer blocks from 2018 and 2020. Her current role at the bottom of the preview shows two bullets while six others vanished into a skills blob. The PDF thumbnail still looks polished. Recruiters filtering on current title won't see those bullets attached to the right employer.
That mismatch is the whole problem. Greenhouse reorders sections relative to your design because it never received your design. It received a parse of text blocks. Until you fix the underlying stream, every re-upload of the same layout reproduces the same scramble.
Edge case: only page one scrambles while page two looks fine. A sidebar on page one dumps once, then body flow returns. Edge case: autofill knockout questions can attach the wrong employer when import order scrambles. Fix layout before you lock answers.
Cause 1: two-column layouts and left sidebars
This is the most common reason Greenhouse reorders sections on US applications. Canva, Figma, Word templates, and Google Docs designs often put Skills, Tools, or Certifications in a left rail beside Experience. Your eyes read the center column first. The importer reads the left rail first because it appears earlier in document order.
Before: Two-column resume. Left rail: Python, SQL, Tableau, AWS. Center: Northwind Health analyst role with four bullets, then prior employer below.
After paste test: Python, SQL, Tableau, AWS, Northwind Health, Analyst, Mar 2022 to Present, bullet text, prior employer. Skills sit between jobs in plain text.
Why dragging boxes fails: visual editors reorder pixels, not DOM. Nudging the skills box on screen does not move its text block in the export stream. The parser still hits the rail mid-file.
The fix direction: rebuild as one column. Put Skills after Experience and Education as a short echo of tools already proven in bullets. No rail. No second column for dates on the right while titles sit left. Title, company, and Month Year dates belong on one line per role.
Read Greenhouse resume upload issues and US fixes that work when upload succeeds but preview stays empty after you flatten columns. Same root cause, different symptom on screen.
Cause 2: text boxes, tables, and icon headers
The second branch is invisible structure. Text boxes in Word, table cells for each job, and icon rows that replace plain headings all fragment the stream. Importers grab the nearest text chunk and map it to a field label they recognize. Chunks that do not match a label float in whatever order they were read.
Before: Operations analyst resume built with one text box per job in Word. Dates live in a right-aligned mini box inside each container.
After import: Greenhouse shows three employers with identical end dates. Bullets from the newest role attach to the oldest company name.
Before: Software engineer resume uses a table: row one is company, row two is title, row three is bullets. A footer table repeats skills in tiny type.
After import: Skills footer imports first. Experience table rows shuffle so Kubernetes and Terraform appear above the employer that actually used them.
Tables feel organized to humans. To a parser they're grids read in row order, not story order. Delete the grid. Use plain lines: Employer | Title | Mar 2021 to Present, then bullets underneath. Icon headers export without the word Experience, so the importer never opens that bucket.
Cause 3: creative headers and merged sections
The third branch is naming and grouping. Greenhouse maps blocks to buckets using heading strings it already knows: Experience, Work History, Employment, Education, Skills. Creative titles break that map. So do merged sections that blend volunteer work, projects, and paid roles under one custom banner.
Before: Section title Where I've Made Impact holds three employers and two volunteer blocks.
After import: Volunteer lines import as jobs. Paid role dates detach. Skills from a subtitle import as a standalone paragraph between blocks.
Before: HR generalist resume labels the main block Professional Story. Skills sit in a footer called What I Bring.
After import: Education rises to the top because it's the only standard label the file contained. Experience field stays empty in recruiter filters.
Split mixed sections. Rename to Experience, Education, and Skills. Move volunteer proof under dated employer lines or a labeled Additional Experience block beneath main jobs, still single column.
How to tell which cause scrambled your file
You don't need to guess from the template brand. Run one paste test and compare it to Greenhouse preview side by side. Export PDF or DOCX, select all, paste into Notepad on Windows or plain TextEdit on Mac. Read top to bottom with formatting stripped.
Column cause: skills or tools appear between two employer names in Notepad. You'll see Python before Acme Corp even though the PDF showed Acme at the top of page one.
Text box or table cause: the same date repeats beside multiple employers, or bullets appear before any company name. Icon-only headers show up as blank lines where Experience should be.
Header cause: Education or Skills is the only recognizable label in the paste, while job bullets float unlabeled. Creative section titles paste as plain text the importer does not bucket.
Five-minute rule: if Notepad order scrambles, Greenhouse will too. If Notepad reads clean but preview scrambles, try DOCX once. Most cases are layout, not format magic.
Open how to preview your resume like an ATS in the US when you want a second check outside the employer portal. Same paste logic, less pressure while the req timer runs.
When Greenhouse reorders sections: fixes that survive re-upload
Match the fix to the cause you diagnosed. All three paths end in the same place: single-column body text, standard headings, Month Year dates on the employer strip, no tables or rails. The rebuild method differs slightly by starting layout.
Fix for column and sidebar layouts
Create a new Word or Google Doc with one column only. Copy employers and bullets from the center column. Paste Skills last as a short list echoing tools already in bullets. Delete the old designed file instead of editing in place.
Before: Canva two-column design. Skills rail on the left. Dates in a right margin column.
After: Single column at 11-point Calibri. Harbor Logistics | Operations Analyst | Jun 2022 to Present on one line. Four bullets underneath. Skills line at bottom: SQL, Looker, Excel, Salesforce.
Fix for text boxes and tables
Strip structure first. Select all, paste unformatted into Notepad, then rebuild in a clean doc from that text. Do not paste back into the templated file. The hidden boxes remain in the old file even when you think you deleted them.
Before: Table layout with one row per job and a skills footer in 8-point type.
After: Plain paragraphs. Beta Corp | Software Engineer | Jan 2023 to Present. Bullets as normal lines. Skills echo after Education, not in a footer grid.
Fix for creative headers and merged sections
Rename sections to Experience, Education, and Skills. Split volunteer proof into its own dated block if it must stay visible. Never bury paid work inside a narrative banner without employer lines.
Before: Professional Story section mixing two jobs and a board role.
After: Experience heading. Acme Retail | Store Manager | Mar 2020 to Present with three bullets. Community Board | Volunteer Treasurer | Jan 2018 to Dec 2019 as a separate labeled block below, still single column.
Copy-paste single-column rebuild skeleton
[Full Name]
[City, ST] · [phone] · [email]
Experience
[Company] | [Job Title] | [Mon YYYY] to [Mon YYYY or Present]
• [Posting must-have in first eight words; metric if honest]
• [Second proof bullet]
[Next employer block, same pattern, reverse chronological]
Education
[Degree], [School], [Mon YYYY]
Skills
[Tools already proven above, comma-separated]
Paste this skeleton into a blank doc before you touch design again. Fill employers first. Run Notepad test. Export PDF for Greenhouse only after order passes. One clean re-upload beats five attempts with the same two-column source file.
When dates scramble even after columns flatten, read how to stop ATS from misreading dates in the US for strip formatting that keeps Month Year attached to the right employer.
Habits that keep sections scrambled on every apply
Re-uploading the same layout after a bad preview. Greenhouse already parsed what it could. A new upload of the same columns reproduces the same order. Change the file, then upload once.
Trusting the PDF thumbnail over parsed fields. Recruiters filter on imported employer and title rows. A pretty thumbnail with empty Experience is invisible in name searches.
Moving boxes in visual editors instead of rebuilding text flow. Pixels move. DOM order does not. Rebuild in a plain doc when paste test fails.
Stuffing keywords into sidebars. Sidebars import early. Keywords detach from the jobs that prove them. Put proof in bullets under dated employers first.
Assuming Greenhouse chose a new order to punish creative design. The portal reflects read order, not taste. Fix the stream and the preview usually stabilizes. I've seen more callbacks after a single-column rebuild than after a fifth re-upload of the same sidebar file.
Test section order before Greenhouse does
Run the free ATS checker with the job description pasted in after your paste test passes. It won't replicate Greenhouse field maps exactly, but it flags missing must-have terms in Experience when Skills looks full on screen.
Rebuilding from a scrambled template? Build a plain single-column resume once, then tailor bullets per req instead of fighting the same sidebar export every week.
When the req also wants a letter, generate a cover letter after preview shows Experience in order. A cover letter does not fix scrambled import on the attachment.
Rebuild order before you re-upload
Greenhouse reorders sections when the import stream does not match the story you designed. Sidebars, tables, and creative headings are the usual triggers on US reqs. The fix is never another hour nudging boxes. It's a single-column rebuild with plain Experience, Education, and Skills labels and Month Year strips recruiters can search.
Export tonight. Paste into Notepad. If skills sit between jobs, flatten the layout before you touch keywords. Re-upload once. Open preview. Only then answer knockout questions. You'll still see silence when reqs close or teams pause. You won't lose as often to a scramble you could have fixed in twenty minutes.
Run a free resume check on the next posting after your paste test passes. Fix order on your side. Let queue timing explain the rest.
Read more
Frequently asked questions
What looks like a deliberate shuffle is usually import order, not a recruiter preference setting. Parsers walk the file in document order. A left sidebar often dumps its text between employer blocks even when your PDF looks fine on screen. Rebuild as single-column body text with standard headings instead of dragging boxes in a visual editor.
Candidates cannot drag sections back into place in the employer portal. Some employers let you edit parsed fields manually, but many reqs lock those boxes after submit. The durable fix is a new export where employers and bullets already appear in reading order before upload. Run a Notepad paste test, then re-upload once.
Greenhouse search runs on imported text fields, not on the PDF thumbnail. When skills land between jobs, keyword hits may attach to the wrong employer row or sit outside Experience entirely. That weakens filters recruiters use for title, company, and duty terms. Order matters because it decides which block owns the proof.
Both formats scramble when the underlying layout uses columns, text boxes, or tables. A plain DOCX from Word can parse cleanly. A designed PDF from Canva can scramble even when it looks polished. Pick the format the posting requests, then judge the paste test result, not the file extension alone.
Export your file, paste all text into Notepad on Windows or plain TextEdit on Mac, and read top to bottom without formatting. Employers should appear before the next employer, with no skills rail between blocks. Then open Greenhouse preview after upload and compare. Matching order means you can focus on keyword gaps instead of layout rescue.