Guide

How to export an accessible PDF from Excel

Excel's PDF export writes no structure tree on most versions. What that means, what to do instead, and why a spreadsheet is often the better deliverable.

About 11 minutes · moderate

What Excel actually exports

Word and PowerPoint write a structure tree when they export a PDF. Excel, on most versions and on both Windows and macOS, does not — File → Save as PDF and File → Export → Create PDF/XPS both produce a file with no /StructTreeRoot at all.

The result is a document where every cell is text painted at a coordinate and nothing else. It is not a badly tagged table; it is not tagged at all, so a screen reader has nothing but the content stream and reads the cells in whatever order the generator emitted them — typically row by row with no indication of where a row ends, no header association, and no table mode.

Check yours rather than trusting this. Open the export in Acrobat, File → Properties, Description tab, bottom right: Tagged PDF: Yes / No. Any free checker answers the same question in a second.

The three honest options

1. Do not ship a PDF

Consider this first, because for tabular data it is usually right.

A spreadsheet is a better deliverable than a picture of one. A screen-reader user can sort a real table, filter it, navigate by cell with the tools they already know, and extract it into something else. A PDF of the same table is a static rendering that removes every one of those.

Publish the .xlsx, or a CSV alongside it, and make sure the workbook itself is accessible: a single header row, no merged cells, named sheets, no information carried only by fill colour. If a PDF is required for archival or signature reasons, ship it as the secondary format.

2. Route through Word

The fastest reliable path to a tagged PDF, and for anything up to a few pages it is faster than remediating the export.

1. In Excel, select the range including the header row.
2. Copy, and in Word paste as a table (Paste Special → Keep Source Formatting or the table paste option).
3. Select the header row → Table Design → tick Header Row, and on Layout tick Repeat Header Rows if the table runs over a page.
4. Add a document title in File → Info → Title.
5. Export with File → Save As → PDF and Options → Document structure tags for accessibility ticked.

Word writes Table, TR, TH and TD elements with Scope on the headers, which is the structure tagging a table correctly otherwise takes an hour to build by hand.

This does not survive a change to the data — the next month’s numbers mean doing it again — which is why option 1 is better for anything recurring.

3. Tag the PDF

When the layout genuinely has to come from Excel — a print-designed report, a form, a certificate — you are tagging an untagged file, and that guide is the process. Two things specific to spreadsheets are worth knowing before you start.

Auto-tagging builds tables from ruled lines. If your table uses cell borders, Acrobat’s autotagger will usually find a table. If it uses fill colour or whitespace alignment instead, it will find paragraphs, and you will be building the table structure by hand.

Header repeats become data rows. Excel’s Print Titles repeats the header row at the top of each page. Visually that is correct; structurally each repeat is another set of cells, and unless they are tagged as TH a reader announces them mid-table as if they were data. Decide which they are and tag them accordingly — it is condition 15-003 territory.

Prepare the workbook before exporting, whichever route

These cost nothing at authoring time and remove most of the downstream work.

Unmerge everything you can. Merged cells are the single most damaging thing in an exported table. They break the grid model a screen reader builds, and a table containing them cannot express its header relationships with Scope alone — every data cell then needs explicit Headers references, which is one dialog per cell in Acrobat. Centre across selection instead of merging where it is only visual.

One header row, at the top. Stacked header rows are legal and expensive. If the table genuinely has two levels of heading, consider whether it is two tables.

Name every sheet. Sheet names become the only landmark a reader has for where it is in a multi-sheet export.

Do not carry meaning in fill colour alone. Red cells meaning overdue are invisible to a screen reader and to anyone who cannot distinguish the hue. Add a column, or a symbol, or text.

Use a real header row, not a bold first row. Excel’s Format as Table marks headers structurally; bold text does not.

Give every chart a description. A chart exported to PDF is an image. It needs alt text that carries what it shows, not “Chart 1” — and writing that is its own skill. Where the chart summarises data that is in the sheet, say so and point at it.

Set the document title. File → Info → Title in Excel, or in Word if you route through it. Without it the reader announces the filename (condition 07-001).

Check the result

Whichever route you took, check the file before it goes out:

• Is it tagged at all? The one-line answer above.
• Do the tables have headers? 15-002 is the condition; a free checker reports it.
• Is the reading order right? For a wide table split across pages, this is where it goes wrong.
• Listen to it. Enter table mode in a screen reader and move cell to cell. If the headers are not announced with the cells, the associations are missing regardless of what a checker said. A test script.

Check a PDF free → · Tag a table correctly → · The whole process →

Related failure conditions

• 01-005 Content is neither marked as Artifact nor tagged as real content Every text-showing, image-painting and path-painting operator on a page must sit either inside a marked-content sequence that maps to the structure tree, or inside an artifact sequence. Content that is in neither is orphaned.
• 15-002 A column has a header cell, but that header cell is not tagged as a header The top row labels the columns but is tagged as ordinary data cells.
• 15-003 A TH cell does not contain a Scope attribute Header cells must declare whether they head a row, a column, or both, unless the table instead uses explicit /Headers ID references on every data cell.
• 09-004 A table-related structure element does not conform to ISO 32000-1, Table 337 The table grammar is broken: a TD outside a TR, a TR outside Table/THead/TBody/TFoot, a Table with no rows.
• 09-001 Tags are not in logical reading order The order of elements in the structure tree does not match the order a human would read the page. This is the single most common substantive defect in tagged PDFs and the one Acrobat's own tooling handles worst.

Last updated 2026-09-12 .