You need to change a PDF and the original Word file is long gone. Converting it back to Word is the obvious move, and for a lot of documents it works well. For some it produces something you'll spend half an hour tidying.
Which one you get depends mostly on how the PDF was made. This guide explains why, so you can tell in advance whether converting is worth it, and what to check afterwards.
A PDF doesn't know what a paragraph is
This is the root of every conversion problem, so it's worth a minute.
A Word document stores structure. It knows this line is a heading, these sentences form a paragraph, and this grid is a table with three columns. Change the page size and Word reflows everything.
A PDF stores almost none of that. It stores instructions like "draw these characters in this font at this exact position on the page". A paragraph is just a set of lines that happen to sit close together. A table is often text placed in a grid plus some lines drawn around it. Two columns are two stacks of lines side by side, and nothing in the file says you should read the left one first.
So any converter, from Microsoft Word's own PDF import to Adobe Acrobat to ours, has to look at the positions and guess the structure back. When the layout is simple, the guesses are good. When it isn't, they aren't.
What PDF to Word on EddiesPDF actually produces
PDF to Word rebuilds the document as flowing, editable content. It detects headings, paragraphs and tables, and on multi-column pages it puts the text in reading order (down the first column, then the second) rather than reading straight across.
What it doesn't try to do is reproduce the exact visual layout. Some converters place every line in its own text box at its exact position. The result looks identical, and it's miserable to edit, because adding a sentence doesn't push anything down. I'd rather get a document that types like a normal Word file and fix a few details than one that looks perfect and fights every edit. That's the trade-off we made, so expect some cleanup.
To run it:
- Open PDF to Word and add your PDF.
- Leave "Pages" empty for the whole document, or type a range like 1-3.
- Decide whether to keep "Include images" ticked.
- Click Convert to Word and download the .docx.
- Open it in Microsoft Word, Google Docs or LibreOffice and check it against the PDF.
What usually converts well, and what needs work
These tend to come out clean:
- Letters, memos and reports that were written in Word or Google Docs and exported to PDF.
- Single-column text with headings.
- Simple tables with clear rows and columns.
- Two-column articles, as long as the columns are clearly separated.
These usually need cleanup:
- Complex layouts: brochures, newsletters, magazine pages with text wrapped around pictures, sidebars and pull quotes.
- Forms. The printed labels and boxes come across as text and lines, but the fillable fields don't carry over as fields you can fill in.
- Charts and diagrams drawn as shapes. In the PDF they're a pile of lines, rectangles and text labels, and they may need redrawing or replacing with a screenshot.
- Text that's part of an image, like a logo with the company name in it, or a screenshot of a table. It stays an image.
- Tables with merged cells, no borders, or cells that span pages.
If the PDF was designed in InDesign or similar, set your expectations low. If it came out of Word, set them high.
Scanned PDFs have no text in them
If you convert a scanned document and get a Word file full of pictures and no editable text, that's not a bug. A scan is a photo of a page. There are no characters in it for any converter to find.
PDF to Word doesn't do OCR itself. Run the file through OCR PDF first, picking the document's language, so it gets a text layer. Then convert the OCR'd file. Quality depends heavily on the scan: a clean, straight 300 dpi scan of printed text works well; a skewed phone photo of a crumpled receipt less so. The guide to making scanned PDFs searchable goes into this.
Tips that save cleanup time
Convert only what you need. If you only need to rework section 4 of a 60-page report, put that page range in "Pages". You'll have less to check, and you won't be fixing tables you were never going to touch.
Turn off "Include images" if you only want the words. It's ticked by default and places the PDF's pictures in the document where they appear on the page. Unticked, you get text and tables only, the file is smaller and there's nothing to accidentally shove around.
Check numbered lists. A PDF stores the number "3." as plain text, so the converter has to work out whether it's a list. Sometimes you get real Word numbering, sometimes typed numbers, and the two don't renumber the same way when you add an item.
Check tables. Count the columns against the original, look for text that got split across two cells, and look at any table that runs across a page break.
If you only need the raw text (to paste into an email or feed into another tool), PDF to text is simpler and skips all the layout guessing.
Sometimes you don't need Word at all
If the job is fixing a typo, changing a date or updating a phone number, converting the whole document and then exporting it back to PDF is a lot of round trip. Every conversion step is a chance for the layout to shift.
Edit PDF lets you make small changes directly on the PDF, with "Edit text" and "Find and replace". There's a catch you should know about, though: these cover the old text with a white box and write the new text on top. The original text is still in the file underneath, so someone could copy it out. That's fine for fixing a typo in a flyer. It's not fine for removing a salary figure or an account number before sending a document to someone else; for that, read how to redact a PDF properly.
My own cut-off: a few words, edit the PDF; a few paragraphs or more, convert to Word. And whichever way you go, compare the final PDF against the original page by page before you send it, because the bit that moved is never the bit you were looking at.
Also useful: OCR PDF · Edit PDF · PDF to text
