How to convert a PDF to Word
Two reliable ways to do it, and an honest explanation of why the result sometimes arrives broken.
Converting a PDF to Word is a two-minute job that works beautifully most of the time and then, on one document in a handful, produces something unusable: text you cannot click into, a table that behaves like a picture, an extra blank page in the middle of chapter three.
That inconsistency is not bad luck, and it is not a sign that you picked the wrong tool. It comes from the file format itself. Understanding the reason takes about four minutes and it will save you from fighting documents that were never going to convert cleanly.
What a PDF actually stores
Here is the single fact that explains almost every PDF to Word problem: a PDF records where marks go on a page, not what those marks mean.
When you look at a PDF you see headings, paragraphs, a table, a footnote. The file contains none of those things. It contains instructions along the lines of "draw the characters Quarterly in 14pt Helvetica starting at coordinate 72, 690", followed by several hundred more instructions just like it. There is no paragraph object. There is no sentence. There is nothing recording that one run of text is a heading and the next is body copy, or that these nine pieces of text belong to a table, or even which piece of text a reader is supposed to encounter first.
This is by design. PDF was created to make a document look identical everywhere, on any screen and any printer, forever. Throwing away structure and keeping only exact positions is exactly how it achieves that. The format is doing its job well. It is simply the opposite of what Word needs.
Word is a format built around meaning. It stores a paragraph as a paragraph, a table as a grid of cells with relationships between them, a heading as a heading. That is what lets text rewrap when you type, and lets you drag a column wider and have the whole table respond.
Converting a PDF to Word is not a translation between two formats that hold the same information. It is a reconstruction. The converter reads a set of positions and works backwards to a guess about what the author originally meant.
When the guess is right, you get a document you can genuinely edit. When it is wrong, you get something that still looks correct and behaves like a photograph.
Every remaining section of this article is a consequence of that one sentence.
Check your PDF first
Before choosing a method, find out which kind of PDF you have. It takes five seconds and it determines everything that follows.
Open the PDF in any viewer and try to select a sentence by dragging your cursor across it.
Born-digital
The PDF was exported from a real document, and genuine text is inside it. Everything in this article applies. Conversion will work, and the quality of the result depends on how cleanly the original was laid out.
Scanned
The page is a photograph. There is no text in the file to convert, so it has to be recognised from the picture first. That is optical character recognition, and it is a different job with different failure modes.
If your document is a scan, the methods below will still open it, but Word in particular will hand you an image sitting in a document rather than editable text. Read converting a scanned PDF to Word instead, which covers what OCR can and cannot do.
One more thing worth checking while you are here: if the PDF is a filled-in form, expect a harder time than the layout suggests. Form field values are stored separately from the page they appear on, and a converter has to merge two different layers to produce something sensible. A blank form usually converts more predictably than the same form with answers typed into it.
Converting with Microsoft Word
Word has been able to open PDFs directly since Word 2013, and it is the most direct route if you already have it. There is no separate converter to install and nothing is uploaded anywhere.
Opening a PDF in Word
- In Word, go to File, then Open, then Browse.
- Set the file type filter to All Files or PDF Files, otherwise the PDF will not be listed.
- Select the PDF and click Open. Word displays a notice explaining that it is about to convert the PDF into an editable document, and that the result may not look exactly like the original. Click OK.
- Wait while Word rebuilds the document. A long or graphics-heavy file takes noticeably longer than a plain one.
- Go to File, then Save As, and choose Word Document (.docx).
Do not skip step 5. Until you save, the open document is still associated with the original .pdf filename, which is how people end up believing their edits vanished. Saving as .docx keeps the editable version separate from the PDF you started with.
That notice Word shows in step 3 is not boilerplate. It is the software telling you, accurately, that it is about to make inferences. Word for Mac works the same way through the same menu path.
Converting with Google Docs
Google Docs converts PDFs from any browser, on any operating system, at no cost. It is the usual answer for anyone without a Word licence, and it is the practical route on a Mac, a Chromebook or a Linux machine.
Opening a PDF in Google Docs
- Upload the PDF to Google Drive by dragging it into a Drive window.
- Right-click the uploaded file and choose Open with, then Google Docs.
- Wait for the converted document to open in a new tab.
- To get a Word file out of it, choose File, then Download, then Microsoft Word (.docx).
Layout is simplified quite aggressively. Google Docs prioritises getting editable text out over reproducing the page. Multi-column layouts frequently arrive as a single column, and precise positioning is generally not preserved. When the text matters more than the layout, that is a feature. When you needed the layout, it is not.
One genuine advantage: Google Docs runs optical character recognition automatically when the PDF turns out to be a scan, so it will produce editable text from a scanned document where Word will not.
The tradeoff is that your document is uploaded to Google Drive to be converted, which leads directly to the next section.
Online converters and confidential files
Search for a PDF to Word converter and you will find a great many web-based services. They are convenient, they need nothing installed, and for a lot of documents they are entirely reasonable.
The question worth asking before using one is not about conversion quality. It is about what happens to the file.
An online converter works by uploading your document to a server you do not control, processing it there and holding the result until you download it. For a restaurant menu or a public brochure that is irrelevant. For these, it is worth pausing over:
- Contracts under negotiation, and anything containing commercial terms
- Medical records, or any document covered by a privacy regime such as HIPAA
- Client files held under professional privilege or a confidentiality obligation
- Employee records, payroll and anything with identifying personal data
- Unpublished financial results and anything under embargo
Retention policies vary and are not always easy to find. Some services delete uploads within the hour, others are considerably vaguer. If a document falls into one of the categories above, the safe assumption is that uploading it counts as disclosing it, and the decision should be made on that basis rather than on convenience.
Word running on your own computer does not transmit the document anywhere, which is worth remembering when the file is sensitive.
Why the result looks wrong
When a conversion disappoints, it is almost always one of three specific failures. Each one is the reconstruction guess going wrong in a particular way, and each one has a tell you can check for.
Text that will not flow
The most disorienting outcome is a document that looks perfect and refuses to behave. You click into a paragraph, type a word, and instead of the text rewrapping, it overflows or vanishes. Delete a line above and nothing below moves up.
What happened is that the converter could not confidently group the positioned text runs into paragraphs, so it fell back on the option that always looks right: it pinned each block of text into a fixed frame at the exact coordinates the PDF specified. A frame preserves appearance perfectly. It also freezes it.
A converter that pins everything will produce a document that is visually flawless and effectively read-only. A converter that flows everything will produce a document you can edit freely that may not match the original page layout. Neither choice is wrong in general, and both are wrong for somebody.
Tables and pictures of tables
Ask what a PDF stores for a table and the answer is unsettling: nothing that identifies it as a table. There are some lines that were drawn, and some text that was placed near those lines. No cell, no row, no column, no record connecting any piece of text to any region.
So a converter has to look at the drawn lines, look at where the text sits, and infer the grid. When it succeeds you get a real Word table: click in a cell and it behaves like a cell, drag a boundary and the column resizes. When it fails you get the lines redrawn as shapes with text boxes floating over them. On screen the two are indistinguishable. The difference appears the moment you try to add a row.
Borderless tables are the hardest case of all, because the drawn lines a converter relies on are simply not there. Column alignment has to be inferred purely from where text sits, and a column of right-aligned figures next to a column of left-aligned labels can look, positionally, a lot like one wide column.
A page count that changes
Converted documents routinely arrive one or two pages longer than the original, or with a blank page wedged into the middle. This one is not really a bug.
A PDF page is a fixed canvas: content was placed onto it at set coordinates and it cannot move. A Word page is the opposite. It is simply however much text happens to fit before the text runs out of room, which means that once the text is flowing again, every small difference compounds down the document.
Font substitution is the usual culprit. If the PDF was set in a typeface that is not installed on your machine, the converter picks the closest available match. A replacement that is fractionally wider changes where every single line breaks, which changes where every paragraph ends, which eventually pushes a line onto a new page. From there the whole document shifts.
The other frequent cause is a converter turning the fixed spacing at the bottom of a PDF page into an explicit page break in Word. Do that on every page and any text that has reflowed shorter leaves a blank page behind it. If your converted document has stray empty pages, turn on paragraph marks and look for manual page breaks. Removing them usually resolves it.
What survives and what does not
Given all of the above, it is possible to predict fairly reliably how a given document will convert before you start.
Usually converts cleanly
- Single-column text documents: letters, reports, essays, manuscripts
- PDFs exported directly from Word in the first place, which often carry structural hints the converter can read
- Documents using widely installed fonts, where nothing has to be substituted
- Simple tables with visible ruled lines and one line of text per cell
- Straightforward numbered and bulleted lists
Converts, but expect to do some repair
- Multi-column layouts, where reading order has to be inferred from position
- Documents with sidebars, pull quotes or text wrapped around images
- Tables without ruled lines, or with cells spanning several columns
- Headers, footers and footnotes, which are frequently converted into ordinary body text
- Anything set in an unusual or embedded typeface
Rarely converts to something you would want to edit
- Scanned documents without an OCR step: there is no text in the file to recover
- Design-led layouts such as brochures, magazine spreads and annual report pages
- Technical drawings, charts and diagrams, which are vector artwork rather than text
- Filled-in forms, where field values live on a separate layer from the page
- Documents in a script the converter cannot map back to characters correctly
If your document is in the third group, converting it is often slower than rebuilding the parts you need by hand. Recognising that early is worth more than any amount of tool selection.