Why PDF-to-Word conversion is imperfect by design
PDF is a fixed-layout format — it describes exactly where each character appears on the page as absolute coordinates. Word (.docx) is a flow layout — text reflows based on margins, font size, and styles. Converting between them requires inferring structure: which groups of characters form a paragraph, which are headings, which text belongs to a table cell. This inference is imperfect for complex layouts and fails completely for scanned PDFs (which are just images with no text layer at all).
What converts well and what doesn't
| Content type | Conversion quality | Notes |
|---|---|---|
| Plain body text | Good | Paragraphs and line breaks usually preserved |
| Simple headings | Good | Detected from font size differences |
| Numbered/bulleted lists | Moderate | Sometimes collapses to plain paragraphs |
| Simple tables | Moderate | Cell boundaries often misidentified in complex tables |
| Multi-column layouts | Poor | Columns frequently merge into single-column output |
| Headers and footers | Poor | Often appear as body text at top/bottom of pages |
| Embedded images | Good | Usually extracted and placed inline |
| Mathematical formulas | Poor | Rendered as images or garbled text |
| Scanned PDFs (no text layer) | Fails | Requires OCR — use a separate OCR tool first |
When to retype instead of converting
If your PDF has complex multi-column layouts, tables with merged cells, or heavy use of text boxes and shapes, the conversion output will require more cleanup time than retyping the relevant sections from scratch. A practical threshold: if the output needs more than 20 minutes of formatting fixes, manual reentry is faster and produces cleaner Word structure for future editing.
For scanned PDFs (photographed or printed-then-scanned documents), you need OCR (Optical Character Recognition) before conversion. Google Docs can open a scanned PDF and run OCR automatically — upload the PDF to Drive, right-click, open with Google Docs.
What PDF elements survive conversion to Word — and what doesn't
| PDF element | Converts to Word? | Notes |
|---|---|---|
| Plain paragraphs | ✓ Fully | Text, font size, and basic styling preserved |
| Bold / italic / underline | ✓ Usually | Preserved if embedded font info is available in the PDF |
| Tables | ✓ Mostly | Simple tables convert well; complex merged cells may need manual cleanup |
| Multi-column layout | ~ Partial | Columns are often extracted as separate text boxes or inline text |
| Embedded images | ✓ Yes | Images are extracted and placed inline in the DOCX |
| Hyperlinks | ✓ Usually | Clickable links preserved in output DOCX |
| Headers and footers | ~ Partial | Content extracted but position may differ |
| Exact fonts (non-standard) | ~ Partial | Substituted with closest available font if not embedded |
| Scanned text (image PDF) | ✗ Not without OCR | Scanned PDFs contain images, not text — OCR is required first |
| Forms (fillable fields) | ~ Partial | Field content extracted but interactivity lost |
| Annotations / comments | ✗ No | PDF annotations are not transferred to DOCX |