DOCX to PDF: Why the Layout Moves, and the Checks That Catch It

A DOCX does not contain pages. It contains text with styles, images anchored to paragraphs, and settings that say how to flow everything — and a program lays that out into pages each time it is opened or printed. Microsoft Word and the office engines used by converters and servers are different programs with different line-breaking rules, hyphenation dictionaries and font-metric tables, so identical input can paginate differently.

That difference is normal and usually small: the same paragraph count, the same fonts, one page break in a different place near the end. It becomes a problem only when a document is tuned to the last line of the last page, which is exactly what happens with a résumé, a two-page brief or a form that must not run to a third page. Convert early enough to see the overflow, not at the moment you send it.

If the converting machine does not have the font, it substitutes one. Office suites ship metric-compatible replacements — Calibri maps to Carlito, Arial to Liberation Sans, Times New Roman to Liberation Serif — which are designed to occupy the same space per character, so the layout mostly holds. Custom fonts have no such twin: a corporate typeface becomes whatever the fallback list says, character widths change, and every line on every page reflows.

If the document uses a brand font, either embed it in the DOCX before converting or accept that a local export from Word will look different from a conversion done elsewhere. And check the page count first thing: if it differs from the original, something reflowed, and the cause is almost always a font.

Tracked changes and comments are the quiet hazard. In Word they are visible as marks; a converted PDF may show them, may show only the accepted text, or may omit the comment margin entirely. Decide which you want — clean copy or review copy — and resolve the tracked changes before converting so the answer is not up to the tool.

Fields are the second category: a table of contents, page numbers, cross-references and date fields are computed by the editing program. If they were stale when you converted, they are frozen stale in the PDF. Update them first. Beyond that, the usual casualties are tables whose rows split across a page break, floating text boxes and shapes that anchor slightly differently, hyperlinks that survive or do not depending on the converter, form controls that become flat, embedded spreadsheets that need their host program, and images using transparency or blend modes that render differently on a non-Microsoft engine.

Compare page counts first, then scan the bottom of each page for a paragraph that used to sit on the previous one. Look at any table that lands near a page break, because a header row separated from its rows is the most common visible defect. Confirm that headers, footers and page numbers are right, click two or three hyperlinks, and check that images have not gone soft — a re-compressed photograph is the one thing in the file that quietly loses quality.

If the PDF is for archival or legal purposes, ask for PDF/A rather than an ordinary PDF: it embeds the fonts and forbids the features that make a document render differently years later. The trade is that PDF/A is a final form — you cannot go back to editing it, so keep the DOCX as the master and treat every PDF as an export.