My chapters were lost or reordered when I imported a Word file

How an importer finds chapters

Almost every program that splits a Word file into chapters — Scrivener's Import and Split, ebook converters, formatting tools — looks at paragraph styles and the outline levels they carry, not at how a line looks. A title in Heading 1 is a chapter. A title typed in bold, enlarged, and centered is just a paragraph that happens to be short, and whatever follows it gets swallowed by the chapter before.

Reordering usually has a different cause. When a book arrives as several files, many tools sort them by filename, and a plain alphabetical sort puts Chapter 10 ahead of Chapter 2.

Fixing the Word file first

1. Turn on View › Navigation Pane and look at its Headings list. That list is roughly what an importer will see. Any chapter missing from it is not styled as a heading.
2. Click in each missing title and apply Heading 1 from the Home tab's styles (Ctrl+Alt+1 on Windows). If there are dozens, select one bold title, right-click it, and choose Select Text with Similar Formatting (Windows Word shows it once Keep track of formatting is on in the advanced options). Then apply the style to all of them at once. Check the selection for stray bold lines before you do.
3. Keep Heading 1 for chapters only. Part titles and scene headings belong at other levels, or the importer makes each one a chapter.
4. Put a page break before each chapter with Insert › Page Break (Ctrl+Enter) instead of repeated Enter presses. It does not define a chapter for most importers, but it keeps empty paragraphs out of the result.
5. If you are importing separate chapter files, rename them with zero-padded numbers: 01, 02, and so on up to 10.

In Tabularia

The rule on the import page is short: each Heading 1 opens a chapter, and Heading 2 and below remain headings within that chapter. A custom style Word treats as an outline level counts too. Bold body text does not, so a file with no heading styles arrives as one long chapter. When several files are imported together, every file starts its own chapter, sorted by filename with numbers read as numbers, so Chapter 2 precedes Chapter 10.

What survives is the readable text: current wording, accepted insertions, and comments with unambiguous anchors, which become manuscript annotations. Tracked deletions do not come in. Footnotes do, each one still attached to the word it annotates. Tables and embedded media are among what is left out, and the import review names every omitted category before anything is created.

If a chapter still lands wrong, there is no need to reimport. Put the caret on its title and apply the Chapter Heading block style; the Outline updates at once. The after-import checks are worth running either way.