Why the Markdown intermediate path is cleaner than direct Word import

Confluence's storage format (XHTML) and Word's OOXML diverge in non-obvious ways. The native importer makes guesses about the mapping, and the guesses fail often enough on real-world documents to require manual cleanup on every import. Markdown, by contrast, is a constrained intermediate: it doesn't carry Word's style baggage, and Confluence's Markdown handling (via "Insert Markup" or paste) is mature.

Convert each Word document on Word to Markdown, copy the Markdown, in Confluence use Insert > Markup > Markdown (or just paste — newer Confluence versions auto-detect). The resulting page mirrors the source document's structure with significantly fewer fixups.

Honest scope note for mass migrations

The web tool converts one document at a time — the right surface for migrating 10-50 important documents progressively. For corporate-scale migrations of thousands of Word documents, the honest enterprise workflows are: (a) Pandoc locally in a shell loop for the .md generation, then Confluence's REST API to upload programmatically, (b) assigning the migration to a small team that converts and uploads documents in a structured queue over weeks, or (c) Atlassian-marketplace migration tools built specifically for the use case. The web tool plus manual paste is the right surface for progressive migration; large-scale migration needs different infrastructure.

Multi-source wiki patterns

Pair with PDF for Confluence for archived specs and contracts, URL for Confluence for clipped reference material, Audio for Confluence for meeting transcript pages, Video for Confluence for recorded session pages.