The most common SharePoint migration is also the least useful: copy the file server into a document library, folder for folder, and call it done. Everyone still navigates six levels deep, the same file lives in three folders, and search returns a hundred near-identical results.
SharePoint works best when documents are described by metadata, meaning columns such as customer, document type, and status, instead of by where they happen to sit in a folder tree.
Why folders stop working
A folder can only answer one question. Is this quote filed under the customer, the year, or the salesperson? Whichever you pick, anyone looking by another dimension has to hunt. Folders also hide permissions (it’s hard to tell who can see what), make long paths that break syncing, and give automation nothing to work with.
Metadata answers all of those questions at once. A quote can have Customer = Acme, Year = 2026, Owner = Sales, and Status = Sent, and a view can group or filter on any of them.
A practical library design
- Split libraries by purpose, not by department. Quotes, contracts, drawings, and policies each get a library, because they have different columns, retention rules, and audiences.
- Add a few required columns. Three to five is plenty: typically a type, a customer or project, and a status. More than that, and people stop filling them in.
- Use choice and lookup columns, not free text. A choice column for Document Type, or a lookup to a Customers list, keeps values consistent so filtering works.
- Build views for how people work. “My open quotes”, “Contracts expiring this quarter”, and “Drawings by project” are saved views, not folders.
- Keep folders shallow, if you use them at all. One level (for example, per project) is fine when it matches a real permission boundary.
Let metadata do the work
Once documents are described consistently, the rest of Microsoft 365 can act on them:
- Search and filtering become precise instead of guesswork.
- Power Automate can route a document when its Status changes to Ready for review.
- Retention labels can apply by document type, so contracts are kept for the required period and drafts are cleaned up.
- Copilot and other AI tools return better answers when the content they read is current, labeled, and not duplicated.
Migrating without a big bang
You don’t have to tag ten years of files on day one. Move active content into the new structure and tag it as you go, and keep the old share read-only as an archive for a while. Where you can, set column values automatically from the old folder path during migration, so files arrive already described.
The goal is simple: anyone should be able to find the current version of a document in a few seconds, without knowing where someone else decided to save it.