All insights

A copy is a fork

Months after a drive migration, the duplicates in search were the good news. The bad news was what had been written into them.

In July I moved a small consultancy's working files out of the founders' private cloud drives into a shared company drive. We copied instead of moving, on purpose. Nothing would be deleted, the old drive would stay as an archive, and if anything went wrong the original was still there. The path comparison afterwards showed zero missing files out of about 1,800. It looked clean. This week a team member asked why several of her documents showed up twice in search.

What a copy keeps and what it loses

A copied file has the same name and the same content. It has a new identity. Every link, bookmark, shortcut, "recently used" entry and automation that pointed to the file still points to the old one.

Automations are the easy part, because they can be searched. We scanned every automation for old file identifiers, found one spreadsheet that mattered and switched it over on the day. People cannot be searched. Their browser history, pinned tabs and the "recent" list on their phone all lead back into the old drive, and the old drive still accepts edits.

We tried to make that visible. The old top folders got a "[MIGRATED]" prefix, and when it turned out that recent-file views do not show the folder, about 190 individual files got an "[OLD]" prefix as well. It helped. It did not stop anyone from typing into the file they opened.

A copy is a fork. From the moment it exists, each side can change without the other.

Moving keeps the identity

A month later we moved a batch of about 260 forms that were still owned by a private account. This time we moved them instead of copying, and the difference was obvious: same identifier, full history, comments intact, and the connection between each form and its response sheet unbroken. Nothing that pointed to them had to change, because they were still the same objects.

Moving has a precondition that copying does not. In most cloud drives, only the owner can move a file out of their own space into a shared one, so the owner has to be a member of the shared drive for the duration. That is more coordination than running a copy script alone at night, which is exactly why copying is tempting.

The cleanup that hid the problem

The duplicates in search had two causes. A sorting rule during the move had placed some files in more than one location instead of moving them, and the team member's old private folder was still shared with her company account, so search showed both versions. There is a simple way to tell them apart: if your own name appears under the file as owner, you are looking at the old copy.

In the meeting she removed the company account's access to the old folder, and the duplicates disappeared. That fixes search. It does not fix what was written into the old copies since July. Those edits still exist. They are now invisible from the account everyone works in, and nobody has merged them.

What I would do differently

  • Move whenever the tool allows it, even if it means coordinating with every owner. Identity is what makes two people look at the same file.
  • If copying is unavoidable, make the old side read-only on cutover day. A renamed folder informs people. Read-only stops the fork.
  • Before removing access to the old side, list everything modified there after the cutover date. Merge first, then cut access.
  • Treat duplicates as a symptom to trace. Each one points to a place where someone kept working after the move.
Key

A migration by copy is complete on the day nobody can write to the old side anymore, not on the day the new side is full.

Hung Mai

Hung Mai is an independent consultant in Weilmünster, Germany, mostly working remotely. He plans, builds and monitors AI automation for marketing and sales for companies without a digital team. About Hung Mai

Let's build something real.

Let's talkResponse within 24h.