File syncing across devices

Every company I have ever worked with has the same file. It is called something like offer_final_v3_FINAL_corrected.xlsx and there are four of them, in three folders, on two computers, and nobody is entirely sure which one went to the client.

That file is not a discipline problem. It is what happens when people copy files manually, because manual copying has no way of knowing which version is newer and no way of merging two people's work. The fix is not a naming convention that everyone promises to follow. The fix is removing the manual copy from the process entirely.

What synchronisation actually does

A sync tool watches a folder on each device and keeps them the same. Change a file on the desktop, it appears changed on the laptop and the phone within seconds. Delete something in one place, it goes everywhere. Add a new device and it receives everything.

That second sentence is the one worth reading twice, and I will come back to it, because it is the source of every disaster in this area.

Underneath, the mechanism is simple in outline. Each device keeps a record of what it has seen — file, size, modification time, and a fingerprint of the contents. When something changes, it tells the others what changed, and only the difference travels. This is why syncing a large file after a small edit is fast, and it is why a sync folder is not the same thing as copying a directory over and over.

Two ways to build it, and what changes

The first arrangement puts a server in the middle. Every device talks to one machine, which holds the authoritative copy — that is what a file server with a desktop client is doing. The advantage is that there is always one place where everything exists, which makes backups obvious and makes sharing with people outside your household straightforward.

The second has no middle at all. Devices find each other and exchange changes directly — your laptop and your phone sync when they are on the same network, or over the internet through an encrypted connection, with no server holding a copy. This is remarkable when it fits: nothing to maintain, nothing to pay for, and your files are never at rest on a machine you do not use.

The trade is that with no central copy, a file only exists where a device has it. If two devices are rarely on at the same time, they drift. And there is no obvious single thing to back up, which in practice means people do not.

I run the first at work and the second for a couple of personal folders, and the deciding question is always whether there needs to be an authoritative copy that exists regardless of which devices are switched on. For a company, yes, always. For a folder of notes that lives on two machines I use daily, no.

The sentence that ruins people

Here is the single most important thing in this article, and it is the thing almost nobody understands until it costs them something.

Synchronisation is not backup. They are close to opposites.

A backup preserves a past state so you can return to it. Synchronisation propagates the present state so every device matches. Which means that when you delete a folder by mistake, or a piece of malware encrypts your documents, or a file is corrupted by a failing drive — synchronisation does its job perfectly and faithfully distributes that to every device you own, in seconds, including the one you were counting on as your safety net.

I have watched this happen to somebody. Not a technical person, not careless — they had files on three machines and reasonably concluded they were safe. The deletion took one click and reached all three before anyone noticed. What saved the situation was a versioning feature on the server, which is exactly the thing I am about to tell you to switch on.

Every serious sync system keeps old versions and a recycle bin, and both are configurable. Turn them on, set a retention period you can live with, and — the step everybody skips — verify that you can actually restore something from three weeks ago. Doing that once, deliberately, on a quiet afternoon, is worth more than any amount of confidence.

What to do about the four spreadsheets

One shared folder, synchronised to every machine, with versioning enabled and a backup running from the server to somewhere else. That arrangement removes the conditions that produce duplicate files: there is one location, it is the same on every device, and nobody needs to email anything to themselves to move it between computers.

It does not solve two people editing the same spreadsheet simultaneously — that needs a collaborative editor, which is a different tool for a different problem, and pretending otherwise is how people end up with a file called offer_final_v3_conflict_laptop.xlsx, which is genuinely worse than what they started with.

But for the ordinary case — files that one person edits at a time, on whichever machine they happen to be sitting at — it removes an entire category of daily friction that people have stopped noticing because they have lived with it for years. That is the quiet value of this one. Nobody praises it afterwards. They just stop having the conversation about which version is current.