The story we hear from photographers is almost always the same. There was a backup. It ran automatically, it had run for years, and then one quiet month it stopped - a full disk, a lapsed subscription, a cable knocked loose - and nobody noticed, because a backup that fails silently looks exactly like a backup that works. The gap only becomes real when a card corrupts or a laptop is stolen, and by then March was six months ago.

The heart of the problem is that a working photographer's whole business lives as files. The edited galleries, the RAW originals, the signed model releases and contracts, the client contact list: lose the drive and you have not lost photos, you have lost the deliverables you are contracted to provide and the paperwork that proves you were allowed to take them. Four hundred gigabytes in one location is not a library, it is a single point of failure.

The rule worth adopting is three-two-one: three copies of anything you care about, on two different kinds of storage, with one of them off-site. In practice that is the working drive, a second local drive it copies to, and a cloud or off-site copy that survives a fire or a burglary taking both machines at once. Cloud backup is the off-site leg most photographers are missing, and it is the one that matters when the loss is physical.

Backups are only real if you test restores. Once a quarter, pick one job at random and actually pull it back from each copy - open the files, confirm the releases are there, check the folder structure came with it. This is also where the silent failure gets caught, because a restore test fails loudly when the underlying backup has quietly stopped.

There is a data-protection edge worth naming. Model releases and client details are personal data, and if your only copy is lost or stolen you may have both a business problem and a reporting duty. Keeping the releases inside the same versioned, backed-up structure as the images, rather than loose on a desktop, means the paperwork survives the same disaster the photos do.