Three Backups, One Restore. The Person Who Will Actually Do It Is Probably Not You

The word backup is doing two jobs at once, and the confusion between them is where most households lose their files. One meaning is a second copy that exists independently of the first, so that damage to the original leaves the copy untouched. The other meaning, the one most people actually have, is synchronization: a folder that mirrors what is on the machine, faithfully and quickly, including the deletion, the encryption by ransomware, and the accidental overwrite. Sync is a convenience feature that borrowed a safety word. The two are priced similarly and behave nothing alike.

Three arrangements, compared by how they give files back

Judge each option by the restore, not the copy, because copying is the part every product does well. A sync folder from a mainstream provider restores individual files easily and recent versions reasonably well, but its retention window is usually measured in weeks, which means a corruption you notice in month three is gone from both sides. A local external drive running image backups restores an entire machine, operating system and settings included, at the speed of a cable rather than a connection, but it sits in the same building as the thing it protects. A hosted backup service keeps versions for longer and lives elsewhere, at the cost of a restore that depends on your upload history being complete.

The honest comparison is therefore not a ranking. It is a division of labor: the local image for speed and whole-machine recovery, the hosted copy for fire, theft and flood, the sync folder for working across devices and nothing more. Each is cheap relative to the loss it addresses, and the combined monthly cost of all three usually lands below what a single data recovery attempt on a failed drive would run. The National Institute of Standards and Technology is the federal body responsible for the cybersecurity framework that most of these products describe themselves against, and its guidance has long treated recovery as a separate function from protection.

Where the copy physically sits, and why that varies by border

Ask a backup provider in Germany where your data rests and you will get a specific answer, often a named facility, because data residency is a selling point shaped by the European Union's general data protection rules and by customers who ask the question as a matter of routine. Ask the same question of a consumer service in the United States and the answer is frequently a region rather than a place, because the assumption here is that location is an operational detail. Canadian providers sit in between, with provincial rules pushing some public sector data to stay in country. None of this makes the American product worse at restoring a file. It does mean that a right of export, which European customers are accustomed to exercising on demand, is something you should confirm in writing before you commit years of records to a vendor.

The practical version of that question is simple. If you stopped paying tomorrow, how long would the copy remain, and in what format could you take it with you? Services that hand back a structured archive, readable without their software, are doing something a European customer would consider ordinary and an American customer often never thinks to ask for. The answer is usually available, buried in a support article rather than on the pricing page, and reading it takes about ten minutes.

The party nobody accounts for is the person doing the restore

Every backup arrangement is a transaction with at least three sides: you, the vendor, and whoever will actually perform the recovery. That third party is the one people leave out. In a household it is frequently a spouse who has never opened the backup software, an adult child handling an estate, or a repair shop that needs the account credentials to put a replacement laptop into service. The vendor's support desk is a fourth participant, and its identity verification process is the step that quietly decides whether a legitimate restore succeeds. Vendors that publish a clear account recovery path, including what documentation an executor needs, are giving you something the price comparison does not show.

This is also where encryption choices bite. End to end encryption with a key only you hold is genuinely stronger, and it means the provider cannot reset your access for anyone, ever, including the person cleaning up after you. That is a reasonable trade if the key is written down and stored where the right person will find it, which usually means a sealed envelope with estate documents or a password manager with an emergency access contact named. Where the key lives is a decision about a person, not a technology.

The restore drill that takes an afternoon

Testing means recovering something and confirming it opens, not confirming that a green checkmark appeared. Pick three files of different kinds, a photo, a tax return, and a document from more than a year ago, then restore each one to a new folder from each backup you maintain. Time it. Note which restores required a password you had to hunt for, and which required a second factor tied to a phone you might not have during a burglary. Then hand the instructions to the person who would have to do it without you and watch them try, saying nothing. The gaps surface in the first ten minutes, and they are almost always credentials rather than software.

Do this twice a year, on a date you already remember, and write the result down with the date. Two lines is enough: what restored, and what needed a password reset. A household that can produce that note is in a materially different position from one that believes the same thing without evidence, and the work involved is roughly a long lunch.

The comparison that matters is not between vendors. It is between a copy you assume exists and a copy someone has actually pulled back onto a working machine, with the person most likely to need it standing at the keyboard.