Software takes the backup. Rules decide whether the backup exists, survives, and actually restores on the worst day of your year.
Reflect standalone and Site Manager produce the same kind of image files, so the survival rules are identical on both sides of the comparison. These are the rules disaster stories are made of — every one of them was ignored by someone whose backups looked fine right up until the restore. None of them cost much. All of them are cheaper than the alternative.
Keep three copies of anything that matters: the working original on the PC, plus two backups. The reasoning is statistical, not dramatic — drives fail, files corrupt, and a backup made from an already-dying disk can be quietly worthless. With two independent backups, a single bad copy is an inconvenience. With one, it is the whole story.
The two backups should not share a fate. An image stored on a second partition of the same disk dies with that disk. An image on the same NAS as the only other copy dies with that NAS. Spread copies across different devices — internal drive images to a NAS, plus a rotated USB drive, or local storage plus cloud. Different device, different failure, different day.
The third rule is the one ransomware taught everyone. Malware that encrypts a PC will happily encrypt every backup it can reach — the always-connected USB drive and the always-mounted network share included. Keep one copy offline or offsite: a rotated drive that lives unplugged or at another location, or a cloud copy. Site Manager repositories and standalone destinations are equally reachable when mounted — the rule is about reachability, not product.
An untested backup is a hope with a file extension. Every administrator who has done a real disaster restore says the same thing: the first restore you ever attempt should not be the one your business depends on.
Finally, keep retention honest. An image chain answers the question how do I get this machine back — it is not where old project files live forever. Decide how far back a restore point must reach (a week? a month?) and let retention rules clear the rest, on standalone Reflect and in Site Manager repositories alike. Storage filled with ancient images is how destinations silently fill and new backups start failing — the exact failure the morning check exists to catch.
See both tools working on real machines — Reflect on a single PC, Site Manager running an office full of computers.
▶️ Watch Full VideoAs a convenience copy, yes — it protects against accidentally deleting Windows or a bad update. As your backup, no: it shares the disk's fate completely. One physical failure takes the original and the backup in the same second. Add at least one copy on a different device.
Match it to how much work you can afford to redo. Daily incrementals suit most offices; the system disk of a home PC may only need a weekly image with files backed up more often by other means. The schedule matters less than the rules on this page — copies, spread, and tested restores.
It automates the visibility — central status, reports and alerts tell you backups completed and images verified. Booting a test restore is still a human-scheduled exercise in both products. Macrium's image verification helps, but only a real restore proves a real restore.