Data backup serves a number of different purposes. Individual data files can be corrupted, deleted or misplaced. Whole folders can disappear due to an accidental drag-and-drop operation, which usually results in a recovery request. Magnetic storage disks can and do go faulty, and ultimately, a whole building can be lost due to a fire or similar disaster.
Each of these problems requires a different type of solution. A corrupt file requires that a file level backup was scheduled before the corruption, and then a file level restore. A dragged folder simply needs a search operation to find it and put it back. The problem is that the customer often cannot remember exactly what the folder was called, nor when she last used it. RAID will usually protect from a broken disk, but sometimes full disk restores are still required.
The backups required to fix these problems are usually traditional full / incremental copies to tape, although NAS sparse snapshots are very useful for missing files.
At the next level, the loss of a building or site is a major disaster, and they do occasionally happen. Recovery of significant anounts of data to several servers is a bit of a challenge, especially in short timescales. The ultimate protection is to have a second, hot standby site with data synchronously mirrored from the primary site. It is possible, if your company will spend the money required, to transfer all processing to the second site with virtually no loss of service.
The Backup and Disaster Recovery section discusses some of the backup products, and snapshot copy and remote mirroring options.