Calculate the manpower, lost time and/or sales, and other costs that would be incurred if the file server or workstation crashed right before the next backup was scheduled to take place (always assume the worst scenario). If the cost is excessive, the strategy should be adjusted accordingly.
For example, the cost to re-create an extensive database system that is continually updated by several database operators would be quite substantial. On the other hand, the cost to re-create the data for a user creating one or two inter-office memos would be considerably less. In this scenario, the network administrator would probably opt to back up the database several times daily, and set up daily jobs for the user's workstation.
In an ideal environment, one full backup should be performed on workstations every day and servers should be fully backed up more often. Important data files and directories that constantly change may need to be backed up several times a day. Because of time and media constraints, this is not feasible for many environments, so a schedule including incremental or differential backups must be implemented. For safety reasons, a full backup should always be performed before adding new applications or changing the server configuration.