Can you see the danger here?
A ministry faithfully backs up their computer data every day. They have two or three sets of media. Perhaps it is a tape that’s swapped every day or several different thumb drives that are alternated. When it’s time to make a new backup, the staff person faithfully grabs a backup media, making sure it is the oldest backup, puts it in the computer and successfully makes another backup. Next day the process is repeated.
What’s wrong?
If the hard drive crashes in the middle of the day, then everything is fine. The ministry has yesterday’s backup to work from.
Let’s say that the problem is instead some sort of corruption in the data and not a catastrophic failure. Data could be corrupted and you might not realize it for some time. If the data is corrupted you may not know about it today when you make a backup. You may not know about it in a few days when you make another backup and write over the last good backup data in existence.
When you find out there’s a problem in a week or a month you will have no backups to recover from.
Solution?
Keep many generations of backup data.
There should be weekly backups retained for a long time along with monthly or quarterly backups that are retained for much longer time.
That way if the data is corrupted in a subtle way that only (only?) turns your database into garbage instead of in such a severe way that the program completely crashes, you can work you go to a backup that wasn’t corrupted. Then you only have to reconstruct everything from that point through today.
Keep many generations of your backup data.