We've probably all heard this phrase: "Once is an accident, twice is a coincidence, three times is a pattern." What this means is that when we see an actual pattern of behaviors or results (and not just random one-offs), there is an underlying issue to address.

Here are three areas where this applies to data management:

1. Process change: When staff or customers consistently complain that a particular process is too cumbersome, it's probably time to revisit how something is done. In many cases I find that processes have been in place for many years without questioning whether all (or any!) of the steps in the process are still necessary.
2. Possible bug: If we can consistently repeat a process that produces the wrong result (e.g., clicking a certain button always  returns an error message) then we have a bug that needs to be addressed.
3. Training issue: If we find that staff is consistently making the same error, don't assume ill intent, assume poor training. The staff person may be doing it the way he or she was (or wasn't!) taught. Correct the error with proper training.

Keep in mind that this is about pattern recognition. I would suggest not reacting immediately to a single error or complaint, but to look for patterns. If there is a consistent pattern, something needs to be addressed.