I first heard this advice from Alan Weiss, consulting guru. The idea is that extrapolating a rule from a single instance is a bad practice. What you may be experiencing is the exception, not the rule.
I see this far too often with my clients. For example, someone on staff (usually a senior staffer who rarely uses the database) will find a data error and declare “This database is a trainwreck! All the data is wrong!”
Another common example is to say “The vendor didn’t deliver on promise X, thus the vendor is bad and can’t be trusted.”
Rather than generalizing from a specific, we should be looking for patterns. Are we finding multiple instances of inaccurate data? And are the inaccuracies consistently wrong, or randomly wrong? After all, people move jobs, people die, they change names, and so the data is never 100% accurate.
And with our vendor, is there a pattern of over-promising and under-delivering? Or did they miss on this particular issue but generally are good about keeping promises?
Use patterns to identify problems. A single issue may be just that, a single issue. Don’t generalize from it.