How do you know when a strategy needs improvement instead of more time?

A few losses do not necessarily mean the strategy has stopped working. But ignoring meaningful evidence can also be costly.

What data would convince you to review or change your approach?