A tool score should say which user task it improves. A higher number is weaker when the use case stays unnamed.