Contained in the quippy wisdom of George Box is a reminder to stay aware: any model is an imposition of order where it may not exist, and contains assumptions that may not be appropriate or accurate. In fact, your own models almost certainly do not describe everything perfectly. Choose useful ideas, be wary of statistical measurements, and discard them when needed.

Always operate with an active and constantly revised working model of the world before you, and how to create the best good for the company, enterprise, customers, or whatever you’re endeavoring toward.

In animal ethology, a signal is something an animal does to change another’s behavior—a rattlesnake’s rattle warns you to back off. A cue is something an animal does that others can observe and learn from—a rattlesnake coiling up just before it strikes. Signals are intentional; cues are inadvertent.

Distinguish between what people are telling you and what they’re showing you. Actions speak louder than words.

Instead, know your audience. Do not “go along to get along,” for that is how the leaderless perish as proverbial lemmings. Knowing the people you’re speaking with will take you miles farther as a leader than joining the crowd consensus.

While there are commonalities and “main effects,” specific context determines which solutions are most actionable and productive. Always accommodate peculiarities and contingencies of your unique situation rather than whapping a complex system with an all-purpose hammer.

Brevity is the vital core.

When interested in knowing about prior work, go directly to the person who created it, whether in books, over email, on the phone, or in person. Various help forums can be useful for a first pass on discovering what might be posing a challenge, but there’s no substitute for contacting the people who have been there before, for any kind of deep technical or operational assistance.