Most documents called strategy are goals with a deadline attached

Rumelt's kernel is diagnosis, guiding policy, coherent action. The test he offers is simple and brutal: if your strategy does not identify a specific obstacle, it is not a strategy.

By that standard most corporate strategy is bad strategy in his precise sense. Ambitious targets, values statements and a list of initiatives that do not reinforce each other. The tell is that you could swap the company name and the document would still read fine.

The uncomfortable implication is that real strategy requires saying what you will not do, which is politically expensive in a way that goal-setting is not.

What is the best diagnosis you have seen written down? Not the plan, the diagnosis.