Richard Rumelt argues that every sound strategy contains a kernel of three parts: an honest diagnosis of the central challenge, a guiding policy for addressing it, and a coherent set of actions that follow from that policy. The kernel doubles as a test. Can you state all three? If you can name only a destination, you are holding an aspiration — and an aspiration commits no resources.
What Rumelt calls bad strategy is the familiar substitute: goals presented as if they were plans, motivational language standing in for thought, and — above all — a refusal to name the real obstacle. The diagnosis is the part most often skipped, because naming the obstacle honestly is uncomfortable; but a strategy that will not say what is actually hard is a strategy about a different, easier problem. The coherence requirement does the other half of the work: the actions must reinforce one another and follow from the policy, or the "strategy" is just a list wearing a title — activity mistaken for direction.
See also: Policy, Strategy, and Tactics, Choosing What Not to Do.