Get feedback from real use, not projections. Iterate on actual problems, not hypothetical ones. Direction is a filter, not an order-dependent roadmap.
Decide as late as you can without losing options. Favor near-action decisions for reversible, loosely-coupled choices where fresh information is highest quality. Decide earlier when a choice is hard to unwind, tightly coupled, or has long lead times.
Certainty isn’t about horizon; it’s about relevance. Plan this month, finish it, then plan the next. Short horizons deliver more certainty; long projections miss. Make more short-term calls—the information is fresh.
Keep the compass, drop the itinerary. Let the far objective constrain what you will and won’t do; let near-term evidence set the sequence. Use intent to narrow options; use learning to pick the next step.
Start with values and principles.
Define what matters at the core (e.g., “remove administrative work to save teachers time on x, y, and z,” “x% of students enroll in y class,” “x% of students pass y exam within ZPD”).
Measure what matters by the work itself.
Is the product good? Does it solve the problem? Does it help users/customers? If someone isn’t doing good work, the team sees it quickly. If you can’t tell who’s contributing, you have a communication or clarity problem—not a measurement problem.
If you need elaborate systems to ensure people are working, either you hired poorly, failed to clarify what matters, or built an environment where people can’t do their best work. Fix root causes instead of adding management overhead. Teach people to set and achieve their own goals so they feel ownership within the directional space set forth and to sustain motivation when it suffers.
Use rolling-wave planning; decide late for reversible/loosely-coupled choices and decide earlier for irreversible/high-coupling/long-lead choices.