The mistake in multi-year planning is applying the same rigor to year three that you apply to the next quarter. Detailed milestones two years out are not planning, they are fiction written in a planning format. They give the appearance of commitment while consuming the time that should go to the near-term work that actually determines the trajectory.
A useful multi-year roadmap has three layers. Near term, twelve months or less, is detailed: named deliverables, dependencies, owners. Medium term, twelve to twenty-four months, is directional: the capabilities you are building toward, the bets you are making, the things you are deliberately not doing. Long term, beyond twenty-four months, is principles: the outcomes you are optimizing for and the constraints you will not violate.
The medium and long-term layers are where strategy lives. What you choose to commit to in year two is a signal about what you believe will be true about the market, the technology, and the organization. Making those bets explicit, rather than hiding them in milestone dates, makes them legible to the leadership that has to approve them and the teams that have to execute them.
Review the multi-year roadmap quarterly, not annually. Not to update every milestone, but to test whether the bets you made still hold. Markets move, technology moves, priorities shift. A roadmap that is only reviewed annually is a document, not a plan.
The value of multi-year planning is not the plan. It is the thinking the plan forces. Done well, it surfaces the decisions that need to be made before the work starts rather than in the middle of it.