Crafting a Product Vision
A product vision is the single most leveraged document a PM writes, and in my experience it's also the one we write worst. Done right, it makes a thousand decisions for you and aligns people you'll never be in a room with. Done wrong, as a feature list with adjectives, it's wall decoration. This note is how I actually write one.
#Vision vs strategy vs roadmap: the three layers people conflate
These three words get used interchangeably, and the confusion is the source of most "we have no strategy" pain. They're different altitudes of the same idea.
| Layer | Question it answers | Time horizon | Changes |
|---|---|---|---|
| Vision | Where are we going, and why does it matter? | 3-10 years | Rarely; it's the North Star |
| Strategy | How will we win, and what won't we do? | 1-3 years | Occasionally, on real new information |
| Roadmap | What are we doing, in what order? | Now / Next / Later | Continuously |
The vision is the destination. The strategy is the route and the deliberate detours you're not taking. The roadmap is the next few turns. A vision without a strategy is a daydream; a roadmap without a vision is a feature factory with a calendar.
If your "vision" could be flipped to its opposite and some other company would proudly claim it ("we want to delight customers with quality products"), it's not a vision. It's a platitude. A real vision is specific enough to be wrong.
#What a good vision actually contains
A vision is fundamentally about the customer's better future, not your product's feature set. An everything-store kind of vision (the store where you can buy anything you can think of, delivered fast) is a vision. "Add 14 new categories" is not. I anchor every vision in three things, straight from the Product Vision and Roadmapping behavior: it's grounded in a real customer need, it reads the market honestly, and it connects to company strategy, while balancing aspiration against feasibility so it's a stretch, not a fantasy.
#Write it as a narrative
The format that aligns people is a story, not bullets. My go-to is a short narrative of a specific customer's day, before and after: what's painful now, and what their world looks like once we've built toward the vision. Amazon's famous future-press-release ("working backwards") is the same move. Write the press release for the thing as if it already shipped, and the vision becomes concrete and testable. Narrative works because humans align around stories and argue about bullet points. This is a storytelling skill, and it's why I treat vision-writing as a communication competency as much as a strategic one.
A vision is not a tagline you brainstorm in an offsite and stencil on the wall. A tagline is marketing. A vision is a decision-making instrument: it has to be detailed enough that an engineer or designer can use it to resolve a real trade-off without you in the room. If it can't do that, it's not done.
#Evangelism: the half everyone skips
Writing the vision is maybe 30% of the work. The other 70% is repetition. A vision stated once at an all-hands is forgotten by the next sprint. I repeat it everywhere (in spec preambles, in stakeholder updates, in exec reviews, in 1:1s) until I'm genuinely tired of saying it. That fatigue is the signal that the org is finally starting to internalize it. A vision you're bored of is usually a vision that's working.
And it evolves. Not the destination, which should be stable, but the framing, the examples, the emphasis as the market moves. The Product Vision and Roadmapping behavior calls this "reinforce and evolve," and the distinction matters: you evolve the expression of the vision while holding the intent steady. Whiplash on the destination destroys trust; refreshing the story keeps it alive.
#How AI is changing it
AI will draft you a polished vision narrative in seconds, and that's exactly the trap. The words were never the hard part. A model can produce a beautiful, generic vision because it's trained on a million beautiful, generic visions; what it cannot supply is the conviction and the specific contrarian bet that makes a vision yours and worth following. Use AI to pressure-test and sharpen (ask it to "argue the opposite of this vision," or "what would a competitor's version be") and to stress the feasibility-aspiration balance. But the bet itself, and the credibility to stand behind it when it's unpopular, is the part of this job AI makes more valuable, not less.
#Continue Reading
- Building and Managing a Roadmap for turning the vision into a sequenced, communicable plan.
- Product Vision and Roadmapping is the competency this craft note sits under.
- The DHM Model is a test for whether the strategy beneath your vision is actually defensible.
- Storytelling with Data is the narrative skill that makes a vision land.
- Executive Communication for how to evangelize the vision upward without losing it in translation.