Strategy, Vision and Roadmap

The idea in one minute#

Three different artifacts get called strategy, and conflating them produces the most common organisational failure in product work: a prioritised list of features with a mission statement on top, which nobody can use to make a decision. A vision describes a future state and is motivational. A strategy is a choice about where to compete and how to win there, and it is only a strategy if something was refused. A roadmap is a sequence, downstream of both. The test that separates them is simple: if your strategy document contains no sentence beginning “we are not”, it is a vision or a plan wearing the word.

A picture#

flowchart LR
  V[":i-eye: <b>Vision</b><br/><small>where this ends up, and why it matters</small>"] --> S{":i-compass: <b>Strategy</b><br/><small>where to play, how to win, what we refuse</small>"}
  S --> R[":i-calendar: <b>Roadmap</b><br/><small>sequence, consequence of the above</small>"]
  R --> W[":i-wrench: <b>The work</b>"]
  S -.->|"no refusals in it"| FAKE[":i-triangle-alert: <b>A backlog with a mission on top</b>"]
  V -.->|"used as a strategy"| DRIFT[":i-triangle-alert: <b>Everything is on strategy</b>"]
  class V neutral
  class S queue
  class R io
  class W compute
  class FAKE,DRIFT warn

The two dotted paths are the failures, and the second is the subtler one: a vision broad enough to be inspiring will justify any piece of work, which is why it cannot do a strategy’s job.

How it really works#

What each one is#

A strategy is a choice, so it costs something#

The definitional property is that choosing one thing forecloses another. “We will serve mid market companies in Europe, winning on implementation speed, which means we will not pursue enterprise deals that require a six month security review this year” is a strategy. It names the segment, the basis of winning, and the thing refused.

“We will delight our customers with an innovative platform” is not a strategy. Nothing is foreclosed, so nothing can be decided with it.

Why the vision cannot do the strategy’s job#

A good vision is deliberately broad so it survives changes in plan. That same breadth means it cannot discriminate between two proposals, because both will fit under it. When a company uses its vision as its strategy, every project can be justified, prioritisation becomes a negotiation about who argues best, and the loudest function wins repeatedly.

Why the roadmap cannot either#

A roadmap is a consequence. If you have a strategy, the roadmap is an argument about sequencing and dependencies, which is a tractable conversation. Without one, the roadmap becomes the place where strategy is implicitly decided, one quarter at a time, by whoever has the most urgent request. The symptom is a roadmap that changes shape completely every quarter while everybody insists the strategy is unchanged.

How to tell which one your company has#

Ask three people in different functions what the company is not doing this year. If the answers agree, there is a strategy. If the answers are vague or contradictory, there is a vision and a backlog, and the gap between them is being filled by argument.

What a product manager does about this at the team level#

You usually cannot write the company’s strategy, and you can always write your team’s: the segment you are serving, what you are winning on, what you are not doing this year, and what would make you change your mind. One page. It is also the fastest way to discover whether the company’s strategy exists, because the moment you publish a refusal, anybody who disagrees will say so.

What fills the box in 2026#

Nothing useful. This is organisational writing, and the tooling question is a distraction. The one practical habit worth adopting is keeping the strategy page dated and versioned next to the decision log, so that when it changes you can see what changed and why, which is the thing that builds trust in strategy over several quarters.

Sector lens#

Remember this#

• Vision is a future state, strategy is a choice with refusals, roadmap is a consequence.
• If no sentence begins “we are not”, it is not a strategy.
• A vision cannot discriminate between proposals, which is why using it as strategy makes the loudest function win.
• Without a strategy, the roadmap becomes where strategy is decided, badly, one quarter at a time.
• Ask three people what you are not doing this year. Agreement means a strategy exists.

Try it#

1. Ask three colleagues in different functions what the company is not doing this year. Write down the answers verbatim.
2. Write your team’s one page strategy: segment, basis of winning, refusals, and what would change your mind. Publish it and see who objects.
3. Take your current roadmap and try to derive it from the strategy. Anything you cannot derive is either a gap in the strategy or work that should not be there.

Check yourself#

1. Why can a good vision justify any piece of work, and why is that a feature of visions rather than a flaw?
2. What is the single test that distinguishes a strategy from a plan?
3. A roadmap changes shape completely every quarter while leadership insists the strategy is stable. What is actually happening?