Open enough strategy decks and you will meet the same excellent intentions.
We will be customer-centric. We will use data. We will innovate. We will create a seamless experience and grow sustainably. Nobody objects because the opposite choices sound absurd.
That is the problem.
A strategy is supposed to help people choose under pressure. A list of virtues tells them how the organisation would like to be described.
The opposite test
Roger Martin argues that strategy is an integrated set of choices, including choices about what not to do. He offers a sharp test: if the opposite of a statement is stupid on its face, the statement is likely an operating imperative rather than a strategy.
“Understand our customers” matters. “Ignore our customers” is obviously foolish. The original statement therefore cannot distinguish your approach from a competent competitor’s.
Try the test on “build a high-quality product,” “make evidence-based decisions” or “collaborate cross-functionally.” Keep them as principles if they improve behaviour. Do not ask them to carry the strategy.
Strategy needs a constraint
A useful product strategy makes at least three things clearer:
- Where are we concentrating? A customer, situation, need or market boundary.
- How do we expect to create an advantage? A capability, experience, business model or position that reinforces the choice.
- What will we not pursue? At least for now, despite plausible value.
Without the third part, every initiative can claim strategic alignment. The strategy becomes a nightclub with no door policy.
“Serve small professional teams who need an auditable workflow, winning through unusually fast setup and opinionated defaults” can be challenged. It excludes enterprises that demand deep customisation. It creates implications for onboarding, architecture, sales and pricing.
That discomfort is evidence that a choice exists.
The initiative list in disguise
Another common substitute is a strategic plan made of projects: launch AI assistant, redesign onboarding, expand to Germany, improve platform reliability.
Projects are actions. Strategy is the logic connecting actions to an advantage. Without that logic, teams can deliver every initiative and still have no idea whether the company became more strategically coherent.
Ask of each initiative: because of which choice?
If the AI assistant exists because every competitor has one, that may be defensive parity. If it exists because the strategy depends on turning complex expert workflows into guided decisions, it may reinforce a distinctive position. Same feature category, different strategic meaning.
A strategy should change everyday arguments
The test of product strategy is not whether the leadership offsite felt aligned. It is whether the strategy changes ordinary decisions later.
Does it help a PM reject a reasonable feature? Does it help design choose between flexibility and speed? Does it tell sales which customer request is informative and which points away from the intended market? Does it help engineering invest in one capability over another?
If every option remains equally compatible, the strategy is scenery.
Begin with a problem worth solving
In his account of how he approaches strategy, Roger Martin describes strategy as a problem-solving technique and spends substantial effort defining the problem before generating choices. This is a useful antidote to strategy exercises that begin with a template.
“Growth is slowing” is a symptom. The strategic problem might be that the product serves a maturing segment with little remaining expansion, that differentiation has eroded, or that acquisition succeeds while customers fail to reach value. Each diagnosis implies different choices.
A team that skips diagnosis can create an energetic roadmap whose initiatives solve incompatible versions of the problem.
Metrics should express the choice
Amplitude’s discussion of product strategy connects strategy with customer needs, competitive context, business goals, product vision and measurable progress. The metric matters because it turns part of the strategic story into an observable claim.
But the relationship runs from choice to metric, not the reverse. Starting with an available metric can make the strategy conform to what is easy to count. A strategic choice around trusted collaboration may require measuring meaningful shared work, not merely sessions or messages because those events already exist in the warehouse.
Ask what customer behaviour would indicate that the chosen advantage is working. Then decide what evidence could represent that behaviour and what balancing measures protect against gaming.
Write the refusal beside the choice
For every strategic statement, add a paired refusal. “We will optimise fast setup for small professional teams; we will not pursue deep enterprise customisation this year.” The refusal makes resource implications visible and gives teams language for saying no consistently.
Revisit refusals when the environment changes. Strategy is not stubbornness. It is a coherent set of current choices, adjusted deliberately rather than eroded request by request.
The one-page test
A product strategy should fit on one page without becoming vague. State the diagnosed problem, chosen audience or arena, mechanism of advantage, capabilities required, measures and explicit non-choices. Supporting analysis can be extensive, but the decision logic should be repeatable by the people expected to use it.
If teams need the original presenter to interpret every trade-off, the strategy has not yet become organisational context.
Questions to take back to your team
- Which sentence in your strategy has an opposite that is plausible rather than obviously stupid?
- What specific diagnosis explains why your current position is not good enough?
- Which customer, arena or opportunity receives disproportionate investment because of the strategy?
- What attractive request can a team reject confidently without escalating to leadership?
- Does your chosen metric express the strategic mechanism, or merely reflect what was easy to instrument?
Read your strategy without the company name. Could it belong to three competitors? If so, where are the choices that make your bet recognisably yours?
My take
Keep the good intentions. Organisations need principles. Just label them honestly.
Then write the choices in language specific enough to make someone nervous. Name the customer you are organising around, the mechanism you believe will win and the attractive work you will leave undone.
Good strategy does not make the future certain. It makes the bet legible.