Vision & Strategy

It sounds decisive. It is, in fact, empty — and there is a simple test that shows why. Try to imagine a sensible company that would state the opposite. Nobody sets out to offer the worst support at the highest price in one region. If the reverse of your statement is absurd, your statement is not a choice. It is a description of wanting to do well.

A real strategy contains things you have decided not to do, and those refusals should be slightly painful to read. Budget airlines are the standard illustration: no assigned seats, no free meals, one aircraft type, secondary airports. Every one of those is a genuine sacrifice, and every one of them is what funds the low fare. A competitor who copies the fare without copying the refusals simply loses money faster.

The reason this matters day to day is that a strategy without refusals cannot settle an argument. When every team can find justification in the document for whatever they already wanted to build, the document has stopped doing work. You will know this has happened because prioritisation debates get decided by whoever is most senior or most persistent in the room, rather than by anything written down.

So when you read a strategy, look for the sacrifice. If nothing has been given up, nothing has been chosen.

The hardest sentence in product work is not “we will build this”. It is “we will not build that, and we are comfortable losing the customers who need it.”

Strategy is mostly this sentence, repeated. Every genuine choice closes a door, and closing doors feels like failure — which is why so many strategies quietly avoid it and end up as a list of everything, ranked by enthusiasm.

The mechanism is simple once you accept it. Your resources are finite: a fixed number of engineers, weeks, and pounds. Saying yes to something is automatically saying no to whatever those same engineers would otherwise have done, whether or not anyone acknowledges it. The choice is not between saying no and not saying no — it is between saying no deliberately and saying it by accident.

A useful habit is to make the refusals explicit and written. Not “we are focusing on mid-market”, which sounds like emphasis, but “we will not build the SSO and audit features that enterprise buyers require this year, and we expect to lose those deals.” Now the trade-off is visible, someone can disagree with it in advance, and the sales team is not surprised in six months.

This also gives you a way to spot a strategy that has stopped being one. If a new request can always be absorbed — if the answer to “can we also do this?” is never no — then the document is not constraining anything, and the strategy has become a wish list wearing a serious font.

A company decides its strategy is to win the mid-market. Everyone nods. The next quarter’s roadmap contains four enterprise features, because a large enterprise customer complained loudly in week two.

Nothing dramatic happened here. Nobody overruled the strategy. It simply was not used — and a strategy that is never used to reject anything quietly stops existing, usually within a quarter of being written.

The gap between a company strategy and a roadmap is where most strategies die, and closing it needs a mechanism rather than good intentions. The practical version is a rule: every roadmap item must name the strategic choice it serves, and items that cannot are not scheduled. Not rejected as bad ideas — many are perfectly good — but not scheduled, because the point of a strategy is that good ideas outside it still lose.

This is deliberately a little bureaucratic, and it is worth it. It moves the argument to the right level. Instead of debating whether a feature is valuable — where the loudest customer story always wins — you are debating whether the strategic choice was right, which is a discussion the company already had and can refer back to.

When an item genuinely does not fit but everyone still wants it, that is real information: either the strategy is wrong and should be changed openly, or the item should be dropped. What should not happen is the third option, which is what usually happens — the item is built anyway, the strategy is left untouched, and everyone learns that the document does not mean anything.