The planning meeting ends, as these meetings often do, in a warm glow. Sales has the feature its biggest prospect asked for. Support has a promise that the bugs clogging its queue will be fixed. The chief executive has the new market. Engineering has, at last, a quarter set aside to repair the foundations. Marketing has something to announce in the spring.
Nobody lost. That is precisely the problem.
Three months later, the same people meet again. Almost nothing on the slide has shipped. Each department is disappointed, and each, privately, blames another. The verdict, when it comes, is familiar: no focus, no strategy, weak leadership.
It is a fair verdict. It is also an incomplete one. Capable, experienced executives produce this roadmap year after year, in companies large and small. When clever people keep doing something that looks foolish, it is usually doing a job nobody has named. The interesting question is not how to stop it. It is what the roadmap that says yes to everyone is actually for.
A plan with no enemies
The textbook answer comes from strategy. Michael Porter, the Harvard economist, argued in 1996 that the essence of strategy is choosing what not to do. A year later, freshly returned to a struggling Apple, Steve Jobs told a room of developers that focusing is about saying no, and then proved it by cutting the company's sprawling product line to a handful.
By this standard the yes-to-everyone roadmap is not a strategy at all. It is an inventory. It lists activities without ranking them, and so it is silent at exactly the moment it is needed: midway through a quarter, when two good things compete for the same scarce people.
The textbook answer has the virtue of clarity and the vice of condescension. It assumes the people drawing up the plan simply do not understand strategy. In my experience they understand it very well. They say yes anyway. Something else is at work.
The peace treaty
In 1963 two economists at Carnegie, Richard Cyert and James March, offered an account of the firm that has aged remarkably well. A company, they argued, is not a single mind pursuing a single goal. It is a coalition of groups that want different things and never entirely agree. Firms survive this not by settling the argument but by sidestepping it. They aim for acceptable performance on each goal rather than the best on any, and they attend to their competing goals in turn, one after another.
Read in that light, the roadmap that pleases everyone is not a failure of strategy. It is a peace treaty. Each department signs because its interests appear on the page, and the coalition holds for another year.
This reframing is uncomfortable for the strategist, because it means that tearing up the treaty is not a planning decision. It is a political act, and will be resisted as one.
Yet Cyert and March also explain where the treaty fails. Their firms kept the peace by taking goals in sequence: the customer this season, costs the next, the product after that. The modern roadmap tries to honour every goal at once. It preserves the promise of the coalition while discarding the very mechanism that made coalitions workable.
The lesson is subtle. The most useful word in a planning meeting may not be "no" at all. It may be "next".
The arithmetic of queues
There is a third way to look at the problem, and it has the advantage of being indifferent to anyone's feelings.
Queueing theory offers a modest-looking result known as Little's Law. The average time a piece of work spends in a system equals the amount of work in the system divided by the rate at which work leaves it. Hold a team's capacity steady, start more things at once, and every one of them takes longer to finish. This is not a matter of effort or talent. It is arithmetic.
The arithmetic is also generous, because it ignores friction. Gerald Weinberg, a long-time observer of software teams, estimated in the early 1990s that each additional project a person juggles consumes a slice of their time in the switching alone, and that by the third project a large part of the day is lost in the gaps. His figures were estimates, not measurements. Few people who have managed engineers would dispute their direction.
The consequence is counter-intuitive. A roadmap that says yes to everyone does not deliver everything slowly. It delivers nothing for a long while, then a great deal at once, and late. Half-finished work is worth nothing to a customer. A feature that is mostly built is not mostly useful. It is simply unavailable.
In defence of yes
Having made the case against the crowded roadmap, it is only fair to make the case for it.
In 1991 James March, by then at Stanford, published a paper on the tension between exploration and exploitation: searching for new possibilities versus refining what already works. Organisations, he observed, drift towards refinement, because its rewards arrive sooner and more surely. Over time they become very good at things that matter less and less.
From this vantage point, a roadmap studded with small and varied items is not necessarily unfocused. It may be a portfolio. A young company that does not yet know which market will answer its knock ought to knock on several doors. Saying no too early can mean committing, with great efficiency, to the wrong idea.
So the meaningful line is not between long roadmaps and short ones. It is between a portfolio and a wish list, two things that look identical on a slide and behave entirely differently in practice. A portfolio is made of small bets, each with a deadline and an agreed point at which it will be abandoned. A wish list is made of large commitments with no end date, which nobody may cancel, because cancelling would break the treaty.
Most companies believe they are running the first. A surprising number are running the second.
The cost of a quiet no
The final perspective is the most human, and perhaps the most neglected.
A "no" in a planning meeting is rarely heard as a judgment on a feature. It is heard as a judgment on the person who asked for it, or on the department behind them: your problem matters less. That is why it stings, and why leaders go to such lengths to avoid saying it.
But consider the alternative from the same person's chair. A "yes" that is quietly allowed to lapse teaches a harsher lesson: that the plan cannot be trusted. Sales begins making its own promises to customers. Support builds its own workarounds. Each department stops relying on the shared plan and starts defending itself against it. The treaty survives on paper and dies in practice.
Seen this way, the kindest answer is often the clearest one. "Next quarter, and here is why" is easier to live with than a "yes" that slowly turns into a "no" without anyone having the courage to say so.
Next, not no
None of this yields a formula. A company still searching for its market and one scaling a product that already works should not plan in the same way, and anyone who offers a single template for both is selling something.
What the different perspectives offer instead are better questions. If half the items vanished from your roadmap tomorrow, who would protest, and would they be speaking for the customer or for the coalition? Which items have a point at which you would stop them, and which have become impossible to stop? Is your plan a sequence or a pile? And when it slips, as plans do, who decides what gives way: someone with the authority to choose, or whoever pushes hardest that week?
The answers will not tell you what to build. They will tell you whether you are holding a strategy, a portfolio, or a peace treaty whose upkeep now costs more than the peace is worth.
A roadmap, in the end, is not a list of what a company will do. It is a record of what it was willing to give up.
