Originally written in French. Translated by AI — the meaning has been preserved, not the prose.
Angle
A Now / Next / Later roadmap does not align a company through the subjects it displays, but through the number of slots it refuses to open: the Now is worth exactly what the team's real capacity is worth — two large subjects, three if they are smaller — and anything that wins a slot in no column is out of scope for the current horizon. It is that rationing, and that alone, which produces focus; the three-column format has nothing to do with it. Hence the fragility of the setup: one place downstream that accepts rejected subjects without limit is enough for the renunciation to stop holding. A Next left unbounded, a backlog collecting what the roadmap declined to take, a Later card shown to a customer who turns it into an expectation — each of these reservoirs hands back to the discarded subject the slot the roadmap had just taken from it, and the roadmap becomes a better-presented inventory again. The only work outside the columns that does not undo the rationing is work that calls for no arbitration: a ring-fenced, bounded slot reserved for subjects whose value is already settled.
Synthesis
The Now / Next / Later format presents itself as a change of measure: you stop sorting subjects by date and sort them by degree of certainty. That change is real, and it produces no alignment on its own. Three columns contain no criterion, no weighting, no strategy; they publish decisions taken elsewhere. What produces the effect is a property of the setup that gets talked about far less: the number of slots is finite, and it is fixed by something other than the merit of the subjects.
The mechanism is easier to see in reverse. As long as slots are free, a subject gets added without having to defend itself: it only has to be interesting, and it almost always is. The discussion then bears on its intrinsic quality, ground on which nobody has any reason to object. As soon as slots are counted, the question changes nature — no longer "is this subject good?" but "is it better than the one holding the slot?". Scarcity does not sort on its own: it turns merit into comparison, and it is comparison that forces you to name the strategy, the risk reduced, the business objective served. A roadmap not tied to a strategy rations too, but it allocates by balance of power.
The renunciation thus produced is written nowhere. No column is called "no"; the refusal can be read only in the complement of the three lists. Yet it is the only part doing any work — and it is also what makes the whole thing undoable without anyone having to contest anything. A discarded subject does not need to be reintroduced into the roadmap to keep existing: it only needs to find somewhere to land.
Those places are many, and none presents itself as a transgression. The first is internal: the candidates tier, which goes unbounded because nothing there is promised. It fills with subjects each person drops in to keep them visible, circulation stops, and the column becomes a political pre-backlog. The second is downstream: the backlog, which takes in what nobody wanted to kill and holds it until someone happens on it. The third is external: an uncommitted subject shown to a customer becomes an expectation, and an expectation gets quoted at renewal as a promise — a renunciation held internally can be cancelled by a slide presented outside. The fourth is the most discreet, because it looks like prudence: keeping a subject visible in Later without writing down what would bring it back, which is to say archiving it while claiming to watch it.
There remains what the rationing costs, and it costs twice. It is paid for on the product: once the slots are allocated, an obvious small improvement gets none, whatever its value. It is exactly the gesture that guards against dispersion that forbids handling anything not worth a debate. And it is paid for by those who apply it, except that here the accounting reverses: what the rationing removes is a cost nobody was measuring — a subject neither committed to nor abandoned keeps coming back in meetings, and working on a subject whose prioritization is unknown lowers the intensity of the work. Settling things hands the organization back the attention it was spending on not settling them.
Hence the form the only admissible way out must take. Any prioritization framework always produces subjects it cannot handle; the place outside the framework is necessary, and it degrades in both directions — absorbed if it becomes a column, diverted if it stays an informal tolerance. What holds it is not its size but its entry criterion: it takes only subjects whose value is already settled, therefore never subjects that would have called for an arbitration. An exception that arbitrates nothing cannot serve to get around the arbitrations. The same rule applies on the technical side: a known architectural weakness does not enter the allocation of slots until someone has decided to deal with it — before that decision, it is not a subject, it is a discomfort.
Tensions / contradictions
Later is presented both as a column of the roadmap and as a potential reservoir. The two readings coexist in these notes without being settled: keeping a subject visible without commitment is described as a legitimate gesture, and the same gesture, stripped of its trigger for reassessment, is described as the most discreet way of cancelling a renunciation. Nothing indicates who, in an organization, notices the passage from one state to the other.
Second tension, on the way out of the framework. It is justified by what the rationing prevents — improvements that are small, obvious, expected — and it is at the same time barred from any subject that would call for an arbitration. This assumes that rejected subjects split cleanly between "value already settled" and "value still to be investigated". The subjects that fall in between are handled by neither mechanism.
Finally, the capacity constraint and the requirement of a strategy do not size the same thing. The first fixes how many slots exist, the second says which ones to fill. A team whose capacity collapses sees its rationing tighten without any strategic choice having been made.
Questions
- Can an outside observer establish that a reservoir has opened in an organization, or does the discrepancy only show once the subjects have come back out?
- What holds the rationing when whoever arbitrates reports to the people whose subjects he refuses?
- Can the attention a deliberate renunciation hands back to an organization be established other than by asking those who lived through it?