Insights
Prioritization When Execution Becomes Free
For about twenty years, nearly every method we used to decide what to work on had the same shape: divide the value of a thing by the effort required to build it, then sort the list. RICE does this. Weighted shortest job first does this. So does the value-versus-effort quadrant that every team has drawn on a whiteboard at least once. The denominator was always some estimate of human labour, because human labour was the scarce input worth protecting.
Something quiet happens to that arithmetic when the denominator approaches zero. The ratio stops discriminating. Everything scores well. Everything looks worth doing. A roadmap becomes, in effect, a to-do list with better formatting. Most teams are reading this as good news, and in one narrow sense it is. But a prioritisation framework was never only a ranking device. It was a refusal device. Scarcity was quietly doing two jobs at once: it rationed capacity, and it forced justification. Only the first of those has been automated away.
The second job deserves more respect than it usually gets — and more scepticism. An experiment published in the Journal of Management and Governance (2024, 360 participants) found that when managers were required to justify a project choice, they shifted measurably away from the higher-risk, higher-return option toward the one that was easier to defend. Justification is not a neutral instrument. It disciplines, and it also biases toward the explainable. So the answer to a collapsed denominator is not simply “make people explain themselves more.” That trade — discipline in exchange for ambition — is a bad one to make by accident.
Making became cheap
Meanwhile the cost of a wrong choice has not fallen at all. New Relic’s 2026 State of AI Coding report, surveying 200 technology leaders, found that 94% rate AI-generated code as higher quality than human-authored code at the moment of review — while 74% say at least a quarter of it requires significant rework after deployment, 86% report more senior-engineer firefighting, and 82% have had a production failure traced to it in the past six months. Read those two findings together and the picture is clear enough. Making became cheap. Owning did not.
What goes in the denominator now?
So the practical question for anyone rebuilding how their organisation chooses: what goes in the denominator now? Three candidates that are still genuinely scarce:
- Attention after launch. Not what it costs to build, but who must keep this alive — review it, explain it, repair it, answer for it — for the next two years. This is the one estimate that has become more honest, not less.
- Reversibility. How cheaply can we be wrong about this? When building was expensive, we bought certainty before starting. Now we can buy it after, but only for the decisions we can unwind.
- Conviction. What would we have to believe for this to matter? If nobody can answer in a paragraph, the item is not ready to be scheduled. It needs thinking, not sequencing.
A declined register
And one instrument I would add to any team’s quarter: a declined register. The backlog used to be an organisation’s written record of its own judgement. Everything below the line was a decision — a considered refusal, kept in a place where it could be revisited and argued with. When everything can be built, nothing sits below the line, and that record silently disappears. So write it down separately: one line per refused item, with the reason. Read it back a year later. It is the cheapest available test of whether you were disciplined or merely busy.
None of this is a productivity problem. Choosing what is worth doing was always the part of the work that could not be delegated — we simply never had to be good at it, because capacity made most of our refusals for us. That crutch is gone. Cheap execution does not make choosing easier. It removes the last excuse for choosing badly.