Why public programmes stall at execution, and what fixes it
The failure rarely sits in the strategy. It sits in the unglamorous space between the decision and its delivery, and that space can be engineered.
Ask why a public programme underperformed and you will usually be handed a story about strategy: the wrong target, the wrong design, the wrong assumptions. Occasionally that is true. Far more often, the strategy was sound and the programme simply stalled on its way to the ground. The decision was made, the money was approved, and then somewhere in the long handover from capital to district, momentum quietly drained away.
Stalling is the most common failure mode in public delivery, and the least examined. It does not produce a scandal. It produces a result that lands a little below target, a report that blames capacity in general terms, and a quiet resolution to try harder next cycle. To fix it, you first have to see it for what it is.
A stall is a process, not an event
Programmes rarely collapse in a single moment. They lose speed in increments. A disbursement is released but not confirmed received. A site falls a week behind, then a month, with no one positioned to notice. An agency waits on another agency that is waiting on a signature. None of these is fatal alone. Together, compounding over a delivery cycle, they are the difference between a programme that reaches its people and one that reaches a fraction of them.
Because the losses are gradual and distributed, they are easy to miss until the end, when they are no longer fixable. The work of preventing a stall is therefore the work of seeing it early, while there is still time to intervene.
The three places programmes break
Release. Money and authority move from the centre outward, but confirmation rarely moves back. The centre knows what it released; it often does not know what actually arrived, or when. That blind spot is where the first weeks of a programme silently leak.
Coordination. Most public outcomes depend on several institutions acting in sequence. When each holds its own records and none shares a common view, the seams between them become the slowest part of the system. Work waits at boundaries no single agency owns.
Verification. A milestone marked complete is not the same as a milestone done. Where completion is self-reported and rarely checked independently, the programme's official picture and its real one drift apart, and decisions get made on the wrong one.
“You cannot accelerate a programme you cannot see. Visibility is not a dashboard. It is the precondition for every correction you will ever make.”
What actually fixes it
The fix is not exhortation or another round of capacity-building in the abstract. It is operational infrastructure that closes each of the three gaps. A confirmation loop, so release is matched by proof of receipt. A shared operating picture, so coordinating agencies work from the same facts instead of reconciling four versions after the fact. And independent verification built into the flow of work, so the official status and the real one stay the same thing.
What these have in common is that they shorten the distance between something going wrong and someone with the authority to fix it knowing about it. A programme that surfaces a stalled site in days rather than months is not marginally better; it is a different kind of programme, because almost every problem it encounters is still recoverable when it is found.
Strategy decides what is worth doing. Execution decides whether it gets done. The programmes that deliver are not the ones with the cleverest plans; they are the ones built so that drift is visible early and momentum can be restored before it is lost.
