Research famously suggests that the large majority of strategies fail — not in the thinking, but in the doing. The execution gap is rarely a strategy problem or an effort problem. It's a conversation problem, and that makes it fixable.
It is one of the most quoted statistics in management, and one of the most sobering: by various estimates, somewhere between two-thirds and ninety per cent of strategies fail to execute. Leaders respond to this the way they respond to most failures — by reaching for a better strategy, more resources, or more accountability. But look closely at where execution actually breaks down, and you rarely find a flawed plan or people who didn't care. You find a conversation that broke at a specific, identifiable point.
Organisations don't run on tasks. They run on a network of commitments that are made, kept, and closed in conversation. Coordinated action moves through a loop with four parts, and the execution gap is simply what opens when one of those parts is missing. The value of seeing it this way is that “why can't we get anything done” — an unanswerable complaint — becomes a precise diagnostic question: which move broke?
A real request names two things people routinely leave out — what would actually satisfy it, and by when. “Can you look into the pricing issue?” is not a request; it's a mood with a question mark. “I need the three pricing options costed, with a recommendation, by Thursday midday” is something a person can act on, decline, or renegotiate. Most coordination failures are seeded right here. Picture a team where the head of product asks an engineer to “tidy up the onboarding flow.” Three weeks later the engineer has rebuilt a screen the product head considered fine, and left untouched the drop-off point that was the actual concern. Nobody was lazy. The request never specified what “done” looked like, so two competent people filled the gap with different pictures.
A request is answered by a commitment — and a real commitment is one the other person could have declined. The quiet killer in most teams is the nod that isn't a yes: the agreement given to end a meeting, with no intention or capacity behind it. A culture where people can't say “no,” or “not by then,” or “yes, but only if we drop something else,” is a culture running on promises that were never promises. The leader who prides themselves on a team that never pushes back usually has a team that quietly doesn't deliver, because the pushback that would have surfaced the real constraint got swallowed. Make it genuinely safe to decline and renegotiate, and the promises that remain start to mean something.
A culture where people can't say “no” is a culture running on promises that were never promises.
Then the work happens — and then comes the step almost everyone skips: someone declares it complete, and someone declares themselves satisfied. This sounds like ceremony and is in fact the difference between work that feels finished and work that drags an invisible tail of ambiguity behind it. Without an explicit close, half the people involved assume the thing is done and half assume someone is still on it, and the cost surfaces a week later as a nasty surprise. The completion is a small act of language, two sentences long, and it closes a loop that otherwise stays open indefinitely, quietly consuming attention.
There is an uncomfortable corollary here for anyone in charge. The vague request, the unfreely-given promise, the unclosed loop — these are not just things that happen to a leader; they are frequently things a leader causes. The executive who fires off “can someone own this?” into a crowded thread, who reacts badly enough to bad news that people quietly stop bringing it, who never once models closing a loop themselves, is manufacturing the exact execution gap they later complain about. Which is, in its way, good news: the single largest lever on a team's execution is usually how its own leader handles these four moves. Tighten that, and the rest tends to follow, because people coordinate the way they are coordinated with.
The next time something important slips, resist the urge to conclude that the strategy was wrong or the people weren't committed. Walk it back through the loop instead. Was the request clear, with conditions of satisfaction and a real deadline? Was there an actual commitment, freely given, that could have been declined? Was completion declared, or left to assumption? Nine times out of ten the break is at one identifiable point, and naming it is most of the repair — because the fix for a vague-request problem is completely different from the fix for a nobody-can-say-no problem, and trying to fix execution “in general” is how leaders spend quarters solving the wrong thing.
None of this requires a new system, a new tool, or more meetings. It requires treating the conversations themselves as the place the work lives, and getting four moves right. Teams that learn to do it drop fewer balls, do markedly less rework, and end discussions with commitments instead of a shared, vague sense that something is now somebody's problem. The strategy was probably never the issue. The conversation was.