What published data does and does not say about strategy execution
Public outlets were searched for surveys and studies on execution failure. None contains one. This piece sets out what they do cover and how to judge the evidence you are shown.

None of the public sources we checked contains a management survey or peer-reviewed study of why strategy fails to reach the front line. This article therefore cannot open with the figures a data brief usually leads with. What it can do is set out what the sources do say, where they fall short, and how to judge the evidence you will be shown instead. That is less than a leader would like, and it is the honest position.
What the sources cover
The World Bank's governance pages are the most substantial material we found. The Bank says its governance work seeks sturdy, answerable state bodies, and it reports that public trust in state institutions is declining. It lists three strategy priorities for governments: managing public financial and human resources, delivering public services, and assessing capacity on the basis of evidence. Each is a sound list for a finance ministry. None describes how a listed company or a multi-division group turns an approved plan into work.
The National Bureau of Economic Research item is narrower still. It reports that entrepreneurship training may help founders recognise when to stop investing in a venture, and it draws on a field experiment with small startups. That is useful for a founder deciding whether to close a product line. It says nothing about how a regional head of a fifteen-thousand-person business responds to a target set in headquarters.
The gov.uk and IFRS Foundation pages were searched in the same way and produced nothing on strategy execution. The research note refers to four sources but lists five outlets. We have treated all five as checked, and the finding does not change either way.
What the data shows
On the evidence available to us, the data does not show why strategy fails in large organisations, how often it fails, or which practices are linked to better follow-through. Anyone who tells you otherwise is working from a source we have not been able to see, and the first question to ask is what that source is.
What it means for a leader
The gap matters because execution failure is a claim many firms make about themselves, and it is easy to repeat without a source. In our editorial judgement, the causes that recur in board discussions are familiar. The plan is written in terms the people doing the work cannot act on. The budget and the plan were agreed in different meetings. Nobody owns the handover between a strategy team and the division that must deliver. These are judgements drawn from experience, not findings from a survey, and this magazine has no data that ranks them.
The first test for any execution study is what it measured. Does it track what happened to a stated plan against dated milestones, or does it record what executives said about the plan afterwards? Surveys of managers usually ask the second question, because it is cheaper and faster to ask. A study that follows named initiatives answers the first. Ask which one you are reading before you read the headline.
The second test is the sample. A survey of chief executives in one sector tells you about that sector's confidence, not about the typical rate of failure across firms. Check how many organisations were included, how large they were, where they operate, and who answered for each. A small, self-selected panel of executives will report a different picture from a random sample of business units, and the difference often matters more than the finding.
The third test is whether the causal claim is in the study or only in the press release. A correlation between a practice and better follow-through can be real and still tell you little about what to start on Monday. Where a study says a practice is linked to better results, read the method section to see whether firms chose the practice or were assigned it. Only the second kind gives you a reason to believe the practice causes the result.
The fourth test is the publisher. A research body with a published method and a named author team is in a different position from a vendor whose survey was run to sell a training course. Neither is automatically wrong. Both can be wrong in ways that flatter their interests, and the document should tell you enough to see which.
What to do about it
If your board intends to act on execution, do not start by buying a benchmark. Start by writing down the three or four commitments the board approved in the last twelve months. For each, record who owns it, when it was due, and what has changed since. That is your own data, collected from your own organisation, and it is the only kind that reliably answers the question you actually have.
Then ask the finance lead and the heads of the operating divisions to produce the same list independently, without seeing each other's version. Where the two lists disagree about the owner or the date, you have located an execution problem. You will not need a literature review to tell you it is important. The disagreement is the finding.
Keep external reading for judgement, not proof. When a consultant, a trade body or a conference speaker cites a figure on execution failure, ask for the publisher, the year, the sample and the exact question asked. If the answer is not in the document, treat the figure as an opinion. That is a fair description of most of what circulates on this subject.
Limits of the data
This piece rests on what the sources contain, and they contain no execution data. We have not filled the gap with figures from memory or from general industry reports, because a number we cannot attribute is worse than no number. The article is short because the evidence is thin, not because the subject is unimportant. Anyone who needs quantified answers about strategy execution should commission a survey with a published method, a stated sample and a named publisher, or find one that already exists. Once that evidence is in hand, the article should be rewritten as a proper data brief.

