Why this matters
Most organisations never decide how decisions will be made. They inherit it. Early on, authority sits close to the work: whoever needs to decide something is within a short conversation of whoever can approve it, and the arrangement is efficient precisely because it is unwritten.
Growth removes the conditions that made this work. Headcount rises, sites multiply, functions specialise, and the short conversation becomes a scheduled one. The implicit allocation of authority does not disappear; it persists, unexamined, while the organisation around it changes shape. What was a sensible shortcut becomes an unstated rule that different parts of the business now read differently.
The cost rarely presents as a single visible failure. It presents as a tax. Decisions take longer than their content warrants. Matters that were settled reopen, because nobody can point to where they were settled or by whom. Work stops at boundaries between functions, or between a head office and a site, while each side waits for the other to move. Escalation reaches senior level not because the question is senior, but because nobody below is confident they hold the authority to close it.
Because the symptoms are behavioural, the diagnosis usually is too. Slow decisions get attributed to particular individuals, to a cautious culture, or to a leadership team that has not communicated clearly enough. Those explanations are sometimes right. They are also the ones most readily to hand, and they point away from structure. In practice, organisations often find the same ambiguity survives a change of personnel — the observation that makes the structural reading worth taking seriously.
Where decision rights come from when nobody writes them down
Unclear decision rights are almost never absent decision rights. Authority is allocated continuously in every organisation; the open question is by what.
Where no explicit allocation exists, several mechanisms do the work. Precedent is the strongest: whoever decided something similar last time acquires a claim to decide it again. Proximity is next, and the person organisationally closest to the executive team acquires authority that was never conferred on them. Budget custody carries decision rights whether or not anyone intended it to, and so does possession of information, since the person holding the numbers frequently ends up making the judgement the numbers were meant to inform. There is usually also some allocation by persistence, where a matter settles in favour of whoever keeps raising it.
None of these is illegitimate on its own and all are fast. Their weakness is that they neither scale nor travel: they work only while the population of decision-makers is small enough to carry the same unwritten map. Once it is not, parts of the organisation operate from different maps, each internally consistent and each derived honestly from local experience. This is the point at which ownership of decisions becomes unclear without anyone having changed anything.
Jay Galbraith, writing in Interfaces in 1974, framed organisation design as a problem of information processing: the more uncertainty a task carries, the more information has to be handled while the work is under way, and organisational form is largely a response to that load. Applied to authority rather than workflow, the observation is uncomfortable. The informal channel that carried decisions in the early period was never designed for the volume it eventually absorbs. It does not announce when it reaches its limit. It degrades quietly, and differently in different parts of the organisation.
Larry Greiner’s account of organisational growth, published in Harvard Business Review and later reissued as an HBR Classic with his own commentary, makes a related point about time rather than load: the arrangement that resolves one period’s difficulty tends to generate the next one. Delegation introduced to relieve an overloaded centre creates, in due course, a problem of control that delegation itself cannot answer. The mechanism is not broken when this happens; it has been outgrown, which is a different condition and is routinely misread as the first. It is also part of why founders struggle to let go of decision-making in ways that look like reluctance but are often an accurate reading that the replacement is not yet there.
The consequence is duplication and gaps inside the same system: two functions both believing a decision is theirs in one place, neither believing it in another. Both errors share an origin, which is why treating them as separate problems tends not to work.
A question worth sitting with: when a decision in your organisation is made well, is it clear afterwards why that person was the one who made it — and would a colleague in another function give the same answer?
The written rule and the working rule
Organisations that recognise the ambiguity often respond by writing things down. Policies are drafted, thresholds stated, approval routes circulated. It is a reasonable response, and it produces a second problem that is less often anticipated: the written rule and the rule actually in force begin to diverge, and the divergence is invisible from the document.
Consider an organisation where an expenses policy specifies a threshold above which pre-approval is required. In practice most claims above that threshold are submitted retrospectively, and approved. At face value this looks like a weak compliance culture. What complicates that reading is that retrospective approval has never been refused. Staff have not misunderstood the policy. They have correctly inferred, from consistent evidence, that it describes a preference rather than a requirement. The policy as written and the policy as enforced have separated, and the enforced version is the operative one. Nobody chose it.
The pattern is not confined to expenses and is not principally about rule-breaking. Erik Hollnagel, Robert Wears and Jeffrey Braithwaite, in a white paper published jointly by the University of Southern Denmark, the University of Florida and Macquarie University, drew a distinction between work as imagined by those who specify it and work as done by those performing it. Their argument, developed in safety-critical settings, is that the gap is not simply an aberration to be eliminated: everyday work involves continuous adjustment to conditions the specification could not anticipate, and outcomes are good largely because of those adjustments rather than in spite of them. Governance documents describe imagined work. The decisions people actually take are done work.
Steven Kerr’s 1975 paper in the Academy of Management Journal, “On the Folly of Rewarding A, While Hoping for B”, supplies the mechanism by which the two settle into a stable divergence rather than correcting. Kerr’s observation was that people work out what is actually rewarded and orient towards it, often to the exclusion of what is merely stated. No reward is paid here in any conventional sense, but an approval that is never refused is a signal, it is repeated, and it is more credible than the document because it is evidenced. The organisation has taught its own rule without deciding to.
The consequence is that such an organisation now runs two governance systems. One is documented and auditable, and describes decision rights a board or an external reviewer would recognise. The other is undocumented, learned by observation, and determines what actually happens. Only the first can be inspected. When the two are finally compared — usually after an incident, an audit or a change of leadership — the discovery is rarely that the rules were unclear. It is that they were clear, and different. This is also why informal governance works until it doesn’t: nothing fails while the two versions agree.
The common misconception is worth naming plainly. Divergence of this kind is usually read as a discipline problem, which locates the cause in individual behaviour. The evidence more often points to people behaving rationally in response to the signals available to them. In practice, organisations often find that the population changes and the pattern does not.
“The enforced version of a rule is the real one, and it is rarely the version anybody chose.”
— Zoom Riddle
A question worth sitting with: which of your organisation’s stated rules would survive being tested — and what would the last twelve months of actual practice say the rule is?
Why the ambiguity is hard to see, and harder to raise
Ambiguity about decision rights is unusual in that the people closest to it are often the least able to name it.
The first is that it presents as its own consequences. What a leadership team notices is not that decision rights are unclear, but a scatter of surface observations: agendas that repeat, closed matters reopening under a different heading, escalations arriving without a recommendation, approval sought for things that plainly did not need it alongside significant commitments never surfaced at all. Each has a plausible local explanation. Collectively they have a common one, but seeing it means holding them together, and they are reported separately, by different people, at different times. It is also how reporting lines become confused during rapid growth: the formal chart is intact, and the traffic on it no longer matches.
The second is that the ambiguity is asymmetrically visible. Someone unsure whether a decision is theirs feels that uncertainty acutely. The senior team receiving the escalation sees something quite different: a colleague who appears reluctant to take responsibility. The same event supports two incompatible readings depending on where in the structure you stand, and the more senior reading is the one that gets recorded.
The third is that naming it carries risk. Amy Edmondson’s study of work teams, published in Administrative Science Quarterly in 1999, found that psychological safety within a team predicted learning behaviour — the raising of questions, errors and doubts — in a way that the team’s confidence in its own capability did not. Saying “I do not know whether this is mine to decide” can be heard as an admission of inadequacy in one direction and an accusation of poor management in the other. Silence is the lower-risk option, and it reads upward as clarity. In practice, organisations often find the question surfaces only in exit interviews, or after something has gone wrong and the account has to be reconstructed.
There is a genuine tension underneath, and it does not resolve cleanly. Any tightening of decision rights makes some decisions slower and some people less able to act; any loosening makes the organisation faster and less predictable. Both directions carry real costs, they fall on different groups, and the trade-off is not the same in a twelve-person team as in a four-hundred-person one. It is closely related to the difference between oversight and bureaucracy, likewise a distinction that only exists in a particular context. Zoom Riddle does not treat that tension as a problem with a general answer, and this guide does not offer one.
Context changes the shape of the problem more than is usually assumed. In professional and knowledge-based organisations, authority attaches to expertise as much as to position, so the formal structure and the operative one can diverge without anyone perceiving a contradiction. In multi-site operations distance does the work, and sensible local adaptation is hard to distinguish from erosion of a common standard. In regulated sectors some decision rights are not the organisation’s to allocate at all, which can produce false confidence that the remainder is equally well defined. The same dynamic underlies the alignment gap between leadership and the front line, where the decision a leadership team believes it has delegated and the decision the front line believes it holds are not the same decision.
A question worth sitting with: in the last decision your organisation reopened, was the disagreement about the substance — or about who was entitled to settle it?
What the research shows
The research here comes from traditions that do not usually cite one another, which makes the points where they converge more interesting than those where they agree by construction.
The most direct evidence comes from Paul Rogers and Marcia Blenko, writing in Harvard Business Review in 2006 on the basis of a Bain survey of executives at more than 350 global companies. Only about 15 per cent of those surveyed said they had an organisation that helped the business outperform its competitors. Their descriptive contribution is that decision-making stalls at recognisable seams rather than randomly — between the centre and a business unit, between one function and another, between global and local, and at the boundary with outside partners. That locates the problem at interfaces rather than in individuals. Their article also advances a proprietary model for assigning decision roles; that model is not set out or examined here.
A second body of work approaches the same territory from execution rather than governance. Donald Sull, Rebecca Homkes and Charles Sull, drawing on a multi-year study of more than 250 companies and several thousand managers, reported in Harvard Business Review in 2015 that only 9 per cent of managers said they could rely on colleagues in other functions and units all of the time, with around half saying most of the time. Their broader conclusion is the one that matters most here: the dominant execution problem in the firms they studied was not vertical, since alignment between levels was largely sound, but horizontal. Coordination across units was where commitments failed.
Read together, the two point the same way from different starting positions. The seams Rogers and Blenko identify and the lateral unreliability Sull and colleagues measure are arguably the same phenomenon seen through different instruments. Galbraith and Greiner supply plausible reasons why it should intensify with scale, the territory covered in more detail in why decision rights become blurred in growing companies. Hollnagel and colleagues, working in a separate field, explain why the documented allocation and the operative one drift apart even when both are held in good faith.
The picture is less settled on direction and magnitude. Rogers and Blenko report an association between clear decision roles and organisational performance, but an association measured through executive self-report is not a demonstrated causal effect, and organisations that decide well may be better at many things for reasons clarity of decision rights does not exhaust. Their evidence is also drawn largely from very large firms. The finding on lateral coordination is stronger as a description than as a diagnosis, since it does not by itself separate authority that is unclear from authority that is clear but contested, or clear and simply not honoured.
What remains genuinely open: whether ambiguity in decision rights is better understood as a cause of coordination failure or as one of its symptoms; whether the divergence between documented and operative authority is a defect to be closed or, as the safety literature would suggest, partly a load-bearing adaptation; and at what point in an organisation’s growth the informal allocation stops being an asset. None has a settled answer, and organisations that claim otherwise are usually describing their own case rather than a general finding.
Looking for your own sector? The Other Sectors page points to the closest match.
Questions worth considering
- In the last decision that was reopened, was the disagreement about the substance or about who was entitled to settle it?
- Which of the organisation’s written rules would survive being checked against the last twelve months of what actually happened?
- Where an informal arrangement is doing the work of a rule, can anyone say whether it is an asset to protect or a risk waiting to surface?
This guide sits within Governance. A related guide: Why Nothing Ever Leaves the Priority List.
Your team is facing similar challenges?
Subscribe to The Riddle Ledger
Perspectives are sent as they are published, with no promotional email, and you can unsubscribe at any time.