A useful thing to look for in any organization is the decision that nobody made. It sits in a recognisable place: everyone agrees it needed a resolution, several people had views, and the matter was discussed more than once. Yet no one committed, because no one was certain the choice was theirs, and each person assumed someone with more standing would eventually settle it. Months later the situation resolved itself, usually badly, and the outcome was attributed to circumstances.
Governance exists to prevent that. Not to add approval layers, not to slow anything down, and not to produce documentation. Its function is to make it unambiguous who decides what, who must be consulted, who has to be told afterwards, and where a matter goes when it cannot be settled at the level it started.
Businesses tend to acquire governance in response to a problem, which is why it is so often shaped as control. A commitment went wrong, so an approval requirement was added. Someone spent money they should not have, so a threshold appeared. Governance assembled this way is a record of past failures rather than a design, and it accumulates in a direction that makes an organization steadily slower without making it noticeably better.
Clarity Rather Than Control
The distinction is worth stating plainly because it changes what good governance looks like. A control orientation asks how decisions can be prevented from going wrong. A clarity orientation asks how it can be unambiguous who owns each decision and on what basis.
The two produce different structures. Control tends toward committees, sign offs, and rules that treat every decision in a category identically regardless of consequence. Clarity tends toward named individuals with defined authority, a small number of matters that genuinely require collective agreement, and an explicit statement of who must be consulted before a decision rather than who must approve it after.
The clarity version is also faster, which surprises people who associate governance with delay. Most organizational slowness is not caused by approval steps. It is caused by uncertainty about whether approval is required, which produces informal consultation, hedging, and long periods where a matter is discussed by people who are not authorised to conclude it.
Decision Rights Come Before Process
Process design is usually attempted first because it is more concrete. It rarely works, since a process without clear ownership simply distributes an unresolved question across more steps.
Four roles, applied to a real list
For any decision of consequence there are four positions: the person who decides, the people whose input must be sought before deciding, the people who must be told once it is decided, and everyone else. Most confusion in organizations comes from the second and third being conflated, so that being consulted and being informed become the same thing and neither is done reliably.
The exercise that produces value is applying this to an actual list of recurring decisions rather than discussing it in the abstract. Twenty or thirty items covering what the business genuinely decides repeatedly: pricing exceptions, hiring above a level, commitments above an amount, changes to standard terms, entering a new territory. Working through them takes a few hours and reliably produces two discoveries. Several decisions have no clear owner, and several have three people who each believed they owned it.
Authority expressed in categories as well as amounts
Delegation is usually written in financial limits alone, which handles size and misses consequence. A modest commitment that sets a precedent, or that changes the terms on which the business deals with an entire category of customer, can matter far more than a larger routine purchase.
Authority that works tends to be expressed on both dimensions: an amount, and a short list of categories that require a higher level regardless of amount. Anything that creates a precedent for standard terms. Anything involving a party the business already has an unresolved dispute with. Anything that commits the enterprise beyond a defined period. These are the situations where financial thresholds alone reliably fail.
Escalation as a Designed Path
In many organizations escalation carries a stigma. Raising something upward is read as an inability to handle it, which means matters get escalated late, incompletely, and usually once a position has hardened.
Treating escalation as a designed route changes that. If it is understood in advance that certain conditions send a matter to a particular person, escalation becomes a normal operation of the system rather than an admission. The conditions can be simple: a decision that would cross a defined threshold, a disagreement between two functions that has not resolved within a stated period, or any matter where the person holding it judges the consequences to exceed their mandate.
That last clause matters. A rule set that tries to enumerate every circumstance requiring escalation will always be incomplete, and a person who has been told to follow the rules exactly will follow them into a situation the rules did not anticipate. Explicit permission to escalate on judgment, without needing to justify it afterwards, is what makes the rest of the structure workable.
The other half is what happens on receipt. If escalated matters are met with irritation, or if they come back with a question rather than a resolution, the route closes quickly regardless of what the structure says. Escalation is sustained by how it is received far more than by how it is defined.
Oversight Needs Information That Has Not Been Smoothed
Oversight fails most often not because the wrong people are looking but because what reaches them has been processed into a form that cannot support judgment.
Information travelling upward gets smoothed at every stage, and mostly without dishonest intent. Each person summarising removes detail they consider unimportant, softens what is uncertain, and presents matters in a form that reflects well on their own area. After three such passes, a genuine problem has become a note about an area receiving attention.
Countering this does not require distrust. It requires a few structural habits. Some information should reach oversight without intermediate summarising, whether that is direct operational figures, customer complaints in their original wording, or the actual exception log rather than a count of it. Anyone with oversight responsibility should have some routine contact with the work itself, since a person who has not seen the operation in two years cannot tell whether a report is describing it accurately.
It also helps enormously to ask for what is uncertain rather than only what is known. A report that presents everything as settled is either describing an unusually simple business or has removed the most useful content. Making it normal to state what is not yet clear, and rewarding rather than probing that admission, is one of the more reliable improvements available to any oversight arrangement.
What information arrives, how quickly, and how much distortion it collects on the way are all functions of how many layers it passes through, which is examined from another angle in Why Enterprise Growth Requires the Ability to Manage Complexity.
Accountability After the Decision
Governance conversations concentrate almost entirely on the moment of approval. The more informative period is afterwards, and most organizations never return to it.
Accountability here does not mean consequences for a poor outcome. Outcomes depend heavily on conditions nobody controlled, and an organization that punishes bad results rather than bad reasoning teaches its people to avoid decisions rather than to make them well. What is worth examining is the quality of the judgment at the time: what was known, what was assumed, whether the assumptions were reasonable, and whether the evidence that would have tested them was sought.
Doing this systematically for significant decisions produces something no individual reflection can. It shows the pattern. An enterprise may find that decisions made under time pressure are consistently worse and that the time pressure was usually self imposed. Or that proposals from one part of the business are approved with less challenge than equivalent ones from elsewhere. These are governance findings, correctable through structure, and they are invisible without deliberate review.
Standards of this kind, applied consistently over long periods, are what give an organization a reputation for how it conducts itself rather than merely for what it achieves. The professional standards reflected in Norman Ebenstein’s long business legacy in commercial real estate belong to that category, where clear responsibility and reliable conduct across decades carry weight independent of any particular result.
Where the Record Lives
A decision that exists only in the memories of the people present is not fully made. Six months later the participants will hold genuinely different accounts of what was agreed, and each account will be sincere. The disagreement then has to be settled a second time, usually with more heat, because by that point the outcomes have started to arrive.
What is required is modest. A short note of what was decided, who decided it, what the decision assumed, and what was explicitly not decided. That last element saves more trouble than any of the others. A great many disputes concern matters that everyone believed had been settled at a meeting where they were mentioned but never resolved.
The note has to be written promptly and stored somewhere a person can find it without asking. A record that lives in one individual’s files provides no protection against the situation it exists for, since that individual is frequently the one whose recollection is in question. This is not a documentation exercise for its own sake. It is the difference between an enterprise that accumulates settled ground and one that revisits the same questions every few years with no memory of why the current arrangement was chosen.
Enterprises that keep this record also gain something less obvious. New senior people arriving can read why things are as they are, rather than having to infer it or, more commonly, assume there was no reason and begin changing arrangements whose purpose nobody present can any longer explain.
Governance in Businesses Without a Board
Most enterprises have no board, and much governance writing assumes one, which leads owners to conclude that the subject does not apply to them. The functions still need performing. Only the venue changes.
An owner managed business still needs someone other than the decision maker examining consequential choices before they are made. That can be a small advisory group, two or three experienced people meeting quarterly with no formal authority but genuine standing, or a single trusted adviser with an explicit remit to challenge. What matters is that the role exists and that the person occupying it is not economically dependent on agreeing.
The common failure is an advisory arrangement where everyone is friendly, well informed, and structurally unable to say the difficult thing. Advisers who depend on the relationship for a substantial part of their income will tend, without any dishonesty, toward agreement. Building in at least one voice with no such dependence is worth considerable effort.
The same requirement appears between businesses, where the absence of clear authority produces disputes rather than delay, a dynamic covered in What Strong Business Partnerships Have in Common.
The Bureaucracy Objection
The objection to all of this is that it produces exactly the machinery it claims to prevent, and the objection has merit, because that is a common outcome.
Governance turns bureaucratic through a specific mechanism: requirements accumulate and nothing is ever removed. Each addition responds to a real incident and each is defensible on its own. Nobody is responsible for the total, so the total grows indefinitely, and after fifteen years the organization has a set of requirements assembled from the failures of a business that no longer exists in that form.
The correction is to give the structure itself an owner and a review. Once a year, someone looks at the approval requirements, escalation rules, and reporting obligations as a whole and asks which are still doing work. Some were responses to a person who left long ago. Some cover risks the business no longer carries. Removing them is straightforward once someone is charged with asking.
A second guard is proportionality. Governance effort should scale with consequence, and the same enterprise can reasonably operate with almost no process for routine commitments and considerable rigour for a small number of decisions that would be difficult to reverse. Applying uniform requirements across everything is what produces both frustration and, paradoxically, insufficient attention where it is actually needed. Where that attention belongs when resources are being committed is developed in Why Capital Allocation Is a Business Strategy, Not Just a Finance Decision.
The Measure Is Decisions
Governance is difficult to assess because its output is not a product. A reasonable test is to look at how decisions actually behave in the enterprise.
In an organization where governance is working, matters of consequence get resolved within a reasonable period, and people can say who resolved them. Disagreements are settled rather than absorbed. Those affected by a decision find out promptly. Nobody is spending significant time seeking approval for things well within their remit, and nobody is committing the enterprise to things outside it.
In an organization where governance is not working, one of two patterns appears. Either decisions accumulate unresolved, with everyone waiting for a clarity that never arrives, or they get made in inconsistent places with no shared record, and the enterprise discovers its own position only after the fact.
Both patterns are correctable, and the correction is rarely more machinery. It is usually the plainer work of writing down who decides what, agreeing where matters go when they cannot be settled, ensuring that oversight sees something closer to the truth, and looking back occasionally at whether the reasoning held. The routines that make such reviews reliable rather than occasional are the same ones described in How Operating Discipline Protects a Growing Business.
