Scope does not creep: it is moved by people, one reasonable accommodation at a time, and in most organisations nobody has established who is permitted to move it.
Ask the sponsor of a large program a simple question: who is allowed to change what this program is for?
The answers come back in a predictable order. First the sponsor themselves. Then the steering committee. Then, after a pause, a qualification — that certain stakeholders would obviously need to be consulted. Then a longer pause, and a list that grows as the sponsor works through the people whose objections have, in practice, changed the program's direction over the past year.
By the end of that list, the honest answer is usually a good deal longer than the governance documentation suggests — and most of the people on it have never been told they hold that authority, nor do they exercise it through anything the organisation would recognise as a decision.
This is not a documentation gap. It is a governance condition, and it is the mechanism behind most of what organisations describe as scope creep — a phrase that conveniently implies scope moved by itself.
The Strategic Context
Stakeholder management has been thoroughly professionalised. The artefacts are familiar: registers, influence-interest grids, engagement plans, communication matrices. They are produced diligently and they answer a genuine question — whose attention must we manage?
What they do not answer is the question that governs outcomes: whose view of the intended result actually prevails?
Portny offered a classification that cuts closer to this than the usual grids. He sorted the people around a project into three groups. Drivers hold some degree of say over what results the work is meant to deliver. Supporters enable or carry out the work itself. Observers take an interest without doing either.
Portny's own wording carries a deliberate qualifier: drivers define the intended results only to some extent. That hedge should not be sanded off. He is describing influence over the definition of the result, not formal authority. The extension proposed here, that this classification should be hardened into explicit decision rights, is ERANORTH's argument rather than his. But the classification is the right foundation for it, because it sorts people by the one property that determines whether a program delivers what the enterprise intended: their relationship to the definition of success.
What Leaders Commonly Misread
That influence and interest are the useful axes. They describe attention, not authority. Two stakeholders can be identical on both axes and differ entirely on whether their objection can change the program's purpose. That difference is the one that matters, and the standard grid cannot express it.
That being consulted and being a driver are different things. In practice they converge. When a stakeholder's concerns have repeatedly reshaped a program's objectives, that stakeholder is a driver, whatever the governance documentation says. Organisations frequently maintain a formal decision structure alongside an informal one that actually operates, and manage the program through the second while reporting against the first.
That more drivers means better alignment. It means the opposite. Every additional person who can redefine the intended result multiplies the number of definitions the program must simultaneously satisfy. Beyond a small number, the program stops having a purpose and starts having a negotiated position, which is a different and much weaker thing — it cannot be tested, and it cannot be used to refuse anything.
That observers are harmless. Observers become expensive when they are treated as drivers out of courtesy. A senior figure with genuine interest but no accountability for the outcome will offer views, as anyone would. Whether those views change the program depends entirely on whether the organisation has established that they should not.
Reframing the Issue
The reframe: this is not a stakeholder-analysis exercise. It is an allocation of decision rights, and it should be made with the same deliberateness as any other allocation of authority.
Decision rights are ordinary in other domains. Boards define delegations. Finance functions set approval limits. Nobody would accept an arrangement in which twenty people could commit capital because they had strong views about it. Yet the equivalent arrangement for program purpose — the thing that determines whether committed capital produces anything of value — is common, and is usually described as being collaborative.
The reframe carries an uncomfortable implication. Naming drivers means naming, explicitly, that certain senior people are not drivers. That is a political act, and it is why the work is so often avoided. But the avoidance does not remove the decision; it delegates it to whoever is most persistent.
Where the Categories Blur, and What It Costs
Public-sector programs show the pattern with unusual clarity, because their stakeholder environment is genuinely crowded and legitimately so.
Consider a hypothetical program to reform how a service is delivered across a jurisdiction. The accountable department is a driver. The minister's office holds a legitimate claim to define what the reform is for. Central agencies impose constraints. Delivery partners are supporters with strong views. Peak bodies, unions, oversight bodies and affected communities all have real standing. Several of these can, in practice, cause a change of direction.
The failure mode is not that the environment is complex. It is that complexity is used as a reason not to make the distinctions. When everyone is treated as a driver, three things follow reliably:
The program's objectives become additive. Each accommodation is individually reasonable. Collectively they produce a scope that no capacity plan can support and no benefits case can justify.
Trade-offs stop being made. Trade-offs require someone with standing to say that one objective outranks another. Where authority is diffuse, the program cannot decline anything, so it accepts everything and fails at the margin.
Escalation loses meaning. Escalation assumes a level at which a matter can be settled. Where the drivers are not defined, escalation becomes circulation — the issue moves between forums, acquiring commentary, until it is resolved by exhaustion or by a deadline.
The same dynamic operates in private enterprise; it is merely less visible, because the stakeholders are internal and the negotiation happens in corridors rather than in submissions.
Decision Framework
A method for establishing decision rights on a program, applied at initiation and re-tested at each major boundary.
| Step | Action | Test of a good answer |
|---|---|---|
| 1. List | Everyone whose objection has changed direction in the last year | Based on evidence of what happened, not on the governance chart |
| 2. Sort | Driver, supporter or observer | Every person in exactly one category |
| 3. Constrain | Reduce drivers to the minimum defensible number | Ideally under five; each must carry accountability for a benefit |
| 4. Test accountability | Does each driver bear consequences if the result is not achieved? | If not, they are an observer with influence — reclassify |
| 5. Declare | Tell people their classification | The uncomfortable step; skipping it voids the exercise |
| 6. Route | Define where a driver-level disagreement is settled | A named forum with authority to decide, not to discuss |
| 7. Re-test | Reconfirm at each phase boundary | Drivers change as programs move between phases |
Step 4 is the sharpest instrument here. Driver status should follow accountability for the outcome, not seniority. An executive who will not be answerable for whether the benefit is realised is not a driver, however senior — and an organisation that cannot enforce that distinction has not allocated decision rights at all.
From Strategy to Execution
Immediate. For the enterprise's most contested program, reconstruct the last twelve months of direction changes and identify who caused each one. That list is the real driver set. Compare it with the governance documentation; the gap between them is the governance debt.
Medium term. Make driver designation part of program initiation, alongside the sponsor and the benefits owner. It should appear in the same document, be approved by the same authority, and be as difficult to change informally.
Long term. Build the organisational norm that senior interest and decision authority are separate things. This is the hard part, and it cannot be achieved by documentation. It is established by what happens the first few times a program declines to accommodate a senior observer, and whether the sponsor is supported when it does. [Related article: Accountability Without Authority: How Organisations Design Delivery Leadership to Fail]
Signals to Monitor
- Objectives that only ever get added. A program whose scope has never contracted has no functioning refusal mechanism.
- Escalations that return without resolution. Evidence that no forum holds settlement authority.
- Direction changes with no traceable decision. The clearest sign that drivers exist outside the formal structure.
- Steering committees that have grown steadily. Past a certain size a committee tends to become a briefing audience, and real decisions migrate elsewhere; the trend is more telling than any particular threshold.
- Sponsors describing their role as building consensus. Consensus-building is valuable; where it replaces deciding, the program has no driver at all.
Questions for the Leadership Team
- Who is permitted to change what our largest program is for — by name, and does that list match who has actually changed it?
- For each named driver, what happens to them personally if the intended benefit is not realised?
- Where does a genuine disagreement between two drivers get settled, and how long does that take?
- When did this program last decline a senior stakeholder's request, and what happened next?
- Have we told people their classification, or only recorded it?
Closing Perspective
Programs do not usually fail because the work was done badly. They fail because the definition of what constituted success moved while the work was under way, and moved through a process nobody designed and nobody owned.
Deciding who holds that authority is unglamorous and politically costly, and it is done in an afternoon. The alternative is not neutrality. It is an unallocated authority that will be exercised anyway, by whoever is most present when the question arises.
Related article: What a Stage Gate Is Actually For
Related article: Accountability Without Authority: How Organisations Design Delivery Leadership to Fail
Related article: Whose Knowledge Does Your Governance System Actually Hear?
About the author
Kevin Jogin is Founder & Principal Advisor at EraNorth. Meet the Founder.