Who decides when a major programme changes?

Imagine a transport programme with several construction packages, a live rail network and a deadline for opening. One team discovers that a planned sequence will disrupt services for longer than expected. The people closest to the work can see an alternative, but approving it requires a chain of reports through the programme office and sponsor. By the time the decision arrives, the opportunity to change the sequence has passed.

This is often presented as a choice between central control and local freedom. For a major programme, neither is a complete design. The sponsor needs a coherent purpose, financial discipline and authority over decisions that affect the whole investment. Delivery teams need enough room to respond to conditions that cannot be specified in advance. The real question is who decides what, with whose information and at what speed.

Keep the decisions that bind the whole programme at the centre

Large programmes have reasons to concentrate authority. A public sponsor must account for expenditure and commitments extending beyond any one contractor or work package. It may have to balance competing demands from users, communities, regulators and funders. A local choice that improves one package can shift cost, risk or disruption to another.

The centre should therefore own the outcomes sought, the limits on cost and risk, and the rules for changing major commitments. It should decide between options that affect the programme as a whole. It should also make clear which benefits matter and who will be responsible for them after the temporary delivery organisation closes.

Central authority can be valuable when different parties need a common direction. The trouble begins when holding that authority is confused with making every operational decision. A sponsor who approves a strategic change is exercising governance. A sponsor who must authorise each adjustment to a construction sequence may be delaying the very information and action needed to deliver the strategy.

Design for the information the work will generate

Organisation theorist Jay Galbraith argued that greater task uncertainty creates a greater need to process information. His account of organisational design sets out several ways to cope: allow some spare capacity, make tasks more self-contained, invest in vertical information systems that carry information through the hierarchy, and create lateral relations through which peers resolve issues directly.

These choices are practical for programme leaders. A rail package with well-defined interfaces may be able to make many decisions within agreed boundaries. Where design, operations and community access are tightly connected, representatives from those areas need a way to resolve issues together, not just send separate reports upward. A shared dashboard may help them see the problem, but it cannot itself decide who has the authority to act.

The sponsor should ask where decisions are waiting for information, and where information is waiting for decisions. How many approvals does a routine change require? Can the team that discovers an interface problem reach the team that owns the adjacent package? Which exception must be escalated because it threatens an agreed outcome or risk limit? The answers should shape the programme organisation before delay becomes normal practice.

Give local teams discretion with clear boundaries

Decision rights need to be specific. A programme might allow the delivery team to resequence work within agreed cost and disruption tolerances, while requiring sponsor approval when a tolerance is breached or a change alters public commitments or expected benefits. The limits should include who must be consulted and when a decision is recorded or escalated.

This is different from telling teams to be agile. An instruction to adapt is of little use if contracts reward only compliance with an initial plan, if people cannot share information across organisational lines, or if raising bad news is treated as failure. Local discretion works when teams know the purpose of the programme, understand the limits of their authority and can obtain help promptly when a problem crosses those limits.

Amy Edmondson and Jean-François Harvey’s work on cross-boundary teaming is useful here. People with different expertise have to work together across knowledge and organisational boundaries, which can make collaboration difficult even when everyone wants the same result. In a major programme, that work may span agencies, operators, consultants and contractors. A formal reporting line alone will not create the relationships needed to solve shared problems; people need occasions to exchange what they know and test assumptions together.

Nor does a flexible team remove the need for leadership. Someone still has to settle conflicts, protect programme-wide interests and make a decision when the participants cannot agree. Local action and central accountability can support each other when the boundaries between them are deliberate.

Treat the original plan as a decision, not a prophecy

A programme begins with a strategy, but delivery produces information the original plan could not contain. A sponsor may learn that users value a different service pattern, a technical interface is harder than expected, or an early benefit can be delivered in another sequence. Refusing to reconsider the plan wastes that information. Changing it without discipline can lose the purpose for which the programme was approved.

The programme needs a way to distinguish adaptation from drift. For a proposed change, ask what has been learned, which original assumption has failed and whether the expected benefits still justify the revised cost and risk. Record the decision and its effect on other packages. If the change is significant, return it to the sponsor with genuine options rather than a request to endorse an outcome already made inevitable. When a major project’s costs rise, who asks whether it still makes sense? examines that continuing decision in more detail.

Make learning survive the programme

Temporary teams often develop knowledge that disappears when contracts end. The people who found a way to manage an interface move on; the permanent operator inherits the asset without the reasoning behind its design. If local learning is to have value beyond the current delivery problem, someone must capture what was decided and transfer it to those who will maintain the result.

The same principle applies to the sponsor. It should compare forecasts with actual outcomes, including changes in scope and the reasons for them, so that the next programme begins with a better outside view. Information should travel sideways during delivery and forward into operations and future decisions.

Return to the transport team facing a missed window for changing its work sequence. The failure is not necessarily that authority sat with the sponsor. It is that the programme had no timely route for a bounded local decision or rapid escalation of a programme-wide one. The design of decision rights determined what the team could do with what it knew. That is where governance becomes real.

Further reading

• Jay R. Galbraith, Organization Design: An Information Processing View (Interfaces, 1974).
• Amy C. Edmondson and Jean-François Harvey, Cross-Boundary Teaming for Innovation (Harvard Business School Working Paper No. 17-013, 2016; revised 2017).