Portfolio processes are built around one question: what should we start? Intake forms, business cases, scoring models and approval boards all point in the same direction. For organisations whose portfolios hold more work than their people, capital and leadership attention can carry, the more valuable question is the opposite one. What should we stop, pause or defer so that the rest can succeed?
This brief argues that failing to stop is rarely a failure of analysis: weak projects are usually visible well before anyone ends them. They continue anyway, because incentives reward starting, psychology rewards persisting and governance is designed to approve rather than to end. Stopping well therefore has to be designed: criteria agreed before they are needed, reviews that look only at remaining cost and remaining value, evidence that is independent of the project’s sponsors, and an executive forum that holds the explicit right to reallocate.
It is the discipline that follows from the shift described in our brief on why PPM is moving beyond project selection: once a portfolio is managed from first decision to realised value, stopping becomes part of its normal operation rather than an admission of failure.
In brief
- Portfolio prioritisation is incomplete until it decides what not to do. Ranking without stopping produces a queue, not a set of priorities.
- Stopping is rare. Citing government data, the UK National Audit Office reported that of 184 major projects under way in March 2021, four had been stopped or closed early by March 2022.
- Research on the sunk-cost effect and escalation of commitment explains why: people and organisations keep investing in failing courses of action, most strongly when they were responsible for the original decision.
- The alternative to deliberate stopping is often not continuing everything but across-the-board cuts, or smaller projects delayed to absorb cost increases on larger ones, patterns the NAO has documented.
- Stopping becomes routine when stop criteria are set at approval, reviews ignore sunk cost, funding is released in stages and a named executive forum owns the decision.
Why portfolios fill up
The National Audit Office put the core difficulty plainly: prioritisation is challenging because it means deciding not to do some things which may be desirable. In its 2024 review of government planning and spending, it described the incentives behind overload. Departments have reason to seek funds for as many programmes as possible; governments have strong incentives to announce new programmes; and there are weaker incentives, and real obstacles, to cancelling programmes that are not working or represent a lower-value use of resources (NAO, Lessons learned: a planning and spending framework, 2024).
Our reading is that the same asymmetry exists in most large organisations. Every new initiative has a sponsor and a story. A project that is stopped has neither, and the executive who proposes stopping it absorbs the conflict. PMI’s 2026 research describes the organisational side of this friction as unclear decision-making authority, siloed teams, competing priorities and misaligned objectives (PMI, Pulse of the Profession 2026).
The consequence is familiar to anyone who has run a portfolio office. Robert Cooper, creator of the Stage-Gate process, observed that many companies suffer from too many projects and not enough resources to do them well, leaving gridlock in the development pipeline so that not much gets through. Gates, he noted, can kill some weaker projects and free resources, but the fuller answer often lies in resource capacity analysis and portfolio management (Cooper, Journal of Product Innovation Management, 2008). When every project draws on the same scarce specialists and the same executive attention, adding one more slows many.
How rarely organisations actually stop
The UK’s Government Major Projects Portfolio (GMPP) offers unusually transparent data. Citing the Infrastructure and Projects Authority, the NAO reported that at March 2022, of the 184 projects under way 12 months earlier, four had been stopped or closed early and seven had been replaced by other programmes. Its conclusion was direct: government does not often stop programmes (NAO, Lessons learned: Resetting major programmes, 2023).
The latest annual report adds detail. Of the 42 projects that left the GMPP during 2025–26, 26 reported delivery against their objectives, eight no longer met the criteria for inclusion, one was replaced by another GMPP project and seven were brought to early closure (NISTA, Major Projects Annual Report 2025–26).
These figures need careful reading: early closure is not the same as failure, and a low stopping rate does not prove that more projects should have been stopped. But private-sector evidence points the same way. Cooper reported that the most common complaint about stage-and-gate systems was gates without teeth: projects are rarely killed, and once approved, never are. In one high-technology manufacturer he described, passing the first gate put a project into the business unit’s financial forecast, after which it was locked in and every later gate was a rubber stamp. The idea-to-launch process, he argued, should be a funnel, not a tunnel.
The evidence on why projects continue
Several decades of research explain why stopping feels harder than starting. None of it suggests that executives are careless; it suggests that continuing is the path of least psychological and political resistance.
The sunk-cost effect
Hal Arkes and Catherine Blumer defined the sunk-cost effect as a greater tendency to continue an endeavour once an investment of money, effort or time has been made (Organizational Behavior and Human Decision Processes, 1985). In one of their studies, a theatre sold season tickets at randomly assigned discounts; those who received the smallest discount attended the most plays, and attendance fell as the discount rose. In a portfolio, the same instinct appears as the argument that too much has been spent to stop now.
Escalation of commitment
Barry Staw’s 1976 study, in which 240 business school students made a simulated investment decision, found that people committed the greatest amount of resources to a previously chosen course of action when they were personally responsible for its negative consequences (Organizational Behavior and Human Performance, 1976). The practical implication is uncomfortable: the executive who sponsored a project is often the person least well placed to judge whether it should continue. That is not a question of integrity. It is how responsibility bends judgement.
Escalation in technology projects
Mark Keil, Joan Mann and Arun Rai surveyed IS audit and control professionals, and their results suggest that between 30% and 40% of information systems projects exhibit some degree of escalation. Projects that escalated had significantly worse outcomes on perceived implementation, budget and schedule performance. Of the theoretical explanations they tested, a construct called the completion effect gave the best classification of which projects escalated (MIS Quarterly, 2000). Our reading is that the nearer a project appears to its finish line, the harder it becomes to question.
Preferential attachment of funds
Bent Flyvbjerg adds a portfolio-level mechanism. The investments that look best on paper tend to be funded, and these are often the ones with the largest cost underestimates and therefore the greatest need for additional money during delivery. After approval comes lock-in and a point of no return, followed by escalation of commitment as more funds flow to the original investment to close the gap (Project Management Journal, 2021).
The allocation habit
A McKinsey study of more than 1,600 US companies between 1990 and 2005 found that, for one-third of the businesses in its sample, the capital received in a given year was almost exactly what they had received the year before, with a mean correlation of 0.99. Companies in the top third for reallocation, which shifted an average of 56% of capital across business units over the 15-year period, earned on average 30% higher annual total returns to shareholders than those in the bottom third. The authors identified anchoring on last year’s budget, loss aversion and corporate politics among the causes (McKinsey Quarterly, 2012).
Two cautions apply. The study concerns business units rather than project portfolios, though the mechanisms are the ones portfolio leaders face. And over periods of less than three years, the heavier reallocators delivered lower shareholder returns than more stable peers: reallocation is a medium-term discipline in service of a clear strategy, not a licence for churn.
What happens when nothing is stopped on purpose
If a portfolio will not stop anything deliberately, something is still stopped: whatever is easiest to cut. Drawing on its earlier work on England’s road investment programme, the NAO’s 2025 lessons on mega-projects noted that increases in cost estimates on large, complex projects can distort the financial position of the overall portfolio, resulting in smaller projects being delayed or cancelled to help manage cost increases in the larger ones. It added that mega-projects can carry sizeable sunk costs, which makes cancelling less palatable (NAO, Lessons learned: Governance and decision-making on mega-projects, 2025).
The other default is the across-the-board cut. The NAO observed that, when spending reductions are needed, there is an incentive to present them as fair by reducing the budgets of all or most departments by the same amount, even when that is not the best use of scarce funding, and that such cuts badged as efficiency savings can produce the opposite: inefficiency, confusion and waste.
Explicit, portfolio-level stop decisions look different. In 2023 the UK government decided to cancel Phase 2 of HS2, citing the increasing costs of Phase 1, repeated schedule delays and changes in travel patterns since the pandemic. As the NAO recorded, the government stated that HS2 accounted for over one-third of all government transport investment at the time and prevented spending on other transport priorities. Whatever view one takes of that decision, its stated rationale was an opportunity cost across a portfolio, not only the economics of a single project.
A portfolio that never stops anything has not prioritised. It has queued.
Stop, pause, defer or reshape: using the full menu
Stopping is one option among several, and the standards already supply the vocabulary. PMI’s Standard for Portfolio Management notes that change control may include starting, stopping and delaying portfolio components as resources are reprioritised. The UK’s project delivery standard, GovS 002, anticipates that monitoring the portfolio will identify existing work for amendment, rescheduling or termination. Cooper’s gates end in one of four decisions: Go, Kill, Hold or Recycle.
| Option | When it fits | What it releases | What to guard against |
|---|---|---|---|
| Stop | The case no longer holds at the remaining cost, or the same capacity has better uses | Capital, people and leadership attention | Unmanaged exit costs; losing reusable work and lessons |
| Pause | A dependency, bottleneck or external decision must be resolved first | Scarce people for higher-priority work | Holding costs; loss of team and supplier knowledge; pauses that become permanent by default |
| Defer | The work is sound but is not yet the best use of capacity | Near-term capacity and funding | Deferral as a polite word for never; the cost of delay |
| Reshape | The objective is right but scope, approach or timing is not | Part of the budget and some risk | Repeated resets that postpone a stop decision |
| Merge or replace | Two efforts overlap, or a better option has emerged | Duplicated effort and cost | Blurred accountability after the merger |
Reshaping deserves particular care. The NAO defines a reset as a fundamental or substantial change to what a programme will achieve, or how it is delivered, that cannot be managed through routine change control. It urges decision-makers to compare the value of resetting a programme against stopping it, and notes that stopping, with costs potentially written off, can be the right decision. Its example of the government’s Verify programme is instructive: the programme was subject to more than 20 internal and external reviews, and the NAO concluded that it was difficult to justify sufficiently the successive decisions to continue.
Designing the stop decision before it is needed
These design choices are our synthesis of the evidence above, not a published standard.
- Set stop criteria at approval. Cooper’s gates combine must-meet criteria, designed to weed out misfit projects quickly, with scored should-meet criteria used to prioritise. Every approval should also record the conditions that would trigger a stop-or-continue review: benefits falling below an agreed threshold, a critical dependency failing, delivery confidence remaining low without a credible recovery plan, or the strategic rationale lapsing.
- Look only forward. A continuation decision should compare the cost still to be spent with the value still achievable. Money already spent is irrelevant to that comparison; the cost of stopping, from contract exits to write-offs, is not, and belongs in the calculation.
- Separate the reviewer from the sponsor. Staw’s finding on personal responsibility is the case for review by people who did not make the original decision, with independent assurance for the largest items.
- Make each stage re-earn its funding. The NAO recommended that, for mega-projects, approval processes might fund a project only to its next stage of development, and that those charged with governance should advise stopping a project if early work shows it to be too risky or costly, or its benefits too uncertain. Our brief on connecting capital to project outcomes covers the funding mechanics.
- Give gates teeth and a deadline. Cooper describes gates as go/kill and resource-allocation meetings where senior management decides whether to keep investing or cut its losses, and he argues that a decision should be made on the day of the gate meeting, with the resources that a go decision requires actually committed.
- Make released capacity visible. When a project stops, report what it frees and where that capacity goes. Otherwise the organisation sees only the loss, and the next stop decision becomes harder.
Who holds the right to say stop
In our reading, stopping fails most often for lack of an owner rather than lack of analysis. The decision right should sit with the forum that owns the whole portfolio, typically an executive portfolio board or investment committee, not with the programme that would be stopped. Each role then has a distinct job:
- The sponsor recommends continuing, reshaping or stopping, and is expected to raise the question first.
- The portfolio office supplies comparable evidence across projects: confidence, capacity, cost to complete and benefits still achievable.
- Finance validates the cost to complete and the cost of stopping.
- Independent assurance reviews the largest and riskiest items.
- The executive forum decides, records the reasons and reallocates the released capacity.
Cadence matters as much as structure: quarterly portfolio reviews, plus reviews triggered by pre-agreed stop criteria, keep the question live rather than leaving it to the annual budget round.
Where the stop decision sits across the disciplines
The cost and meaning of stopping differ by discipline. In capital projects, stopping is usually cheapest before the final investment decision; our brief on why capital projects fail before construction begins examines that stage. In technology and transformation, the escalation evidence from information systems projects applies directly, and benefits depend on adoption that may never arrive (see transformation and technology investment). In R&D, stopping is part of the design: Stage-Gate was developed for new-product development, and Cooper argues that tough gates eliminate poorer projects early and produce a better portfolio (see R&D portfolio management). Staged funding makes stopping cheaper, and every stop that releases capacity to a stronger project is a gain in value that never appears on the stopped project’s record.
Decide, fund, deliver, prove, and decide again
The project economy runs from Decide through Fund and Deliver to Prove. The stop decision belongs to every stage: before funding, when the case is weak; during delivery, when confidence falls; and after early results, when the benefits are not appearing. Project Economy Forum, which convenes the leaders who decide, fund and deliver projects, treats portfolio leadership as one of its eight areas, and its approach to project portfolio management focuses on those decisions themselves.
Stopping as the discipline behind honest business cases
The most important effect of a portfolio that stops projects may be on the projects it starts. If cost underestimates are rewarded with funding and never with a stop decision, as the preferential-attachment argument suggests, optimistic business cases are a rational strategy for sponsors. When stopping becomes a routine and respected outcome of review, the incentive changes: a case that overstates value or understates cost is more likely to be tested and ended. The portfolios that stop well are, in our reading, the ones most likely to be offered honest proposals in the first place.
Frequently asked questions
What is project portfolio prioritisation?
It is the ongoing ranking of an organisation’s projects and programmes against each other by value, urgency, risk and dependency, within the limits of available capital and capacity. It is only complete when it also decides which work to stop, pause or defer.
Why do organisations find it so hard to stop projects?
Incentives reward starting new work more than ending weak work, and research on the sunk-cost effect and escalation of commitment shows that people keep investing in failing courses of action, especially when they were responsible for the original decision.
What is the difference between stopping, pausing and deferring a project?
Stopping ends the work and releases its resources. Pausing suspends it until a dependency or bottleneck is resolved. Deferring delays a sound project because other work is a better use of current capacity. Pauses and deferrals need restart conditions, or they become stops by default.
Who should have the authority to stop a project?
The forum that owns the whole portfolio, typically an executive portfolio board or investment committee, with the sponsor recommending, the portfolio office and finance supplying evidence, and independent assurance for the largest projects.
Sources
- National Audit Office, Lessons learned: a planning and spending framework that enables long-term value for money (2024). NAO
- National Audit Office, Lessons learned: Resetting major programmes (2023). NAO
- National Audit Office, Lessons learned: Governance and decision-making on mega-projects (2025). NAO
- National Infrastructure and Service Transformation Authority, NISTA Major Projects Annual Report 2025–26 (2026). GOV.UK
- Project Management Institute, Pulse of the Profession 2026: Driving Success in Complex Projects (2026). PMI
- Project Management Institute, The Standard for Portfolio Management, third edition (2013). PMI
- Cabinet Office and Government Project Delivery, Government Functional Standard GovS 002: Project delivery (2025). Government Project Delivery
- R. G. Cooper, The Stage-Gate Idea-to-Launch Process: Update, What’s New and NexGen Systems, Journal of Product Innovation Management 25(3) (2008). Robert G. Cooper
- H. R. Arkes and C. Blumer, The psychology of sunk cost, Organizational Behavior and Human Decision Processes 35(1) (1985). RePEc record
- P. N. Herrmann, D. O. Kundisch and M. S. Rahman, Beating Irrationality: Does Delegating to IT Alleviate the Sunk Cost Effect?, working paper describing the Arkes and Blumer studies (2012). arXiv
- B. M. Staw, Knee-deep in the big muddy: A study of escalating commitment to a chosen course of action, Organizational Behavior and Human Performance 16 (1976). ScienceDirect
- M. Keil, J. Mann and A. Rai, Why Software Projects Escalate: An Empirical Analysis and Test of Four Theoretical Models, MIS Quarterly 24(4) (2000). AIS eLibrary
- B. Flyvbjerg, Top Ten Behavioral Biases in Project Management: An Overview, Project Management Journal 52(6) (2021). arXiv
- S. Hall, D. Lovallo and R. Musters, How to put your money where your strategy is, McKinsey Quarterly (2012). McKinsey