Major projects frequently exceed their budgets, miss deadlines or fall short of the benefits used to justify them. This happens despite the involvement of intelligent, experienced executives and project leaders who know the principles of good delivery.
Why do capable people continue to make decisions that violate principles they already know?
Decades of research by Oxford professor Bent Flyvbjerg, brought together with Dan Gardner in How Big Things Get Done, help explain the pattern. Our brains make us overly optimistic, underestimate uncertainty, disregard relevant experience and become attached to decisions already made. Flyvbjerg also shows that project decisions are shaped by power, the competition for resources and the incentives to make a proposal look attractive.
Based on our work with large organisations and research into organisational behaviour, we want to take that last point further. A project manager who puts forward an unrealistic timetable may know perfectly well that it is optimistic, yet also know that a realistic one will never secure approval. An executive may continue funding a troubled initiative because stopping would damage their credibility. To understand the decision, we need to examine the conditions that make it reasonable for the person taking it.
There is a leadership dimension too. Large initiatives involve legitimate competing demands: ambition and realism, commitment and flexibility, speed and thoroughness, alignment and constructive disagreement. Overplaying one can eventually undermine the very value it was meant to protect.
We have organised these problems into eight recurring traps, following the decisions executives make from the first proposal to the final review. The eight traps are our synthesis, informed by Flyvbjerg's findings and practical guidance. They are not a taxonomy published by Flyvbjerg. They also provide the structure of our Project Blind Spots Check.
What Flyvbjerg says about bias
In his 2021 paper Top Ten Behavioral Biases in Project Management, Flyvbjerg identifies ten biases that can undermine project planning and delivery. Nine are principally cognitive tendencies. The tenth, strategic misrepresentation, concerns the deliberate distortion of information to serve an interest, such as obtaining approval or funding. He argues that political pressures can compound cognitive biases, especially where the stakes and hierarchy are substantial.
That distinction matters. Someone may genuinely believe an ambitious forecast. Someone else may know it is unrealistic and still promote it. The resulting business case can look much the same. Sensible governance must be designed for both possibilities.
Flyvbjerg identifies the following ten: strategic misrepresentation, optimism, uniqueness, planning fallacy, overconfidence, hindsight, availability, base-rate fallacy, anchoring and escalation of commitment. His paper defines each and explains how they affect projects.
The eight traps and their behavioural roots
The table connects each trap in our assessment to the closest biases in Flyvbjerg's ten. These are interpretative links: several biases can act together, while some traps also concern project design or organisational practice. The four sections below explain the biases in context.
| Stage and trap | What happens | Closest biases from Flyvbjerg |
|---|---|---|
| 1. The solution trap | We become invested in a preferred answer before testing whether it is the best route to the outcome we need. | Anchoring; overconfidence |
| 2. The uniqueness trap | We treat our initiative as exceptional and discount what actually happened in comparable projects. | Uniqueness; base-rate fallacy; availability |
| 3. The false certainty trap | Precise budgets and dates hide weak assumptions, dependencies and exposure to severe downside risks. | Optimism; planning fallacy; overconfidence; anchoring |
| 4. The premature commitment trap | Public promises and major investments close options before the most consequential unknowns have been tested. | Anchoring; planning fallacy; later escalation |
| 5. The expertise illusion | We mistake seniority or technical credentials for proven delivery experience and overlook gaps in the wider team. | Overconfidence; availability |
| 6. The alignment illusion | Formal agreement can conceal unresolved objections and conflicting incentives. When pressure rises, colleagues discover they cannot rely on one another to deliver. | Strategic misrepresentation; sometimes overconfidence |
| 7. The big-bang trap | We bundle too much into one launch, delaying feedback and allowing complexity and interdependence to grow. | Planning fallacy; overconfidence; also a project-design problem |
| 8. The escalation trap | Past spending, public commitments and personal credibility make it harder to change course when the evidence turns. | Escalation of commitment; hindsight can distort later reviews |
1. Define the case: the solution and uniqueness traps
The first mistake may happen before anyone has written a project plan. A favoured system, building, technology or reorganisation starts to define the problem it was meant to solve. This is anchoring: the first idea receives more weight than it deserves when alternatives are considered. Overconfidence can reinforce it when sponsors trust their own judgement more than the evidence warrants. Flyvbjerg's instruction to start by asking why is particularly useful here. If the preferred solution disappeared tomorrow, could the sponsor still state the outcome and compare credible alternatives?
Even with a clear outcome, the team may assume this project is different. Uniqueness bias makes us see our initiative as more exceptional than it is. The base-rate fallacy shifts attention from the actual distribution of cost, time and benefit outcomes in comparable projects to the appealing story of this particular case. Availability bias compounds the problem: a vivid success story is easier to recall than a representative sample of outcomes. An outside view asks what happened to comparable projects, including the unsuccessful ones, and what that evidence changes about the present plan.
2. Commit resources: false certainty and premature commitment
Budgets and schedules become increasingly precise as a proposal approaches approval, often while its underlying assumptions remain poorly tested. Optimism bias makes favourable outcomes seem more likely and setbacks less likely than the evidence suggests. The planning fallacy shows up in the estimates themselves: costs, delivery times and risks are understated, while expected benefits are overstated. Overconfidence reinforces these tendencies. Anchoring adds another difficulty: once a number has appeared in a board paper or public statement, later estimates tend to orbit around it.
There is a further issue that averages conceal. Some projects face extreme downside risks. Testing the assumptions that could invalidate the business case, examining severe scenarios and changing the design to contain exposure matter more than adding another entry to a risk register. Flyvbjerg's watch your downside principle is particularly relevant.
The pressure to show momentum can then lock the organisation into contracts, public promises or a full programme before the hard unknowns are resolved. Thoughtful preparation, staged commitments and evidence-based decision gates allow a team to act quickly once the route is sound. A formal gate achieves little when there is no plausible way to say stop.
3. Organise delivery: expertise, alignment and the big-bang trap
An accomplished executive may have led a large business without ever having delivered this particular kind of programme. The same applies to a technically impressive team that has never handled the relevant integrations, dependencies or political constraints. A few memorable successes can make someone appear more experienced than their full track record justifies, another expression of availability bias. The question is whether demonstrated experience matches the risks of the work, and whether specialists can materially change decisions when they have contrary evidence.
Next comes apparent agreement. A steering committee may have signed off while important stakeholders remain unconvinced, functions protect different targets, and uncomfortable facts travel slowly up the hierarchy. That becomes a delivery problem when teams cannot rely on other functions to honour commitments under pressure. Strategic misrepresentation means deliberately distorting or withholding information to secure an advantage, such as approval, funding or political support. Flyvbjerg includes it as a political bias. Establishing it requires evidence of intent; silence or disagreement alone is insufficient. Independent challenge, explicit obligations and a record of delivering on commitments offer stronger evidence of alignment than unanimous meeting minutes.
Finally, large ambitions can create a single tightly coupled launch. Interdependence grows and usable feedback arrives too late. The underlying problem is partly architectural, which is why Flyvbjerg's build with Lego guidance matters. Build and test useful components where possible, with real-world users and dependencies, before exposing the entire organisation.
4. Reconsider the case: the escalation trap
The more money, status and public commitment invested in a decision, the more difficult it becomes to change direction. Escalation of commitment, sometimes called the sunk-cost fallacy, is the tendency to keep investing because of what has already been spent, even when new evidence points the other way. Hindsight bias tends to appear during later reviews: once the outcome is known, people remember it as more predictable than it was. That can undermine honest learning about the original decision.
The useful test is to agree, early on, what evidence would justify reducing, redesigning or stopping the project, who can make that decision and what happens to people who bring the evidence forward. A fresh appraisal should examine the remaining costs and expected benefits. It should also allow for an initiative that still deserves funding, even though its original case has changed.
Our additional lens: the organisational rules behind the behaviour
Flyvbjerg explicitly addresses political pressure and strategic misrepresentation. We build on that analysis by tracing the organisational machinery that makes certain behaviours attractive or risky for individuals.
The starting point is work on the unwritten rules of the game, associated with Peter Scott-Morgan. People learn what gets rewarded, who controls access to resources and promotion, and what actions help or endanger their interests. Formal structures, targets, approval processes and leadership signals create an informal logic of behaviour.
Business requirement and management practices → rewards and decision power → unwritten rule → individually rational behaviour → collective project consequence.
Consider a funding process that consistently rewards confident business cases and punishes delays. The unwritten rule may become: bring an ambitious number and keep doubts for later. A forecast that looks irrational for the company may be entirely understandable for the person seeking approval. Simply instructing that person to be more realistic leaves the source of the pressure intact.
This approach asks executives to look beyond individual intentions. Which approval rules favour a particular answer? Who pays a career price for raising a problem? Does stopping an initiative count as disciplined stewardship or visible failure? What happens when an expert challenges the sponsor? The answers may explain why information is repeatedly softened, delayed or ignored.
There is a second addition, drawn from our work with Fons Trompenaars on dilemma reconciliation. Many practices that undermine projects begin as legitimate strengths taken too far:
- Ambition and realism: ambition gives a project purpose; realism makes the ambition deliverable.
- Commitment and flexibility: commitment mobilises people; flexibility allows evidence to change the route.
- Speed and thoroughness: speed reduces delay and exposure; thorough testing prevents costly surprises.
- Alignment and constructive disagreement: shared direction makes delivery possible; challenge keeps decisions honest.
The practical question is how each value can strengthen its counterpart. For example, fast, small-scale tests can generate better evidence and allow the team to commit sooner with greater confidence. Clear stop criteria can make sustained commitment more credible because sponsors know it has boundaries. This is the logic behind dilemma thinking.
A blind spot check for people who already know the textbook
Experienced leaders have heard advice to make realistic plans, recruit capable people and manage stakeholders. What is harder is noticing the unspoken assumptions that make the advice disappear when a real decision is on the table.
Our Project Blind Spots Check therefore uses 24 evidence-based questions across the eight traps. It asks for observable instances: when did comparison with completed projects change the forecast? When did specialist dissent change an important decision? What would genuinely cause the sponsor to stop or redesign the initiative?
The results are prompts for investigation and discussion. They do not calculate a validated probability of success or diagnose motives. An unanswered question identifies something worth checking, and disagreement between sponsors may be the most useful finding of all.
The eight traps cover recurring decision risks, though a specific programme may also face technical, regulatory, environmental or market risks that require their own scrutiny. Flyvbjerg's broader guidance includes climate mitigation, which deserves explicit treatment where relevant.
Sources and further reading
- Bent Flyvbjerg (2021), Top Ten Behavioral Biases in Project Management: An Overview, Project Management Journal, 52(6).
- Bent Flyvbjerg and Dan Gardner (2023), How Big Things Get Done, especially the concluding heuristics for project leadership.
- Peter Scott-Morgan (1993), Removing Barriers to Change: The Unwritten Rules of the Game, Arthur D. Little Prism.
- Trompenaars Hampden-Turner, Dilemma thinking and reconciliation.
The eight-trap grouping, its mapping to Flyvbjerg's biases and the assessment questions are an independent IntotheNXT interpretation. The Project Blind Spots Check is not endorsed by Flyvbjerg or Gardner.