Designing the Project Organisation

What a Century of Construction Reform Tells Us About Delivering Complex Projects

Executive Opening

The construction industry knows what good practice looks like, and it has known for more than a century. Its problem is implementation.

Since the early 1900s, governments, professional institutions and industry bodies have asked why projects exceed their budgets, miss their programmes and fail to deliver their intended outcomes. Their answers have been consistent. Projects perform poorly when clients lack capability, objectives are unclear, commitments are made before scope is defined, governance is ambiguous, risk sits with organisations that cannot manage it, commercial incentives conflict with project objectives, interfaces fall between contracts and problems come to light too late. The prescribed remedy has also been consistent, from the early British reviews through Latham and Egan to current practice in Europe, North America, Australia, Japan, Hong Kong and Singapore: stronger clients, better front-end definition, clearer governance, realistic planning, integrated teams, appropriate risk allocation, transparent information and closer collaboration.

The same problems still recur. Better contracts, controls, technology and methods all help, but none is sufficient alone. The deeper problem is that complex projects are rarely designed as organisations.

The physical asset receives intense design attention. Structures are analysed, systems modelled, interfaces coordinated, components tested and the finished asset commissioned. The temporary organisation that creates the asset receives little. It emerges from procurement decisions: a client team forms, consultants and designers are appointed, project managers are added, contractors and subcontractors arrive later, committees accumulate, and contracts divide responsibility and risk. Each organisation may be highly competent in its own field. But competence at organisational level does not produce competence at project level, and the result can be a collection of excellent organisations inside a poorly designed system. This explains a familiar paradox: capable people and firms can together produce a poor outcome without any one of them being incompetent. The fault lies in the architecture that connects them.

A complex project is therefore a temporary enterprise. It brings together hundreds or thousands of people from independent organisations, with different objectives, cultures, commercial interests and contractual duties, and asks them to make thousands of interdependent decisions over several years. The research literature recognises this. Lundin and Söderholm (1995) described projects as temporary organisations with their own logic. Morris (1994) argued that the management of projects must include front-end definition, not only execution. Winch (2014) distinguished the temporary project from the permanent owner and supplier organisations that form it. What is still missing is a practical method for designing the temporary organisation with the discipline applied to the asset. This essay proposes one.

Its central proposition is that complex projects must be designed twice: first as an organisation, and then as an asset.

Five principles follow from the reform record. The client’s capability is the foundation of project performance. Uncertainty must be resolved before commitment, not transferred by contract. Governance must give real authority and let uncomfortable information reach decision-makers quickly. Complexity must be managed at the interfaces between packages. Collaboration must be built into the commercial architecture, because it will not survive a commercial system that rewards the opposite. Technology is an enabler of these principles, not a sixth.

The essay first shows how consistent the diagnosis has been and explains why it has not been acted on. It then examines the five principles and ends by describing what the design of a project organisation contains and how that design is tested.

A Century of Consistent Diagnosis

The diagnosis is older than the formal reviews of contracts. In 1903 the Royal Institute of British Architects (RIBA) agreed a standard form of building contract with the builders’ organisations, and revised it in 1909. The builders withdrew in 1919, and the parties did not agree a common form again until they formed the Joint Contracts Tribunal in 1931. The allocation of risk and payment between client, designer and builder was in dispute a generation before any government reviewed contracts.

In 1918 the Tudor Walters Committee, which set standards for post-war housing, recommended better site organisation, more accurate cost accounting and regular employment for building workers. At the same time the industry’s Joint Industrial Council, known as the Building Trades Parliament, proposed to reorganise building as a public service instead of a casual, adversarial trade. The housing programme that followed showed the cost of commitment before capability. The Housing, Town Planning, &c. Act 1919 placed a duty on local authorities to build, against a political target of 500,000 homes. About 213,000 were completed before the government curtailed the programme in July 1921, amid complaints in Parliament of excessive contractor profits.

Other countries reached similar conclusions at the same time. In New York, the Lockwood Committee (1919–1922) exposed collusion between contractors and building-trades unions that had prevented independent bidding; its work led to hundreds of indictments. In 1921 the Federated American Engineering Societies, led by Herbert Hoover, published Waste in Industry, which placed much of the responsibility for waste in building and five other industries on management rather than labour. In Germany, the Reich procurement committee published the Verdingungsordnung für Bauleistungen (VOB) in 1926 to give construction the balanced procurement rules and contract conditions that the Civil Code lacked. A revised VOB is still in use. These were different instruments: a housing standard, an industrial-relations body, a corruption inquiry, an efficiency study and a procurement code. Each identified problems that later reviews would find again.

In Britain the formal reviews began with the Simon Report, The Placing and Management of Building Contracts (1944), which raised concerns about coordination between the professions and construction. The Emmerson Report, Survey of Problems before the Construction Industries (1962), and the Banwell Report, The Placing and Management of Contracts for Building and Civil Engineering Work (1964), again found fragmentation, poor coordination and weak procurement arrangements. Sir Michael Latham’s Constructing the Team (1994) described a fragmented, adversarial industry and called for cooperation, better contracts and dispute avoidance. Sir John Egan’s Rethinking Construction (1998), drawing partly on manufacturing, argued for client leadership, integrated teams, performance measurement and a focus on value instead of lowest initial cost.

Latham endorsed the New Engineering Contract (NEC), first published in 1993, as the model collaborative contract, and later NEC editions developed under the influence of both reports. The two reports also shaped partnering, framework agreements, Constructing Excellence, the Government Construction Strategy (2011), the BIM Level 2 mandate (2016) and the Construction Playbook (2020). The vocabulary changed; the diagnosis did not.

The pattern is international. In Europe, the Eurocodes gave structural and civil design a common technical basis, and procurement reform let public clients weigh quality and whole-life value with price. In the United States, the Critical Path Method and PERT changed programme management; the Construction Industry Institute built a body of research on front-end planning and benchmarking; Lean Construction and the Last Planner System moved attention to the reliability of production; and design-build and Integrated Project Delivery reduced the separation between design and construction. Canada standardised contracts and addressed payment and rapid dispute resolution. Australia institutionalised business cases, gateway assurance and independent infrastructure assessment. Japan pursued productivity through industrialisation and its i-Construction programme. Singapore combined regulation, digitalisation, Design for Manufacturing and Assembly, workforce development and integrated planning. In Hong Kong, Construct for Excellence (2001), the report of the Construction Industry Review Committee chaired by Henry Tang, called for action across procurement, quality, safety, capability, technology and industry structure.

These jurisdictions differ in law, culture, market structure and the role of the state. Their reform programmes still converge on the same short list: better clients, definition, integration, risk allocation, information and collaboration. The industry does not need to discover good practice. It needs project environments in which good practices operate together.

Why the Diagnosis Has Not Been Acted On

If the diagnosis has been so consistent, the failure to act on it needs an explanation. Five causes recur.

The first is the incentive to secure approval. Flyvbjerg, Holm and Buhl (2002) found that cost underestimation in transport projects had not decreased in 70 years and that technical error could not explain it. They concluded that optimism bias and strategic misrepresentation, the deliberate understatement of cost to win approval, are systematic. Of more than 16,000 projects in Flyvbjerg’s database, 8.5% were delivered on budget and on time, and 0.5% on budget, on time and with the expected benefits (Flyvbjerg and Gardner, 2023). A project approved before it is defined will be defined to fit its approval.

The second is the public announcement. Leaders often announce dates and budgets before programme logic exists. After that, the date is no longer an estimate the project can test but a commitment it must defend. The 1919 housing programme is an early example.

The third is the one-off client. Many clients build a major project once, or once in a generation. The team is assembled for the project and dispersed at completion, and its learning leaves with it. Reformers ask clients for a capability that the structure of client organisations makes hard to keep.

The fourth is the culture of procurement and audit, which favours decisions that are procedurally defensible. The lowest compliant price is the easiest to defend, and at award a price that is low because risk is underpriced looks the same as a price that is low because the supplier is efficient.

The fifth is the way reform has been delivered: as separate initiatives owned by different bodies, such as a contract form, a BIM mandate, a gateway process or a payment act. Each changes one part of the system. The system absorbs the change and continues to produce its previous outcomes.

None of these causes is technical. They are organisational and economic, and only deliberate organisational design will remove them.

The Capable Client Is the Foundation

The client sets the environment in which every other organisation works. It defines the objectives and what counts as success. It chooses the governance, procurement strategy, contract relationships, risk appetite, funding and decision processes. It decides when the project goes to market and how mature the design must be at that point. It determines whether the other organisations are encouraged to cooperate or pushed to protect themselves. No other party has as much influence on the outcome.

This does not mean the client must hold every technical skill in-house. It means the client must remain an intelligent client: one that understands the problem it is solving, owns the objectives, challenges professional advice, understands the consequences of major decisions and stays accountable for the whole project system.

A weak client cannot correct this by appointing more consultants. That creates a further failure. The project accumulates project managers, programme managers, designers, cost consultants, schedulers and specialist advisers until responsibility is spread across a network in which everyone owns a part and nobody owns the whole.

Client capability is not another project-management discipline. It is the first project control.

Define Before Committing

Much of a project’s performance is fixed before construction starts, when objectives, requirements, scope, design criteria, interfaces, constraints, procurement and programme strategy, and risk allocation are set. This is also when the pressure to move fast is greatest. Design starts before requirements are stable. Procurement starts before design is mature. Completion dates are announced before the programme logic is tested, and budgets become commitments before uncertainty is measured.

The project then tries to buy certainty through the contract: a fixed price, a contractual completion date, a transfer of risk. A contract decides who carries uncertainty; it does not remove it. If scope is immature, someone pays to develop it. If interfaces are unresolved, someone must resolve them. If the programme is physically impossible, delay damages will not make it possible. The cost appears later as contingency, qualifications, variations, claims, disputes, extensions of time, lower quality or contractor distress.

Do not confuse transferring uncertainty with resolving it.

The evidence for investment in front-end definition is strong. The Construction Industry Institute (CII) measures scope definition with its Project Definition Rating Index (PDRI), on which a lower score indicates better definition. CII data for building projects compares projects by their score before detailed design. Those scoring below 200 out of 1,000 averaged 3% over budget and 5% behind schedule; those above 200 averaged 9% over budget and 21% behind schedule. Merrow (2011) reached the same conclusion from a different dataset: about 65% of the industrial megaprojects he studied failed to meet their business objectives, and weak front-end definition was the strongest predictor of failure. Gateway assurance, infrastructure business-case frameworks and major-project governance models rest on the same logic.

Before large commitments, a project should therefore pass maturity gates that test objectives, requirements, scope, design maturity, interfaces, consents, cost, programme, risk, commercial strategy, organisational readiness and operational requirements. A gate that every project passes is a reporting exercise. If the evidence does not support proceeding, the client must be prepared to stop, redefine or defer.

Governance Must Be Designed

Complex projects often confuse governance with management. Governance sets objectives, authority, decision rights and accountability. Management coordinates the work. Assurance tests independently whether the information given to decision-makers is credible and whether the project can still meet its objectives. Delivery produces the design, construction and systems. When these functions blur, accountability blurs with them.

The usual response to complexity is to add committees. A committee without real authority is only another reporting interface. Good governance defines who can decide what, within which limits and on what evidence, with explicit delegation and clear escalation thresholds. It also needs disciplined change control. Change is inevitable, so the aim is to understand its consequences for scope, cost, programme, risk and operation together, before the decision.

Above all, governance depends on the quality of information moving upwards. A project can have sophisticated dashboards and still be poorly governed if uncertainty is removed from the report at each level of the hierarchy. The condition is simple to state and hard to meet: bad news must travel quickly. That depends on leadership and behaviour as much as on project controls.

Integration Is the Defining Challenge

As complexity grows, the main management challenge moves from the components to the interfaces between them.

London 2012 shows this. The programme had an immovable date and combined venues, infrastructure, utilities, transport interfaces, public realm and many contractors under a strong client, the Olympic Delivery Authority. At bid stage the estimated gross cost was about £4 billion, with £3.4 billion from public funds. In March 2007 the government set a budget of £9.325 billion, which included programme management, contingency, tax and security provisions absent from the bid. The Games did not hold their first budget. Their achievement was disciplined control once a realistic baseline and contingency existed: in December 2012 the National Audit Office expected an underspend of about £377 million against the £9.3 billion public-sector funding package. Programme-level risk management, integrated controls and scope discipline let each project be managed as part of a whole whose purpose was an operational Olympic Park on a fixed date. Someone was responsible for the spaces between the projects.

Crossrail shows the same principle through failure. Its tunnels and stations were substantial civil engineering achievements. The central section was due to open in December 2018 and opened in May 2022, about three and a half years late. In 2021 the National Audit Office forecast the cost at £18.9 billion, excluding trains and depot, 28% above the 2010 budget of £14.8 billion. The components could be built. The difficulty was to make signalling, trains, communications, power, stations, tunnels, operating procedures, infrastructure managers and contractors work as one railway. Crossrail’s own lessons stress systems integration, testing, commissioning and operational readiness throughout the life cycle, not at the end of construction. The same applies to any complex asset. A hospital is not operational because each building-services package is installed, and a data centre is not operational because each item of equipment has passed its factory test.

HS2 shows the problem one level higher, in strategy. The eastern leg to Leeds was cancelled in 2021, and the northern section of Phase 2 in October 2023. In May 2026 the government reset what remains. In 2019 prices, the cost estimate rose from £35–45 billion to £70.9–82.2 billion, and first services between Old Oak Common and Birmingham moved from 2029–2033 to between May 2036 and October 2039. The government attributed about two thirds of the increase to works omitted from the original scope, underestimation and inefficient delivery, and the rest to inflation. When objectives, funding and scope keep changing, the delivery organisation is always solving a different problem, and no quality of project management can compensate. Strategic governance must come before project governance.

Commercial Architecture Determines Behaviour

A project is an economic system as well as a technical one. Each participant has obligations, interests and risks, and these shape behaviour more than any charter does. If organisations gain by withholding information, they will withhold it. If contractors carry risks they cannot control, they will price them, qualify them, pass them down the supply chain or recover them through claims. If designers are paid only for their own deliverables, they will optimise their discipline instead of the asset. Commercial architecture is therefore part of project design.

Crossrail illustrates the effect. The National Audit Office (2019) found that Crossrail Ltd held to the December 2018 date after it had become unrealistic, so that many activities ran in parallel. It identified the compressed schedule, the contractual model, a loss of downward pressure on costs and the absence of an achievable plan as factors in the failure.

Transferring a risk by contract does not change it. Unknown ground stays unknown, an immature design stays immature and an unresolved approval stays unresolved. A risk given to a party that cannot control it has been relocated, not managed. Good commercial strategy allocates each risk to the party best able to understand, influence and manage it. Collaborative contracts, early contractor involvement, Integrated Project Delivery, NEC, prompt-payment legislation and statutory adjudication differ in form but share that aim: a commercial environment that supports production instead of destabilising it. The commercial strategy is part of the project’s operating system, not a legal wrapper around it.

Collaboration Must Be Engineered

Since Latham and Egan, collaboration has become one of the industry’s most used words, and too often an aspiration instead of a condition. Teams attend workshops, sign charters and display values; contracts refer to mutual trust. None of this survives a project system that rewards different behaviour. Real collaboration needs aligned objectives, transparent information, suitable commercial incentives, clear accountability, early involvement of critical specialists and leaders who welcome challenge.

Collaboration is not consensus, and trust is not the absence of accountability. Complex projects need clear leadership and hard decisions. Collaboration lets those decisions draw on the project’s collective knowledge without blurring who is responsible.

Psychological safety matters here. Edmondson (1999) defined it as a shared belief among team members that they can take interpersonal risks without penalty. Somebody on the project must be able to tell a senior manager that an assumption is wrong, question an established programme, expose an emerging risk or admit that their own firm has made an error. If the culture punishes this, the problem does not go away. It appears later, when it costs more. People must be safe to raise problems and accountable for solving them.

Technology Supports the System

Digital information management, artificial intelligence, digital twins, modular construction, Design for Manufacturing and Assembly and automated controls can all improve performance, and construction can move further from bespoke site work towards manufacture and assembly. Singapore shows what is possible when technology, industrialisation and regulation are combined in one sector-wide programme. Japan’s i-Construction programme, launched in 2016, pursues the same objective, but its effect on sector productivity is not yet clear.

Technology does not correct a defective organisation. BIM can model an ill-defined project in great detail, a digital twin can reproduce unresolved assumptions, and a dashboard can present an unrealistic baseline more efficiently. Technology adds value when it improves decisions, integration and production, and not otherwise.

What a Designed Project Organisation Contains

The elements above depend on one another. Good governance cannot compensate for immature scope, good BIM for adversarial commercial terms, a collaborative contract for weak client leadership, or sophisticated controls for an unrealistic baseline. Performance is a property of the whole system, so the system must be designed as a whole.

That needs design deliverables, as a structure needs drawings and calculations. The deliverables below describe the organisation in sufficient detail to test it. Most already exist on well-run projects in some form. What is usually missing is their preparation as one coordinated set, by one owner, before main procurement.

Deliverable	Question it answers	Baseline before
Purpose and success statement	Which outcome defines success, and who owns it?	Business case approval
Organisation breakdown structure	Which organisation is responsible for each part of the work breakdown structure?	Procurement strategy approval
Decision-rights and delegated-authority schedule	Who can decide what, within which limits, on what evidence? How fast must escalation occur?	Procurement strategy approval
Design responsibility matrix	Who designs each element, to which level of detail, at which stage? Who reviews it?	Tender issue
Interface register	Who owns each technical, contractual and organisational interface? Which interfaces cross contract boundaries?	Tender issue
Risk and incentive allocation	Which party can best understand, influence and manage each major risk? Do the payment and incentive mechanisms reward that party for managing it?	Tender issue
Information architecture	Which information reaches which decision-maker, through which forum, within what time?	Contract award
Assurance plan	Which gates must the project pass? What are their exit criteria? Who tests the evidence independently?	Business case approval
Integration and operational-readiness plan	How will systems be integrated, tested, commissioned and handed to the operator, planned back from the operational date?	Contract award

“Baseline before” is the latest point at which each deliverable should be fixed. As with drawings, each deliverable is revised under change control after that point.

The client owns this design. It may appoint a programme manager or systems integrator to prepare it. It cannot delegate ownership, for the same reason that it cannot delegate its objectives.

The design must also be tested. A structural design is checked against load cases; an organisational design should be checked against scenarios. Before contract award, the project team should run tabletop exercises that trace plausible events through the proposed organisation: a major client change after design freeze, unforeseen ground conditions, a late statutory approval, a failed integrated systems test, a supplier in financial distress. For each event, four questions apply:

Does the information reach a person with authority to act?
How long does it take to reach them?
Is the consequence visible across cost, programme, risk and operation?
Does the party that must act have a commercial incentive to act?

Where the answer is no, the organisation has a defect. It is cheaper to correct the defect on paper than on site.

The design then remains a live document. It is reviewed at every stage gate together with the asset design, and it is updated as parties join and leave the project. Stage gates should test the maturity of the organisation as well as the maturity of the asset.

Conclusion

The reform record leads to an uncomfortable conclusion: the industry largely knows what to do, and has known for a long time. Reform has failed mainly because it has arrived as separate interventions (a new contract, a BIM mandate, a gateway process, a partnering workshop, a new controls system), each useful and none sufficient.

A complex project is a set of nested systems. Its technical systems sit inside an organisational system, which sits inside a commercial system. Commercial behaviour follows governance and incentives. Governance depends on information. Information depends on people who are willing and able to report reality. That is why competent organisations can together produce poor outcomes, and why the temporary organisation deserves the same design attention as the asset: its governance, decision paths, incentives, interfaces, information flows, means of challenge and escalation, and transition into operation.

Complex projects must be designed twice: first as an organisation, and then as an asset. The quality of the second design depends on the first.

One question remains open. The argument rests on a capable client, and the one-off client described earlier is the type least able to be one. Some clients borrow capability from a programme partner; the risk is that ownership leaves with the capability. How a client that builds once can hold enough capability to own the design of its own project organisation is a question this essay cannot settle.

References

Banwell, H. (1964). The Placing and Management of Contracts for Building and Civil Engineering Work. HMSO.
Central Council for Works and Buildings (1944). The Placing and Management of Building Contracts (Simon Report). HMSO.
Construction Industry Institute (1999). Development of a Project Definition Rating Index (PDRI) for Building Projects. RR155-11.
Construction Industry Review Committee (2001). Construct for Excellence. Hong Kong SAR Government.
Department for Transport (2026). HS2 6-monthly report to Parliament: May 2026.
Edmondson, A. (1999). Psychological safety and learning behavior in work teams. Administrative Science Quarterly, 44(2).
Egan, J. (1998). Rethinking Construction. DETR.
Emmerson, H. (1962). Survey of Problems before the Construction Industries. HMSO.
Federated American Engineering Societies, Committee on Elimination of Waste in Industry (1921). Waste in Industry. McGraw-Hill.
Flyvbjerg, B., Holm, M. S. and Buhl, S. (2002). Underestimating costs in public works projects: error or lie? Journal of the American Planning Association, 68(3).
Flyvbjerg, B. and Gardner, D. (2023). How Big Things Get Done. Currency.
Latham, M. (1994). Constructing the Team. HMSO.
Local Government Board (1918). Report of the Committee appointed … to consider questions of building construction in connection with the provision of dwellings for the working classes (Tudor Walters Report). HMSO.
Lundin, R. A. and Söderholm, A. (1995). A theory of the temporary organization. Scandinavian Journal of Management, 11(4).
Merrow, E. W. (2011). Industrial Megaprojects. Wiley.
Morris, P. W. G. (1994). The Management of Projects. Thomas Telford.
National Audit Office (2007). The Budget for the London 2012 Olympic and Paralympic Games. HC 612.
National Audit Office (2012). The London 2012 Olympic Games and Paralympic Games: post-Games review. HC 794.
National Audit Office (2019). Completing Crossrail.
National Audit Office (2021). Crossrail – a progress update. HC 299.
New York State Legislature, Joint Legislative Committee on Housing (Lockwood Committee) (1919–1922). Reports.
Reichsverdingungsausschuss (1926). Verdingungsordnung für Bauleistungen (VOB).
Winch, G. M. (2014). Three domains of project organising. International Journal of Project Management, 32(5).