What makes large projects go wrong (and some suggestions on what to do about it)
*(and some suggestions on what to do about it)
The best laid schemes o’ mice an’ men
Gang aft agley,
An la ’e us nought but grief an’ pain
For promis’d joy!
What is it about large projects’ that makes them go wrong? Cost estimators underestimate cost, planners underestimate time, technologists underestimate technical complexity, and owners and project managers face tremendous difficulties in reorienting, replanning, and redirecting projects. Projects which, on paper, show great promise in the conceptual stages falter, flounder, and cause disappointments and frustrations in their middle and closing phases. Ultimately, many large projects (LP’s) have resulted in partial or complete failure.
Clearly, as LP’s evolve out of their embryonic stages, they come to have a mind of their own. Typically, LP’s originate in a single or a few brilliant men with sharp visions of their scope and mission, but the multiple, simultaneous, and diverse activities set in motion in the later, middle stages reveal problems which no one had foreseen or would have wanted to admit as possibilities. In view of their importance, LP’s deserve a great deal more critical attention than they have been given in the past.
Before the Decision to “Go Ahead”
LP’s demand significant use of land resources or human resources or both. They affect the lifeblood of regions and of nations. Certainly, the justification for initiating LP’s must be overwhelmingly favorable, as society is the benefactor. Many studies, disciplines, and approaches are available to decision makers for evaluating the costs and benefits of LP’s, and their long-term net benefits. In hydroelectric power projects, for instance, the ‘estimated cost per kilowatt hour’ becomes the major criteria of acceptance. The construction of automated mechanical sortation plant is justified on the basis of cost per piece sorted.
But the rationale for LP’s goes further. One aspect that often appears to be overlooked is the social ethic of projects. LP’s inevitably alter the lives of men, or nature (the environment), or both. The new changed state of man and environment after the project completion must be worth the price. Professional groups, industrialists, politicians, and special interest groups have a responsibility to consider fully the ethics of the projects that they promote. To continue with the earlier examples cited, the changing topology caused by major hydroelectric projects may have a broad impact on human and animal habitats. Automation in large industrial plants affects the economy and the distribution of wealth, and most of all the workers. In both these instances, the social and environmental impact must be assessed.
A special consideration within the framework of a social ethic is the private ethic of the promoters of LP’s, be they politicians, technologists, or businessmen. Human nature being what it is, once the idea of a large project is kindled in the hearts of men, it seems all too frequent that subsequent analysis point to predictable conclusions which smack of self-justification and which somehow disregard the overall good. But on this shrinking planet earth, man must increasingly interact and cooperate with man and nature, and there can be no room for petty, sectarian, or narrow-minded evaluations. Those who are entrusted with making the basic, strategic decisions must display statesmanship and wisdom in their deliberations, remembering that man the initiator is inexorably linked to man the inheritor.
Beyond statesmanship, there is the possibility of reinforcing wise decisions at the outset, by the creation of an advisory committee which would be composed of recognized leaders in appropriate fields of specialization and whose express purpose would be to counsel the decision maker. Such a committee would be a safeguard in the decision-making process.
After the Decision to “Go Ahead”
Once the financing of a LP has been approved, the nature of the problems shifts from basic considerations of feasibility, to that of how to get the job done.
A basic model for planning and controlling projects is shown in figure 1 below. Figure 1 illustrates a typical two-phased project in which the first phase consists of conceptual development and the second, the execution phase. For the purpose of this study, the conceptual phase or program development phase is defined as including the detailed system concept, supported by detailed cost analysis, and the elaboration of a project work breakdown structure. The execution phase involves detailed design, specification, contracting for works, and acceptance by the client at project completion. The second “execution” phase may be simply considered as an agglomeration of ongoing managerial, technical and contractual activities, modified and corrected by an information-gathering or measurement and feedback process. This second phase is likened to the classical control or feedback theory.
Figure 1. Project planning and control model.
An analysis of this simple model and a comparison with actual events on LP’s lead to some clues as to what makes LP’s go wrong after the “go ahead.” Firstly, the deliberate phasing of projects into distinct conceptual and execution phases merits critical attention. It is assumed that the conceptual phase provides clear guidelines to designers and builders. For instance, let us take the hypothetical example of an airport conceptual development which might freeze the concept of getting passengers to airplanes by selecting a shuttle vehicle to travel between the terminal and the aircraft. During an initial study, this might prove to be the most time and cost economic method of getting passengers to their aircraft. Alternative cost analyses would probably bear this out on paper. Now, let us suppose that the concept is frozen and that architectural and engineering design of the terminal, runways, and other services proceeds based on this concept (in the execution phase). It is possible to imagine that the agreed-to shuttle system would prove to be much costlier than originally foreseen to be, and that an alternative modified concept would now prove more cost-effective. It is precisely for this reason that a project should not be divided categorically into distinct conceptual/execution phases. LP concepts should be subject to change until the last possible moment.
As concepts deviate from plan, then they must be adjusted and updated. A first guideline for LP’s following the “go ahead” is therefore: don’t phase projects into distinct cut-offs between conceptual and execution phases. Conceptual program development is a continuing process affected by better, more complete information. The baseline of a project is dynamic, not static. The unwillingness to adjust programs may well lead to failure.
Further insights as to why LP’s go wrong can be gained by focusing on project activities, measurement, the so-called “execution” phase (figure 1). LP’s differ from small and medium-sized projects in that their activities are more concentrated and intense, involving greater technical input, financial outlays, and longer time periods for completion. Obviously, the more novel a project is technically, the more prone it is to cost overruns. And the more numerous are simultaneous activities set in motion, the more difficult it is to manage. Surely, the key to reducing risks during execution is to minimize the complexity and intensity of execution.
A second conclusion drawn from consideration of a project model: Package the work of large projects into smaller self-contained projects. In other words, it is wise to establish mini-projects, with their own goals and resources. If possible, the projects should be isolated physically and functionally. Efforts should be expended to limit managerial, financial, scheduling, and technical risks. Typically, projects in the $40 to $50 million range have been managed quite successfully. With smaller projects, it should be possible to delegate appropriate responsibility and authority to the mini-project managers. The LP project manager then becomes the team builder and coordinator among these mini-managers. The power to reward and sanction constitutes the lever with which the project manager guides and directs the mini-project managers.
A final point which merits attention when evaluating the failure of LP’s, relates to the quality of decision-making. Our model (figure 1) assumes that appropriate time and cost “measurements” are instituted by the project manager to help arrive at good decisions. But what if the resource mix is basically wrong? Or the marketplace does not have the appropriate expertise to respond to project needs? Or if the technocrats cannot exercise appropriate technical judgements?
In such instances, planning tools such as the Critical Path Method (CPM) will show that there is a problem but won’t indicate the way to solve it. Financial reports that report planned versus actual expenditures will also indicate a delay in spending, but there will be no obvious conclusions on what is the appropriate remedial action. It is evident that good decisions are enhanced by thorough and systematic evaluations of all conditions.
The complexity of the decision-making process on LP’s suggests a third approach: All aspects of planning and execution of large projects must be evaluated on a continuing basis. Management by time and cost-related exception reports is not enough. LP’s are novel and unpredictable. They demand constant analysis and investigation. Small changes which might be overcome fairly easily on small or medium-sized projects can have disastrous multiplying effects on LP’s. The analysis of project performance is crucial, and the flexibility to adapt and change, equally so. Rigidity, sloppiness, and intransigence are undesirable properties which may lead to disaster.
The evaluation (or audit) process, a formal organizational endeavour, may be usefully categorized in the following areas of concern:
- Program: relates to objectives, goals, the analysis of the “overall” system to be produced; costs and benefits, quantitative and key variable analysis, projections.
- Information systems: including financial and scheduling planning and control data; risk analysis; market intelligence; industrial intelligence.
- The technical system: quality standards, industrial and consultant technical capability, technical risks, critical design elements, configuration control, documentation needs, technical evaluations.
- The management system: the combination of policies, procedures, the management decision process, management style and communications, contractual arrangements/enforcement, manpower planning, contingency planning, performance evaluations, team building, and management reports.
Evaluations of these spheres of concern should answer the question: Are the critical issues adequately and economically dealt with? LP’s must be equipped manager-ially to provide the basis for good decision making. Key managers in LP’s should come to accept this need for continuing probing analysis. LP’s should have access to outside consultants and skills in the event that a particular evaluation skill is found lacking. For instance, in a large automation project, it is essential to have access to an automation expert, with relevant experience.
A modified project planning and control model for large projects is illustrated in figure 2 below. This revised model serves to illustrate the three recommendations made regarding the possible causes of large project failures, once the “go-ahead” has been granted:
Figure 2. Modified project planning and control model for large projects.
- A distinct initial program phase: This is overcome by not separating into conceptual and execution phases, and allowing the concept to vary, as necessary.
- Complexity: Overcome by breaking down projects into mini-projects with mini-managers reporting to a project manager.
- Faulty decision-making: Overcome by accentuating the systematic evaluation over all aspects of LP’s, and equipping LP’s with the staffing to ensure proper evaluation.
Furthermore, it is to be hoped that future strategic decisions to undertake LP’s will be done with full consideration of all aspects, including the ethics of projects being proposed, so that man as a whole will benefit.
- A large project is defined approximately as any project valued at over SI00 million (1978). This ballpark figure is based on the author’s experience and association with projects of all sizes.