Planning and controlling megaprojects

Abstract

Multiple independent research efforts are beginning to show a more consistent approach to developing successful megaprojects in the areas of oil/gas, mining, and construction projects than have been used in the past. These megaprojects are characterized by high value (often defined as greater than $1 billion), comparably high benefits, years-long timelines, and correspondingly high risk. While there have been great advances in both project management methodologies and in the tools the project managers have available (such as CAD/CAM, BIM, and advanced project scheduling and budgeting tools), the complexity of these multi-year programs has advanced even more quickly than the tools have. Construction and engineering projects have become more complex and ambitious faster than our ability to manage them. Oil/gas/infrastructure projects now are much longer in duration and far more complex than even ten years ago, with concomitant increased risks and failures. The International Energy Agency estimates that meeting global energy needs will require investing more than $17 trillion by 2030 (Jensen, 2006).

Introduction

This article will look at the classical project management approaches that focus on delivering the final product within cost and schedule constraints once the project enters the execution phase. We will then examine multiple lines of research that show that the ultimate success of a complex program has very little dependency on how the program is managed once the construction phase begins and far greater dependency on what happens before that phase begins. If a $10 billion dollar refinery runs late and over budget, the failure has started long before the project schedule was created or the engineering/procurement/construction (EPC) process began. All of the serious research in this area shows that the only part of the effort where traditional project management approaches make sense in the later stages, the engineering and EPC stages. Earlier phases take a different approach to ensure success.

We will look at four topics: An overview current project management practices; current research on megaprojects; development stages for efforts on this scale; and some recommendations.

PMI's Standard For Program Management–2nd Edition (2008) defines a program as “a collection of projects managed in a coordinated way to obtain benefits not available by managing them individually.” While this is a reasonable definition, it does not begin to explain the difficult challenges in managing these efforts. While much of the research in this area has concentrated on high-dollar projects in the oil, gas, and mining industries, with some research on infrastructure projects, the approach described here is for such projects at this scale in any industry.

Because of the particular challenges in managing public works infrastructure projects, this approach is particularly beneficial for large projects involving government interfaces, such as major transportation or other infrastructure projects. Examples of infrastructure projects include work the Saline Water Conversion Corporation in Saudi Arabia is doing on desalination plants. The Saline Water Conversion Corporation develops some of the largest desalination plants in the world and distributes fresh water over approximately 5000 kilometers, of piping. They are presently converting from managing multiple individual projects to a PMO-based program management approach and improving their processes prior to the EPC phase. The government of Qatar has planned $146 billion of infrastructure improvements, and this scale is repeated in other parts of the world.

Overview of Current Project Management Practices

The most common way to manage projects is by cropping the project into distinct phases, each phase having its own inputs, activities, deliverables, closing activities, and milestones. At a very high level, the phases might look something like this:

Exhibit 1–Traditional project management approach

The entire project is planned out in detail and the resulting “waterfall” Gantt chart guides the management of the execution phase of the project. This traditional approach to project management puts a premium on gathering a complete set of planning data which is used to accurately forecast future behavior. For a project that is six months long, accurate predictions are difficult enough. For a megaproject that spans years it can be impossible.

As the IT and software industries have slowly developed an appreciation for effective management over the past 10 to 15 years they have evolved a different approach to managing their projects. There are a number of “quick hit” development methods that come under the umbrella of agile methodologies. These approaches don't even try to plan out the entire project at once but only a small subset of it, while keeping the client/customer/sponsor closely in the loop. While these approaches are adequate for short-term projects where the requirements are very likely to change and where the buyer can't easily visualize the final product, they are of limited use on megaprojects where the technical details are well-defined and where the development goes on for years.

The current traditional planning approach focuses on the details of the project itself, defining the activities needed to do the work and predicting how long the project “should” take and how much it “should” cost. For a megaproject there are too many external influences over a long period of time that impact our ability to accurately predict the future to use this approach. The government of Indonesia in early 2014 decided to stop exporting raw materials such as nickel ore or copper, having been convinced that they can make more money by refining the ore locally and exporting the refined ore. This has had a huge impact on ongoing projects as well as on current mining operations. This creates an environment where our traditional approaches to these projects no longer work.

Current Research

There has been an increasing body of evidence from organizations such as Independent Project Analysis, Inc. (IPA), the Construction Industry Institute (CII), research sponsored by PMI, and academic research that the most critical decisions, the ones that are most likely to make a project successful or to fail, are made by the business decision-makers long before the design and construction stages start. With this in mind, this front-end development (FED) approach to stage gating the project should begin long before the engineering design/EPC phases begin. The seeds of project success are sown in the very earliest setup stages of a project, before the engineers ever get involved (Williams, Samset & Sunnevåg, 2009).

One of the first organizations to emphasize a FED approach seriously was Royal Dutch Shell around the year 2000. In 2001 they revised their project management guide to take into account the change in emphasis from a pure execution-oriented approach to a heavier emphasis on the early stages of the project.

Research by Klakegg on Early Warning Signs

Klakegg, Williams, Walker, Andersen, and Magnussen (2010) studied literature on complex projects to attempt to identify early indicators of future problems. They developed a model to use in comparing complex projects across a variety of industries. The model begins with the business development stage where initial decisions are made based on strategic planning goals and identifies several major stages that must be done effectively before the engineering and construction phases begin.

Exhibit 2–Program phases from early warning signs

Before the project begins, and several times during the early stages, go/no-go decision gates are inserted into the process. The decision gates, DG0-DG4, are defined as:

DGO: Decision that the idea is formally recognized in the organization as an acceptable initiative with a person appointed as responsible for following up and spending (allocated) resources in planning the initiative up to the next decision gate.

DG1: Decision that the initiative is acceptable for further investigation and resource consumption. Further investigation will include identifying principal alternative solutions or concepts for decision makers to choose from.

DG2: Choice of concept. The decision implies that the initiative is acceptable for further planning and includes choice of the conceptual solution to be investigated in a pre-project.

DG3: Decision about financing and execution of the project. In this reference model, this gate is associated with the go/no-go decision.

DG4: Decision to accept the project's outputs/deliverables as complete and commence operations.

It is particularly important that the most critical gates, DG0-DG3, occur before the engineering and construction phases. These gates all belong in the realm of business decisions. There is typically no project management, detailed engineering, or contractor involvement. Indeed, the model assumes that once the go/no-go gate at DG3 is successfully passed, there will be nothing requiring additional decision gates. Considering how extremely challenging these projects are, this is a significant shortfall in the model.

Hutchinson and Wabeke

At Shell Oil, Hutchison and Wabeke (2006) came up with a similar model for the phases of complex megaprojects. Their model begins even before Jergeas’ model does and has five phases: Identify and Assess, Select, Define, Execute, and Operate. They plotted the value of each phase as shown here:

Exhibit 3–Value creation as program evolves

The greatest value is provided in the early, business-owned phases of the project. This is the time to identify and assess opportunities, select the “best” opportunity for further development, develop the details of the selected opportunity, and then go through the execution phase, which includes both project planning, detailed engineering, and construction efforts. Phases 1 and 2 belong to the business; phase 3 starts involving the engineers and some project management; phase 4 is all project management and construction; and phase 5 goes back to the business.

No amount of good engineering, management, and construction will provide much value if the project was the wrong one to begin with. Even good project management will not recover the needed value in a poorly selected project.

Jergeas

Jergeas (2008) shows a similar categorization of pre-EPC phases: identify and assess opportunities; select from alternatives; develop the preferred alternative; execute; operate and evaluate. Each of these phases is followed by a decision gate as shown here:

Exhibit 4 - Jergeas program phases

Phase 1 activities (typically one percent of the engineering costs of the project):

• Clearly frame goal and test for strategic fit;
• Preliminary overall plan;
• Preliminary assessment;
• Develop the phase 1 estimate.

Phase 2 activities:

• Generate alternatives and the preliminary development of alternatives;
• Develop expected value;
• Identify preferred alternative;
• Develop the phase 2 estimate.

Phase 3 activities:

• Fully define scope and develop the detailed execution plans;
• Refine estimate;
• Submit funding for approval;
• Phase 3 estimates (+/- 10 % accuracy).

These phase 3 activities may consume up to 25% of the engineering costs. At the end of Decision Gate 3 is when the authorization for expenditures (AFE) is made.

Phase 4 activities:

• Implement execution plan;
• Minimize changes;
• Finalize operating plan;
• Business plan for phase 5.

Phase 5 (operations) activities:

• Operate asset;
• Monitor and evaluate performance.

While this model has been widely quoted, it is slightly inconsistent with other approaches in that a business plan for operations is not developed in Jergeas’ model until after the engineering work begins. In both the CII's Project Definition Rating Index (PDRI) (1996) and in IPA's research results this is much too late to develop a business plan for operations. This document should be developed much earlier in the process.

AngloAmerican

It is often true that for-profit organizations are much better at developing efficient processes than the academics who come along later and study them. That's true in this area also. The international mining company, AngloAmerican, has for many years faced multiple, serious risks on any mining project they engage in as do all other international mining companies. Their approach to the phases includes: Opportunity Identification; Concept Development and Conceptual Studies; Analysis and Pre-Feasibility Studies; Planning and Feasibility Studies; Execution/Implementation; Commissioning/Production; Ramp-Up/Handover; and Closeout.

There again are stage gates at appropriate points as shown here:

Exhibit 5–AngloAmerican program phases

The mining industry faces many of the same challenges as does the oil/gas (both upstream and downstream) industry: high regulatory burden, environmental impacts, strong opposition from environmental NGOs, high profit risks, and so on.

Because of the high costs and high risks in mining, these early phases become highly critical to success of the mine and production facilities. Similar stages can be found in oil exploration and production companies such as Landmark Exploration Inc.

In September 2013 AngloAmerican announced that it was pulling out of a joint venture to develop the Pebble Mine copper and gold project in Alaska (Reuters Business Report). While they walked away from the project after spending $500 million, the predicted risks in the project outweighed the future benefits. Not too many companies have the strength to walk away after spending that much money. This shows a serious commitment to the early phases of these major efforts.

Development Stages for Programs

There is often confusion in the literature about the proper terminology regarding projects. Should projects be divided into phases or stages? For this discussion, we will utilize the term phases for the phases of a specific project and the term stages for the divisions of a program.

The financial performance of these megaprojects is inherently fragile. Due to their complexity and their environment, their response to an input is not a linear relationship to that input since there are interactions among the multiple components. The behavior of these systems is better described by chaos theory than by classical project management. Chaotic projects can suffer huge changes in their behavior with small changes to their inputs. For example, a large greenfield oil development project can fail financially if the local government decides to withhold permits unless more money is paid to certain government officials (a common occurrence in Russia and in some South American countries), or if the feedstock for a refinery is not exactly as expected. Perhaps a better term for such chaotic projects is that promulgated by Rittel and Webber (1973) and by D.J. Hancock (2002) These projects are too often what they refer to as “wicked messes.”

Our normal approach to detailed planning assumes perfect predictability and so builds rigidity into our management approach by writing schedule and cost constraints into the contracts given to the contractors who will do the actual work on the project. Locking in the approach through contracts makes sense from a pure planning standpoint, but it also builds barriers in our ability to respond to future unknowns and to changing circumstances. Freezing the future in such a way ensures construction claims resulting in schedule delays and cost overruns when the future is not exactly how we assumed it would be. As Flyvbjerg (2013) puts it, “The traditional way to think about a complex project is to focus on the project itself and its details, to bring to bear what one knows about it, paying special attention to its unique or unusual features, trying to predict the events that will influence its future.”

With all this independent research by both academics, consultants, and commercial companies we have gained a fair amount of confidence that our normal approach to managing these projects simply doesn't work. There is a better way to approach management. Based on the previous discussion we have created a combined model, with an emphasis on the decision gates as shown here:

Exhibit 6–Combined program stage model

Business Decision Gate 1 (BG1) – Opportunity Ientification

This is the initial decision by the business to begin a new project or not. For the financial resources which can be made available, what projects should the organization invest in?

• For an oil/gas company,which are the most potentially profitable new fields for development (greenfield projects) at an acceptable level of risk? Or is the money better spent refurnishing and upgrading an existing facility (brownfield project)?
• For a mining company, which are the most potentially profitable new mineral resources we can invest in?
• For a pipeline project,where should the pipeline be routed? How big should it be?
• For a public works infrastructure project, what financial resources are available through taxes or bonds? Which possible projects will return the greatest public benefits?

There is a significant amount of pre-project work that needs be done to ensure success later on.

Business Decision Gate 2 (BG2)–Feasibility Analysis

Once a variety of potential opportunities have been identified, there is work done to develop the most beneficial ones. This is an area where IPA's front-end loading (FEL) process works very well and can provide significant information before a decision must be made to continue or not.

This work does not come out of the project's budget but should be funded out of basic research or from overhead. There is no specific schedule to be met here, the effort should take as long as necessary to produce the information needed to make an intelligent decision. For a pipeline project in Canada (Muiño. & Akselrad, 2009), this effort took three months and cost about one percent of the total project budget.

Engineering Decision Gate 1 (EG1)– Detailed Engineering Studies

In preparation for EG1 we need to start getting the technical staff and the project managers involved in the project. Here, we are beginning the design of the facilities (the front-end engineering design or FEED stage) and the planning for managing the execution itself.

This does not mean the business people have backed away from the project, they are still heavily involved in this stage. At some point towards the middle or end of this stage, there must be an AFE to approve the funds that will finance the remainder of the project. (Note: Some organizations arrange the financing much earlier, without knowing the exact project costs but based on internal historical information. This approach is far more inaccurate than arranging funding after project costs have been analyzed.)

Financial arrangements can be a significant effort all by themselves. It can take months to arrange external financing and cost a significant amount of money. It is not unusual for financing to cost five percent or more of the project costs, a large amount of money in a $10 billion project.

Success data for a go decision should be clearly defined. IPA recommends that a full set of process flow diagrams (PFDs) be complete before proceeding onward. For the pipeline project in Canada cited earlier this effort took six months and cost about four percent of the total project budget.

Developing a schedule for this stage can be highly problematic depending on the project. For a mining project or an oil/gas project, it can take many months to do the field work and develop the site's geology report—longer if the work is in an inhospitable area such as northern Canada, Alaska, or Siberia.

If utilizing the CII PDRI assessment tool this stage would be a reasonable place to perform the first survey. The resulting numbers will not be good because much of the work hasn't been done yet, but will improve as the PDRI is repeated in later stages.

Project Management Decision Gate 1(PMG1) –Project Execution

Preparing for this gate is where traditional project management does best. The business people have largely backed away from day-to-day involvement in the project at this point, and the work is controlled by the engineers and project managers with construction contractors involved in the planning efforts.

This decision point is the beginning of the execution phase. This is where the majority (typically 85-90%) of the overall costs will be spent and the greatest amount of time committed. Both the FEL and the PDRI have areas related to this stage. For the pipeline project in Canada cited earlier, this effort took six months and cost the remaining 85% of the project budget. At the end of this stage the plan was ready for operation.

Project Management Decision Gate 2 (PMG2)–Initial Operations

This is a combined decision point of the project management team, the engineers, the contractors, and the operations/maintenance people from the business side. Here the decision is made to start the commissioning process and slowly ramp up production to full scale. For infrastructure projects such as roads and bridges, this is usually not a major area and never receives a no-go decision.

For production plants this is a danger point. If everything did not go as expected in the design/execution stages, the plant will not achieve its designed operational capabilities. A plant that is only capable of 50% production capacity will never return the economic investment and should be shut down before too much money is lost.

One consideration in the decision is how the economic environment has changed since the project started.If the price of the feedstock has risen, the price of the final product has dropped, or the regulatory environment has changed significantly the decision becomes non-trivial. What does the future economic environment look like? This is not a decision for the engineers or the project management team. This becomes a decision on the business side to continue if the future looks profitable or to mothball the plant if it does not. This decision is repeated after the next gate, once the plant is commissioned and ready for operations.

Business Decision Gate (BG 3)

Of all the gates, this one might be the least important. Not trivial, but the gate which has the least impact on whether to go into full operation. If everything earlier has been done with a reasonable degree of success, the only data that can cause a negative vote to go forward will come from the outside environment. If a company has spent six years developing a new oil refinery, then the only thing that can cause them to mothball or to decommission the project at this point is that oil prices have dropped so low that they would lose money by operating the plant. For infrastructure projects, BG3 will always be a positive vote to go forward. Regardless of the actual benefits achieved there is no benefit to not going into operations for roads, bridges, and so on.

At each gate there are specific data that is required and specific criteria that must be met for the decision to go forward. If data is missing or inadequate the decision must be to withhold approval of further work until the data is sufficient. A gatekeeper should be identified whose job description is to track the data needed and ensure it is adequate for the decision makers to make an intelligent decision.

Recommendations

With four changes in the traditional approaches, classical project management works fairly well in the later stages. The changes lie in treating each project as a new business, in risk management, in requirements definition, and in the procurement approach.

Business Approach

To increase the project success rate a different approach must be taken than traditional project management. The first is to treat each new project as you would treat setting up a new business. Because the future is so unpredictable and so subject to outside influences, this requires a highly entrepreneurial approach by all the participants.

Identify the goals of this new business and set up the organization chart as well as the financing to achieve those goals. While there needs to be support by the parent organization (or organizations in the case of a joint venture), the business should have the flexibility to develop its own approach as much as possible to this unique project.

Set up an integrated project team (IPT) that serves as the core management team to oversee each stage of the project. The leadership of the IPT must be with operations and maintenance teams as the owners of the final facility.

While there is a core IPT, the specific membership evolves as the project moves through the different stages. The business should be involved throughout the project from beginning to end. Engineers and project managers should be involved in the team in Stage Gate 2, the feasibility analysis. Once past BG2, the team is grown by adding contractors or consultants to gain input on the engineering and constructability of the facility.

The research clearly shows that to make these projects successful the emphasis must be placed on the pre-engineering stages. Competent decisions made here will significantly improve the success rate so there must be heavy involvement from the business users. A business “owner” should be identified at the very beginning stages and remain with the project into operations. This would be a major cultural change for many organizations because business people tend to move around more and get promotions faster than engineers do. By keeping them on the project through completion, it ensures they learn from their business decisions instead of not seeing the long-term impacts of their own decisions.

However, even with competent business decisions there is no guarantee of success, only an increased probability. Errors in the engineering stage and construction problems can still have an impact.

Increased Risk Management

Project managers have typically spent little time on formal risk analysis. A risk is something that “could” happen in the future, and most project managers are too swamped with planning and running the day to day details of the project to worry about something that might not happen. A more effect risk management approach is to emphasize formal risk management processes with the staff and tools needed to understand those things that can impact the successful completion of the project thoroughly.

The universe of risks should be all-inclusive, with a particular emphasis on external risks such as environmental, regulatory, labor, and so on. These are the risk that are far more likely to be uncontrollable than are the technical risks of the project. Many of these risks are predictable in occurrence, if not in impact, by looking at comparable projects in the same geographic/country area. For instance, mining projects in South Africa consistently deal with labor unrest and strikes. Once these risks are identified, mitigation actions can be taken to deal with potential impacts.

Any approach to risk management should involve identifying business risks as part of the overall effort. The PDRI is a widely used risk identification approach that begins by identifying the business risks very early in the stage gate process. There are currently three versions of the PDRI (construction, infrastructure, and industrial projects). Each has a series of questions designed to drive out the risks on the project. The early questions, and the ones most heavily weighted in terms of risk impacts, are those asked of the business and cover basic business decisions. Later questions involve the details of the engineering, and the last set of questions cover many project management areas. The weightings are based on feedback from real projects and reflect again that the most critical decisions are made before the engineers and the project managers are involved.

Develop the Requirements and the Data

Experienced project managers know that they cannot adequately plan out a project until they understand the requirements that define the final product.

For small- or medium-sized projects, this is often an area where not enough time and effort are allocated. Scope creep is often identified in many research publications as a major cause of projects running over budget and behind schedule. Yet scope creep is nothing more than not having done a thorough job gathering, analyzing, and freezing the requirements before the project is planned out and the execution phase begins. While there are some causes of scope changing outside of the project's control, the primary source is inadequate requirements.

For projects in the construction industry, the requirements flow from the architect who does the top-level design and architecture, to the engineering firms which do the detailed engineering calculations, to the contractor who implements the architect's vision and the engineer's calculation. For engineering projects there is often a FEED performed to produce the top 25% of the design, which then goes to the contractors who develop the more detailed requirements and design. Barring any errors in the process, there is a relatively straightforward flow of requirements from top level to detailed engineering drawings.

But the initial set of requirements comes from the business. These are generally not true requirements as engineers think of requirements; they are often goals to be achieved. These initial requirements are decided on by the business to answer the question, What can we do that will give us the greatest benefit at an acceptable level of risk?

This is where the disconnect occurs that causes significant problems later on in the EPC phase. The business people think in terms of the ultimate goals and pay little attention to the technical requirements. The business just wants the construction to begin so they can start production and obtain revenues from the final operational facility. To the engineers and contractors, goals are insufficient to create a successful project. They need much more than goals to design the end facility. They need detailed technical information before they can even begin design work.

For an oil refinery, that detailed technical information is an accurate chemical analysis of the feedstock that the refinery will process. For a mining facility, that detailed information is the exact composition of the raw ore that will be processed. For a desalination plant that information is the exact salinity of the input. Bad data at this stage will result in a design that does not work as expected, leading to extensive redesign efforts with concomitant impacts to cost and schedule when more accurate data on the feedstock is available.

If the project is being financed externally, the situation is even worse. There is schedule pressure put on the contractors by the financers to complete the project quickly so they can get their return on the investment. Financers are not engineers or project managers. They simply don't realize that putting schedule pressure on the project will do exactly the opposite of what they want, the project is likely to run into even more problems by rushing the work and creating expensive errors and rework. Even if this situation is pointed out to them, they believe that their project is different, and past history has nothing to teach them.

Change the Procurement Approach

In 2009 KPMG (2009) estimated over 50% of all major construction projects used a lump sum contracting approach. This gives the owners the feeling that there is more risk on the contractor (marginally true), and they can pay less attention to the contractor's work progress (not true at all). The only consistency in these contracts is that the contracting price is much higher, there are more claims filed, and there is a strong incentive on the contractor's part to do low-quality work and utilize low-quality materials. As has been said by several authors, “Nobody has ever paid less than the Lump Sum amount, and almost everybody has paid more, often considerably more.”

Any form of fixed price contract (FFP) is a poor choice for megaprojects. Locking in the schedule and budget when the future is highly unpredictable puts significant constraints on our ability to adapt to changes in the environment. This situation is compounded by the fact that there are only a limited number of contractors capable of constructing megaprojects. As Berends (2009) states: ‘Some 75% of the current global LNG capacity has been realized by four contractors, acting alone or as leading contractor in a JV.’ While the situation in most fields is not that restricted, there is a finite supply of contractors capable of doing this complex work, and that constraint needs to be understood during the procurement strategy approach planning.

Instead, the contracting strategy should be phased. During the architectural, FEED, and engineering phases a FFP contract vehicle can be utilized. However, in the actual construction phases a cost-plus contract is more effective. The best contracting strategy will allow for flexibility to respond to external changes and will require more owner involvement in the construction stage.

This mixed approach is similar to that recommended by Merrow (2011). As he states, “Mixed contracting is a strategy that involved reimbursable engineering and procurement, including, in some cases, the procurement of some lump-sum package items, followed by lump-sum contracts of construction or fabrication by constructors or fabricators that are independent of the engineering and procurement firm(s). The construction lump-sum contracts can be a single lump-sum contract to a construction management organization or a series of lump-sum contracts by craft discipline.”

Returning to our five-stage model, we can identify the specific contracting and procurement documents required at each stage. Harris, Formigli, Crager, Eggen, Reed, and Khurana (2004) have laid out a high-level plan for procurement which has been modified slightly below to fit our model. For megaprojects, the C&P process is equally challenging as the actual work itself.

Exhibit 7–Combined program stage model

Stage Gate 1:

The procurement process begins during Stage Gate 1. The preliminary procurement strategy is defined here in accordance with the normal approach for the organization.

Stage Gate 2:

Gate 2:

• Contracting and Procurement (C&P) strategy;
• Long-lead items list;
• Preliminary local industry participation plan (mandatory in countries with government-mandated set-asides for local suppliers and contractors);
• Insurance strategy;
• Materials management strategy.

Stage Gate 3:

• C&P plan;
• Orders placed for long-lead items and equipment;
• Major contracts ready to execute;
• Participation plan for local industry;
• Insurance plan;
• Materials management plan.

Stage Gate 4:

• Monitor contractor performance;
• Close out contracts and purchases orders as appropriate;
• Gather project and vendor data and documentation;
• Local industry participation plan;
• Updated materials management plan;
• Spares management plan.

Stage Gate 5:

• Ensure delivered products satisfy requirements and perform as expected;
• As-Built documentation.

The need to utilize local labor and suppliers is a requirement in almost every country in the world, and is a major source of risk for prime contractors because of the often low quality of the work performed.

When work is done in developing countries, a best practice approach is that used by Tully Oil when working in Africa. They send a team to the area years in advance to provide suitable infrastructure and training for the local population so that when a refinery is built there is strong support for it, and the people have been trained in its operations and support.