Jan-2018

Advances in benchmarking capital project costs

A scope based approach to benchmarking project costs can provide insights to improve competitiveness

SHAWN HANSEN
Asset Performance Networks

Article Summary

Effective execution of capital projects is essential to the financial health of refining, chemical, energy, and midstream companies. Unfortunately, engineering and construction project performance has stagnated over the past decade. Reports of multibillion dollar projects over-running by hundreds of millions of dollars are common and almost unnewsworthy. Some industry owners and operators have responded to disappointing project performance by implementing more elaborate project assurance practices. But in most cases, the changes have not provided the intended benefits. The upshot has been more arthritic and bureaucratic processes, while the industry demands more nimble and competitive strategies. Meanwhile, project benchmarking practices have stagnated with project performance. Faced with increasing challenges and increasingly ineffective tools, owners and operators have found themselves ill equipped to affect tangible change in project performance. To deliver better project outcomes, owners and operators need to gain a better understanding of project risk and competitiveness. To that effect, they need better benchmarking tools.

Project predictability and productivity are falling
Data collected by Asset Performance Networks (AP-Networks) provides an understanding of current project performance and challenges. The AP-Networks Capital Project Database contains information on approximately 2000 capital projects authorised after 2005. More than 900 of these are major projects, with costs ranging from $25 million to more than $3 billion. The major project database comprises information from refining, chemicals, energy, midstream, and power generation projects, with data provided by more than 30 refining and chemical companies. Revamp, debottleneck, brownfield, and greenfield projects are all represented.

This data demonstrates that capital project costs and schedules have remained unpredictable for the past decade. Seventy-two percent of projects fail to achieve their cost, schedule, or safety targets. Approximately 25% of projects can be dubbed ‘train wrecks’, over- running their cost estimates or schedule targets by more than 25%, or suffering significant issues and delays during start-up.

Many factors contribute to project failure. Project characteristics associated with complexity and risk – including size, number of stakeholders, technical complexity, and level of integration with existing assets – continue to increase. One factor worth emphasising is labour productivity. Engineering and construction productivity are key to cost competitiveness and understanding risks. Our analysis shows that over the past decade, engineering and construction productivity has been poor. Figure 1 shows the change in construction labour productivity of refinery projects relative to other industries.

Based on data from the US Bureau of Labor Statistics (BLS), construction labour productivity for the overall US economy in the last decade has improved at more than 1% annually. AP-Networks’ data and analysis show that construction labour productivity in the refining industry has remained flat over this same period. The BLS reports that auto manufacturing productivity has improved nearly 2% per year, leading to cost reductions of about 16% over the decade. Even accountants and lawyers have made more productivity improvements than the projects in the refinery industry.

A recent article concluded that some construction industry practices and structural problems increase costs for buyers, but hamper investment in improving productivity.1 As other industry sectors improve, the world of projects languishes. Indeed, the projects sector lacks effective tools to understand and benchmark engineering and construction productivity.

The limits of current project benchmarking tools
Before funding and executing a major capital project, an owner should seek to answer the following questions:
1. How much risk are we taking on? Is it more than we can handle?
2. Is the project cost estimate too aggressive?
3. Are we spending too much to modernise this plant or add capacity?
4. Is the project cost estimate too conservative?
5. Will this project add to or detract from our competitive position in the industry?

Armed with cost benchmarks to compare their own plans to industry norms, firms look to answer these vital questions. Engineering and construction activities make up between 70% and 85% of a typical project’s costs, with the remainder being equipment and materials (pipe, steel, concrete and so on). At AP-Networks, our conversations with industry leaders have revealed shortcomings with current cost benchmarking methods; benchmarking efforts do not provide the actionable insights that are necessary to affect change.

The traditional approach to benchmarking capital cost is derived from a method known as Lang Factor Analysis. Developed as an estimating tool seven decades ago, the Lang Factor is the ratio of the total project cost to total equipment cost.2 Lang Factors can be found for various types of plants (see Table 1). Project teams and estimators make adjustments for escalation, location and size, and then compare industry Lang Factors to their project estimates and actual costs in order to benchmark performance. The results are interpreted as follows: higher Lang Factors represent a conservative estimate or an overly expensive actual cost, while lower Lang Factors represent a more competitive (and riskier) estimate or competitive actual costs.

Lang Factor analysis is based on the assumption that equipment costs fully describe a project’s scope and can be used to factor up to total capital costs. Statistics show that, in general, as more equipment is installed, total costs increase. However, as Figure 2 shows, there is a large degree of variability in the equipment to total cost relationship. Indeed, that variability is more than 100%.

This traditional approach to project cost benchmarking often provides results that are difficult to decipher. Is my project’s cost high because of inefficient engineering, or is the amount of engineering needed appropriate because of the amount of piping that must be designed?

The Lang Factor method and its variants do not provide insight into such issues. Overall, these benchmarks do not provide adequate specificity to help project teams improve performance and deliver successful outcomes. The conclusion is clear: equipment costs alone do not fully explain total project costs.