EVM for infrastructure projects Part 2 - Better planning delivers better results

Last week we discussed the principles of earned value management (EVM) and the advantages using this approach can bring to major infrastructure projects.

At the core of EVM is development of a common definition of the work between the scope, cost plan and schedule. This is achieved through the use of a work breakdown structure (WBS) which easily captures all the associated works and provides a common framework. These metrics of each element of the WBS are then readily combined in the project baseline. The key questions on an infrastructure project are:

• How should the WBS be structured?
• To what level should the WBS be defined?
• What should be included or excluded from the earned value baseline?

The definition of the project scope starts early in the planning phase and will be progressively refined until the design is complete. This is consistent with how the WBS should be developed. At the beginning of planning, the WBS should be high level – a simple breakdown of the major work elements which make up the project. It’s then refined further in the same way that the cost plan and program are developed during design and planning. In fact, the cost plan and program at each stage should both be structured in the same way as the work breakdown structure.

Structuring the WBS

The WBS should be structured so that it aligns with how the project is planned to be delivered and, as far as possible, procured. This will ensure three things happen:

1. The procurement plan will be developed early in the process
2. The schedule will be easy to assess for progress
3. Costs will be easily attributable to elements in the WBS

We also recommend following these principles:

• Most projects are delivered as a series of work elements. The components of the project scope should be the top level of the WBS. For example, each building, or each civil structure or type of engineering services infrastructure, might be a work element.
• Where work is procured across several elements and is difficult to apportion, this work should be separated into its own element. For example, bulk civil works might be procured as a single contract which may be difficult to apportion to individual buildings which sit on the pad.
• Where work is procured across several elements and is easily attributable to each, these should be treated as a subsidiary element. For example, the electrical trade might be procured across several buildings, but both progress and cost can be easily measured against each building.
• Clearly delineate between direct construction work and support functions as these should be measured separately for earned value analysis.

The WBS shouldn’t be broken down any further than is necessary for the accurate and objective assessment of progress and cost, as it won’t improve the outcomes and will massively increase the impost of preparing and managing the earned value. In most cases, this means the WBS stops at the subcontract level. A simplified example from an airfield project is shown below.

A WBS which reflects the procurement plan is ideal however the reality is that procurement plans for major projects can change over time – as the market appetite for certain sized contracts or other priorities change. This doesn’t mean the WBS needs to be restructured each time the procurement plan is adjusted. It simply means the cost of the contract needs to be distributed across the WBS elements as appropriate. It’s important the WBS dictionary clearly defines what costs should have been assigned to which element to avoid distortion of the earned value metrics.

Defining the earned value baseline

We recommend that only the WBS elements related to construction work be included in the earned value calculation. Supporting functions such as project management, design, preliminaries and construction supervision should be treated separately – for these three reasons:

• Many support tasks don’t have their own timeline and are tied to the completion of other related construction activities. Project management is an example of this – our role is over when the other project activities are completed. This makes the calculation of earned value for that activity difficult and unreliable, and it effectively double counts the schedule variance of the tasks to which it is tied.
• While support tasks are critical to achieving the objectives of the project, it’s the construction works which are important to assess progress. For example, progress on design work, or lack thereof, is only important when it starts to impact actual construction. Unless it’s on the critical path for construction we recommend that it’s not included in the earned value baseline.
• The highest risks are associated with construction costs, and their costs are typically far higher than the cost of support tasks. Consequently, variances in cost and schedule are often almost entirely attributable to the construction work. Excluding the support tasks makes the calculation easier to manage without reducing the reliability of the earned value outcomes.

However, under some contract forms, it may be more convenient to include more of these details. For example, the Managing Contractor form of contract splits management costs from all construction and design costs.

Once the WBS is fully defined, a WBS dictionary should be produced. For each element the dictionary specifies a description, scope, duration, estimated cost, the cost curve to be used (for example a straight line or S-curve), whether it’s a direct cost or an indirect cost, and an objective measure to calculate earned value.

Developing the cashflow

As a general rule, the earned value baseline should reflect the planned cashflow for the project – work should be scheduled as it’s planned to be delivered. However, if management of the critical path is more important to the project than cashflow, then the project manager might consider scheduling work as late as possible. This has the effect of producing a positive schedule variance if the project is ahead of the critical path, or a negative schedule variance if it’s behind.

For each work element, the planned value should be calculated to be consistent with how costs will be incurred. This should reflect the expected payment structure in the contract and the expected cashflow. In most construction contracts, this is simply payment based on percentage complete however there are a few special cases, for example

• Supply contracts - supply of plant and equipment is often based on milestones such as ordering and delivery. The planned value should be based on when these milestones are scheduled to be achieved. The earned value graph for this element should be similar to the graph below.
  
• Design - design contracts are often based on progress payments to a maximum which can only be exceeded on achieving a certain milestone. This arrangement should be specified in the WBS dictionary. The earned value for this type of contract is shown below. A delay in achieving the milestone causes the value earned to pause until the milestone is achieved, as shown in the first milestone below.
  
• Decontamination – decontamination activities are difficult to estimate in scale, rate of expenditure and duration. However, as they often represent a major risk to the project, it’s important to include them in the earned value baseline. We recommend that the duration of decontamination is aligned to the bulk civil works in contaminated areas. The project manager should carefully consider the cost curve to be used based on where the works are being conducted, where contamination is suspected or known to exist, and the expected rate of expenditure. This might look like the curve below, with a higher gradient where more civil works are being conducted or heavier contamination is expected.
  

Design and construction contingency shouldn’t be included in the earned value baseline unless it can be aligned with a particular, definable and programmable risk. Management reserve type contingencies shouldn’t be included at all. The reasons for excluding these contingencies are:

• It’s impossible to predict when design and construction contingency will be spent. If it’s included in the baseline, over-expenditure against the cost plan can be concealed by the buffer that the contingency provides. That is, the retention of contingency to deal with unanticipated issues later on is preferable to spending the contingency on a prorated basis over time. Therefore, the baseline which performance should be measured against shouldn’t include the contingency.
• Earned value can easily be drilled down into to review the performance of construction managers responsible for individual work elements, or groups of elements. Performance at this level should be measured against predicted expenditure for that work element, without contingencies. Therefore, contingencies shouldn’t be included in the earned value baseline for each work element, or by extension the overall baseline for the project.

Look out in the new year for Part 3 where we discuss the how to use earned value reporting to manage the unique characteristics associated with construction projects.

In the meantime, I'm interested to hear about your experiences with EVM on infrastructure or other construction projects. Have you used it? How did you structure the WBS? What did you include or exclude from the EVM baseline? How did you treat the special cases that I've highlighted or others that you identified? What would you do differently next time?

About the author

Daniel Kenny is a Project Director who has worked with Coffey for more than eight years across the Defence, resources, energy, urban development and airfield infrastructure sectors. Daniel is currently leading a $1.5B project for the Department of Defence to deliver a range of airfield, engineering and building infrastructure to support the introduction into service of the new Joint Strike Fighter.