With design-bid-build, the engineer considers only the owner's requirements. The owner then assumes the risk of errors or poor design choices, which are resolved during the construction stage with a contractor typically requesting a change order and the owner paying additional money.
By engaging the contractor in identifying and eliminating risks and applying innovations during design, CM/ GC lowers the owner's risk.
With design-bid-build, project managers budget for risk by using a percentage of the project's total estimated cost. As they gain experience, they tend to increase this contingency factor. A project manager who experienced a situation where 5% wasn't enough to cover an unforeseen problem estimates the risk at 10% the next time; someone who found 10% wasn't sufficient estimates the risk at 15% for future projects. This approach limits available construction funds.
Our largest CM/GC project to date is the Mountain View Corridor, a continuous multimodal north-south route from western Salt Lake County to northern Utah County that's been identified in long-range plans since the 1960s. The system serves 13 communities with 35 more miles of freeway, 20 miles of transit that goes to and from Salt Lake International Airport, and a trail that runs along the freeway.
To extract as much value as possible from the overall $760 million budget — which includes anticipated right of way purchase for a freeway — we addressed risk directly rather than indirectly: Throughout design, we used a formal process to identify, assess, and mitigate each individual risk.
The team conducted four workshops to identify and assess risks:
- At concept phase; participants: UDOT, design team
- At 30% design; participants: UDOT, design team, contractor
- At 50% design
- At 75% design
Each workshop offered a new assessment of the project cost, taking into consideration the most up-to-date information on specific risks. (See chart on page 52.)
During the first workshop, the designers recorded each risk on a “risk register” that:
- Briefly described the risk
- Indicated strategies to eliminate or minimize the risk
- Provided data that could be used to statistically assess risk, such as the team's estimate of how likely a risk was to occur and a range of probable costs in the event that the risk did occur.
Using this data, the team ran statistical simulations to determine each risk's potential impact on total project cost. Using intervals that are typical for statistical analysis, a risk with a 10% chance of not exceeding the price was labeled low, a 50% chance was median, and a 90% chance was high.
The team computed the monetary value of each risk using software based on Monte Carlo simulation, a computerized mathematical technique used to assess the impact of risk. They presented the results as S-curves for:
- Cost probability
- The reduction in probable cost (the change in the median pricing at each stage)
- The overall contingency required for the project (difference between the median cost and the high cost on a given S-curve).
Once 30% of the design was complete, UDOT, the design team, and the contractor met to finalize the risk register. In this workshop the contractor also appraised the risks identified in the risk register, identified new risks, and presented a project cost estimate.
The next step was to tackle the risks.Flexible risk management
The team used two basic criteria to prioritize risks: cost and probability of a risk becoming an actual event.
The highest priority went to risks that were on the project critical path; i.e., those that would impact the sequence and scheduling of work. Risks that weren't on the critical path remained high priority if they carried a high cost. Risks that didn't have as great an impact were deferred, or simply absorbed, into the cost of the project.
We delegated responsibility to the team member most capable of managing the risk effectively. These “risk owners” were encouraged not to view the plans as unchangeable, but to consider the goals that specific provisions were intended to meet. The risk owner was free to propose alternate means of meeting project goals that reduced or eliminated the risk. Risk owners then investigated each solution and produced revised probability and cost estimates for the changes proposed.
The entire team evaluated all proposed solutions and selected the ones that fit best with the project as a whole. We gave new assignments based on the chosen mitigation strategy. With the mitigation strategies in place, the team compiled new high, median, and low cost estimates.
This process was repeated at regular intervals until the design reached 75% completion. At that point, team meetings stopped because risk had been adequately addressed and the team had a clear expectation of price to bid the project. However, risks were monitored and controlled through the completion of design and construction.
The project was awarded for $223 million — 27% below the engineer's estimate at 30% design, which is when a construction contract is typically awarded in design-build projects. We saved $126 million. Now that we know that CM/GC generates greater cost savings over both design-bid-build and design-build, we plan to increase this savings over the next few years by making incremental improvements to the process: applying design-build concepts to the contractor selection process and sharing savings to increase contractor effort in the design process.
— Alder (firstname.lastname@example.org) is engineering manager of innovative contracting with Utah DOT.WEB EXTRA
To access UDOT reports conducted at the end of design and at the end of construction for CM/GC projects, click here.
UTAH DOT COMPARES COSTS ON TWO BRIDGE PROJECTS
We awarded two bridge projects to the same contractor, one via design-bid-build and the other via CM/GC. Both involved offsite construction and a bridge slide into place. The bid openings were within three days of each other.
When unit prices were compared to state averages, the CM/GC project was 14% less expensive than the design-bid-build (see top chart). This is comparable to the savings generated in the Mountain View Corridor project; 27% less than the engineer's estimate and 15% less than design-bid-build (see bottom chart).
After the projects were completed, we added the savings for change orders to the award price savings and found the overall savings increased to about 19%.