Predicting the unpredictable

By accurately estimating the probability and impact of risk drivers, we can create a risk-adjusted project schedule.

FEDERATION TOWER IS ONE OF THE LARGEST construction projects in Moscow, Russia, but it has had a troubled history. The project team initially planned a 62-story west tower and a 93-story east tower, with a 506-meter spire in between. In 2008, the west tower was completed, but work on the partially built east tower was suspended due to lack of funding. Construction resumed in 2011, but the halt had caused technical issues that had to be fixed. The next year, a fire delayed the project again. Then a main contractor was replaced. In 2014, stakeholders decided to demolish the partially erected spire, change the design and increase the number of floors to 95. The plan is now to complete the tower this year.

Project managers viewing this troubled project might wonder: Could improved risk management have prevented any of this delay and rework? Fortunately, modern risk analysis techniques allow us to create an accurate project forecast with ranges that account for risks and uncertainties. By accurately estimating the probability and impact of risk drivers, we can create a risk-adjusted project schedule. Analyzing this schedule provides us with answers to very important questions: What happens to the schedule if certain risks occur, and what is the chance that the project or a certain milestone will be completed on time and on budget?

In the case of Federation Tower’s east tower, the delays and extra costs were caused not only by separate events, but by chains of events: The global financial crisis led to the suspension of construction, which eventually caused design changes and the decision to demolish the spire. Such event chains may lead to a dramatic escalation of project costs and major delays.

So the next step toward a better modeling of project uncertainties is event chain methodology. This focuses on analyzing the relationship between events: Events can trigger other events, which may require mitigation and response activities that create their own events. Event chain methodology also helps us perform a risk analysis of portfolios, since the same risk can affect different projects.

It might sound complex, but it doesn’t have to be. Here is a brief outline of how to perform event chain methodology:

1. Design schedules for projects within a portfolio.
2. Identify risks and enter them into the risk register. Risks may have different properties, including costs.
3. Assign risks to activities and resources. During this stage, the risk probabilities and impacts can be defined. The same risk, assigned to different activities and resources, may have different probabilities and impacts. The same risks also may belong to different categories: For example, the risk of fire may impact safety, cost and schedule.
4. Identify risk mitigation and response plans and associate them with activities on project schedules. These response plans may have risks themselves. For example, during the Alaskan Way Viaduct Replacement project in the U.S. state of Washington, a tunnel-boring machine broke. To repair it, a 120-foot vertical shaft had to be drilled down to the machine’s cutting head. This risk response project delayed construction for more than a year.
5. Identify relationships between risks and determine event chains. Risks can trigger each other or be correlated with each other.
6. Create an event chain diagram (like the one on the previous page), an easy way to visualize event chains. Threats and opportunities assigned to the activities are shown as arrows on the Gantt chart. The size of the arrow depicts risk probability. The color depicts impact.
7. Identify uncertainties in project activities that are not part of the discrete risk events, and define their cost and duration using three-point estimates (best case, most likely and worst case). For example, the quality of concrete has delayed many construction activities. This delay can be modeled using three-point estimates of the duration of affected activities.
8. Perform Monte Carlo simulations of project schedules. This means project schedules will be calculated multiple times with different combinations of events.

After the calculation is performed, the analysis allows us to:

1. Create a risk-adjusted project schedule. The schedule can be associated with a certain probability that a project will be completed on time and on budget. We can also determine a probability that certain milestones in terms of cost and schedule will be met.
2. Rank risks based on their cumulative probability and impact on multiple activities and resources. Higher-ranked risks can be mitigated or avoided first. Event chain methodology allows the ranking of risks belonging to different categories separately or the calculation of a combined ranking based on the relative importance of different categories. We can also identify critical event chains, which can be broken if necessary.
3. Rank projects in a portfolio based on their risk exposure.
4. Identify the probability that certain mitigation and response plans will need to be executed, so these plans can be determined in advance.

Modern software significantly simplifies risk analysis of portfolios. Risks can be constantly tracked during the course of projects. A risk’s probability and impact may change over the course of a project; therefore, project and portfolio risk analysis should be performed on a regular basis. The main advantage of project risk analysis is that, if done properly, it provides realistic project schedules, and improves both the quality of project decisions and project outcomes. PM

Lev Virine, far left, is a consultant at Intaver Institute, Calgary, Alberta, Canada. Michael Trumper is business development manager at Intaver Institute.