There are several phases to the Schedule Risk Analysis process, and it is good to become familiar with each step along the way. Let’s take a look at this in more detail.
Acumen is considered a one stop shop for risk analysis. Not only can you examine the quality of your schedule, you also can perform the actual risk analysis. Acumen performs both functions. The Schedule Risk Analysis process is somewhat involved, so an overview of the major phases or steps is helpful. There are essentially seven steps to Schedule Risk Analysis process (SRA) including the following:
- Evaluate schedule quality and prepare the schedule for SRA.
- Assign uncertainty for duration/cost and define activity correlations then run an uncertainty analysis.
- Define discrete risk events and run analysis to find worst case scenario.
- Consider mitigation including multi-step plans.
- Review the mitigation analysis results and choose the appropriate mitigation plan.
- Update the schedule scope and cost estimate.
- Publish results at desired confidence-levels.
Again, the Schedule Risk Analysis process is somewhat long and involved. But these seven steps are your roadmap to guide you through each phase of the overall SRA effort.
This article discusses each step in the Schedule Risk Analysis process and provides a high-level overview of the overall effort.
1. Evaluate and clean up schedule
The first step in the Schedule Risk Analysis process is to investigate the quality of the schedule and prepare the schedule for risk analysis. A quality schedule is the first and most important element in an accurate SRA. It is not worth proceeding further to other steps until the schedule is thoroughly inspected and cleaned up, accordingly.
Schedule Quality Metric Group
Acumen has a Schedule Quality metric group that includes metrics that Acumen considers the most important metrics for a quality schedule. These critical metrics are the following: missing logic, logic density, critical, hard constraints, negative float, insufficient detail, number of lags, number of leads, and merge hotspots. If your schedule fails to achieve a passing score on this Schedule Quality analysis review the tasks or records that scored poorly. A sample output from a Schedule Quality analysis is in Figure 1.
DCMA 14-Point Assessment Metric Group
The Schedule Quality analysis has one negative; most of its metrics consider completed activities. Because nothing can be done about these activities it would be better to only include remaining work in the analysis. The Defense Contract Management Agency (DCMA) 14-Point Assessment is an industry standard for inspecting schedule quality. Additionally, most of its metrics (not all) only consider remaining work, which is what you want. So, run the DCMA 14-Point Assessment metric group analysis to further inspect the quality of your schedule. A sample output from a DCMA 14-Point metric group Acumen Fuse analysis is in Figure 2.
Core Risk Analysis Metrics
Additionally, the most relevant metrics to help prepare for SRA are the following: constraints, missing logic, leads, lags, and redundant logic. The DCMA 14-Point metric group considers hard constraints, but not soft constraints. Therefore, it is best to include the Acumen constraints metric found in the Advanced | Constraints subgroup in Acumen metrics. The ideal is to have no constraints in the schedule. But a few soft constraints may be acceptable. It is possible to create a metric group containing the four metrics: constraints, missing logic, leads, and lags. There is no metric for redundant logic metric, but this issue can be reviewed in S2 // Logic of Acumen.
After inspecting the schedule using the Schedule Quality metric group, DCMA 14 Point metric group, and a metric group consisting of the most relevant metrics to SRA, make any necessary updates to the schedule. Repeat the analysis until passing scores are achieved. At the end of this step you have a high-quality schedule that provides a foundation for an accurate SRA.
2. Assign uncertainty estimates for schedule, set up correlations, run analysis
Duration/Cost Uncertainty
Risk is inherent in scheduling. If a schedule has no known risk events, it still has duration/cost uncertainty that should be considered. In single point deterministic schedule models, it is assumed that the duration/cost is known with certainty. The reality is that the specified durations/costs are uncertain estimates. It is therefore best to model activity duration/cost using a probability distribution. A beta shaped distribution would be accurate but is difficult to achieve. A more efficient method is the triangular probability distribution from a three-point estimate. A triangular probability distribution specifies the optimistic, most likely, and pessimistic duration/cost estimate for each individual schedule task, Figure 3.
But even a three-point distribution can quickly become unwieldy and cumbersome on large projects. Acumen has a way to make the process more efficient. In Acumen the scheduler provides the most-likely duration and whether this estimate is conservative, realistic, or aggressive. Figure 4 displays the Uncertainty Factor Editor where all levels and corresponding Min, Most-Likely, and Max percentages are defined.
Acumen also has a handy slider system where the project, work breakdown structure (WBS) elements, or individual activities are assigned probability distribution triangle percentages based on whether the scheduler or team member considers the duration estimates to be conservative, realistic, or aggressive. In the Figure 5 the Duration Uncertainty slider is set to Aggressive for the displayed activities.
And for certain tasks the scheduler can override the slider and manually specify Min, Most-Likely, and Max duration estimates, Figure 6.
Activity-to-Activity Correlation
In addition to setting uncertainty estimates in step 2, the scheduler should also define any activity-to-activity correlation in this step. In the instance of causation, you want to define activity-to-activity correlation. Causation at times is not realized because one predecessor’s random number extended duration is canceled out by a successor’s random number shortened duration. Activity correlation helps prevent this cancellation by breaking up the sampling’s randomness in the triangular probability distribution. Correlated activities sample random numbers from the same end of the probability distribution triangle. If the predecessor samples from the upper end of the triangle the successor samples from the upper end. So, the successor takes it cue on where to sample from the random number generated by the predecessor.
A good example demonstrating how activity correlation works is scheduling a new vendor. You want how the vendor performs early on, either early or late, to correlate to successor activities. If your new vendor early on is in the upper tail end of the triangle you want this same pattern to prevail in later tasks. Another example is supplier delivery performance. A supplier that delivers an item late is likely to repeat this late delivery for other deliveries. In Figures 7 we have Design and Testing correlated activities.
In Figure 7 Design’s sample random number is from the upper end of the probability triangle, so the successor activity Testing’s random number is also from its triangle’s upper end.
After duration uncertainty is defined for all tasks and any activity-to-activity correlation specified it is time to run the SRA. The scenario to consider is duration uncertainty only. If you are running a cost model with schedule overlay it is good to first run the SRA with duration uncertainty off and cost uncertainty realistic. Then to investigate the effect of the schedule overlay run the SRA with duration uncertainty on and cost uncertainty realistic. In this way we find out if schedule delays have a material effect on the cost output.
3. Discrete risk events for schedule – run worst case scenario
Project risk is a fact; projects are unique, and, therefore, inherently a risky proposition. Not only do projects come with duration/cost uncertainty, but risk events are a major contributor to schedule unknowns. Acumen risk register lists specific risk events along with the probability of occurrence and impact.
Figure 8 displays a comprehensive probability distribution triangle that considers either duration or cost uncertainty and a risk event.
As you can see, the risk event may have a significant influence on extending the probability distribution triangle. In Acumen risk events may include threats, opportunities, calendar events, and risk windows. Opportunities may have a positive leveraging effect on the schedule.
Calendar events can describe issues, such as adverse weather during specified months of the year. And risk windows can define periods of time when work must suspend do to extenuating circumstances, such as eagle nesting season.
You cannot do work on your outdoor jobsite when eagles are nesting nearby. Another example is freezing weather in Canada that halts dredging work on a river.
The goal is a comprehensive risk register. Historical projects help formulate risks. Team brainstorming sessions are key to a well-thought-out risk register. But do not overdo the risk register. It is best to exclude those risks events that are not very probable and where there is nothing you can do about it.
A well-known example is an airplane crashing into your manufacturing facility. This would have a huge impact on your project, but it is not likely, and you cannot prepare for this airplane crash event.
Once your comprehensive risk register is complete you are ready to run the SRA for duration uncertainty and risk events scenario, excluding mitigation. This provides you the worst-case scenario for the project. The SRA results are a histogram and S-curve, and from this data you can find the project completion date and/or cost at your desired confidence level.
4. Brainstorm mitigation – run option analysis
Now we want to turn to mitigation efforts. We need to find out the maximum benefit from a respective mitigation plan. To do this we first run the SRA for a risk event and no mitigation. This is the worst-case scenario for that risk event. Then we disable the risk event in question (to eliminate it) to find out the most we can gain from a mitigation effort.
We continue and brainstorm mitigation. When we consider mitigation, it is important to note that mitigation can reduce either probability or impact, but generally not both at the same time.
This is known as “double dipping”. Now mitigation in Acumen can be one step or more. A classic multi-step mitigation plan is the steps to mitigate a failed pump risk event. This may include the following steps, each step reducing either probability or impact:
- Reduce probability by implementing a preventative maintenance plan.
- Reduce impact by purchasing a spare pump to swap out with the failed pump.
- Reduce impact by hiring a professional pipe fitter/HVAC engineer to install replacement pump.
Acumen has a Mitigation Analysis feature in the risk register tools that performs an SRA on a multi-phase risk mitigation plan. The Mitigation Analysis feature provides results for no mitigation and then for each following step in sequence. In this way we get a full picture on the benefit of each mitigation step. We can use this data to support our mitigation selection effort.
5. Select mitigation
Results from a sample ‘Boulder in Trench’ mitigation plan are in Figure 9.
This plan is to mitigate the likely risk event of a ‘Boulder in Trench’ during excavation. Clearly displayed on a bar chart are the results for no mitigation, and then each step in sequence. Mitigation analysis may be executed at any confidence level or P-value.
Our example results are for a P-value of P80 or 80% confidence level. Note the duration and cost of each mitigation step at the top of the bars. The duration and cost of the mitigation steps are included in the SRA when the scenario is set to uncertainty and risk events, including mitigation and overhead.
Some veteran schedulers prefer to keep the overhead expense of the mitigation step separate as it provides a more robust proposal. Otherwise, the expense can be diluted or overwhelmed by the overall duration/cost of the project. Once you have your mitigation plan and results you have everything you need to make an informed decision and select the right mitigation plan.
6. Update schedule and cost estimate
After you select your mitigation plan you are ready to update the schedule and cost estimate. Your mitigation plan may require updates to the schedule scope by adding additional activities and relationships. In tandem you should specify the cost of these inserted activities.
Again, Acumen will include schedule and cost impacts when the scenario is set to include overhead costs. But your duration/cost mitigation plan may be more involved and require updates to the scope of the project schedule, which may include additional activities and associated relationships.
7. Publish P-value schedule and cost estimate
Well, the final step in SRA is to publish the results. Acumen Risk is all about finding confidence levels and ranges that we know the schedule/cost is going to fall. Of upmost interest to stakeholders is what date and cost are associated for a respective confidence level or P-value.
A common desired P-value is P80, which, again, is an 80% confidence level. You can also build a scenario at your desired P-value that will generate a copy of your risk adjusted schedule, which you can then export back into your scheduling software.
Summary
It is good to have a roadmap to take along and refer to for any long journey. The Schedule Risk Analysis process effort is somewhat lengthy. A good overview is helpful to keep in mind as you enter each phase of the effort.
The first and most important step is to inspect the quality of the schedule and prepare it for the rigors of an SRA effort. An accurate SRA is most dependent on schedule quality. It is not worth the progressing to other steps until a quality schedule is confirmed.
Second, assign duration/cost uncertainty to tasks using triangular probability distribution. Also, if you have activity-to-activity correlation that should be modeled at this time. The correlation pairs or yokes two activities together so the random values generated in predecessor/successor activities are taken from the same end or location of the probability distribution triangle.
Third, risk events are major contributors to schedule duration/cost. Implement team brainstorming sessions to produce a comprehensive risk register. Then run a scenario including uncertainty and risk events to find your schedule’s worst-case situation.
Fourth, consider and run mitigation analyses. And remember that each individual mitigation step can reduce either probability or impact, but typically not both.
Fifth, select your mitigation plan. When running the analysis to include overhead, mitigation duration and cost expenses are inserted. Some schedulers prefer to separate the mitigation plan for a more robust proposal.
Sixth, update your schedule scope and cost estimate as appropriate for your mitigation plan. Your mitigation plan may require insertion of tasks and relationships.
Seventh, and finally, publish your results. Provide target dates and costs at desired confidence levels. And now that you have a risk-adjusted schedule you have more realistic goals that are more achievable in the real world.