Hi everyone and happy holidays. Welcome to this two-part series on advanced risk management for Project Management Professionals. Some weeks ago, we explored the risk management knowledge area of the PMP exam. In that post, we examined the tools and techniques of all the processes that make up the project management knowledge area. You can review the posts; part 1 here and part 2 here.
In this post, we would further explore the tools and techniques for managing risk, the rationale for using these tools, the quality of information that can be obtained from these tools and techniques and how they can impact on project decisions and outcomes.
As a project manager, you are saddled with the task of ensuring that the project deliverables are met within the project schedule without overrunning the project resources. The reason why this is a complex task (and by extension, the reason why project managers are paid a lot of money) is because nothing is certain. In fact, there are so many things that can affect the project (and can easily happen) that you have absolutely no control over. Off the top of my head, some uncertainties that have affected some of the projects that I worked on include;
- Fluctuations in the weather.
- Fluctuations in exchange rates for imported goods.
- Resignation of one of the critical team members.
- National policies/laws affecting key deliverables in the project.
And so on…
You would have noticed that the list above does not include natural disasters like earthquakes, floods or hurricanes. Those kinds of uncertainties are beyond the scope of this articleJ.
The point I am trying to make here is simple. When making estimations, considering a single set of average estimations can be disastrous. For example, if we are working on a building project, and we estimate that based on average conditions, we should complete the building in 90 days, what are the actual chances of everything working as planned and completing the building in exactly 90 days? The probability is quite low. In fact, in most cases, the probability is less than 1 percent. However, more than 90% of project managers base their decisions on these single scenario estimations. No wonder why many projects fail. Hopefully, as from today, you would join the 10% of project managers who pay more attention to risk dependencies in projects!
In the rest of this article, we would examine three techniques of qualitative risk analysis; Scenario Analysis, Sensitivity Analysis and Probability distributions. These three techniques can be used to help in analyzing risky elements of a project and to provide the project manager with more information to make better decisions.
Scenario analysis involves considering multiple scenarios when forecasting various aspects of the project. A mild example of a scenario analysis that all project managers are familiar with is the PERT (Program Evaluation and Review Techniques) method of estimation. In the PERT analysis, the optimistic, pessimistic and most-likely scenarios of an estimation are considered and averaged using an algorithm, (P + 4m + O)/6, to determine the best part.
Similarly, in scenario analysis, different scenarios are analyzed to give a spectrum of results. For example, let us consider a small project to develop a website with the following criteria.
|Project Schedule||100 hours|
|Subcription Costs||400 Dollars|
|Personnel Costs||50 Dollars/hour|
For a scenario analysis, let us consider the best case and worst case scenarios;
|Project Variable||Base Case||Best Case||Worst Case|
|Project Schedule||100 hours||80hours||150 hours|
|Personnel Costs||$50 /hour||$40 per hour||$80 / hour|
The best and worst case scenarios show two possible extremes of the project outcome being measured. In the case of the project above, even though the costs are less than $6000 in the base case, the Worst case show a significant risk of the costs rising to over 200% and that needs to be investigated before embarking on the project.
When there are a few variables to be measured (like in the example above), the project manager can easily perform scenario analysis using a simple calculator. However, as the project dependencies become complex, more sophisticated tools might be needed to be employed to easily perform scenario analysis. One of those tools is the Excel Spreadsheet software developed by Microsoft Inc. Using Excel; you can input various scenarios of a value and then estimate the impact of those scenarios on a selected variable. This method can be really useful in complex analysis of dependent variables where a rough estimate can get confusing. A sample output of a scenario analysis of different scenarios created using Excel is shown below;
In the scenario analysis above, we can see the estimation of best and worst case scenarios for different variables that can affect the NPV of a venture. If you recall the article on Project Selection, one of the factors that can be used to determine which project should be selected is the NPV.
So how do we do this in excel? You use a tool known as the What-if analysis tool. You need a basic knowledge of how Excel works to do this. You just need to input your base estimations and link them together using excel formulas. Then you can run scenarios using the What-if Analysis tool. You can find the What-if Analysis tool under the Data Menu in excel (It’s the same for Excel 2007 and above). If you use a lower version of excel, just Google for where it is, or better still, upgrade your spreadsheet program J
Key Point: Navigation: Data —> What-if Analysis –> Scenario Manager
For detailed steps on scenario analysis using excel, you can use this tutorial from Microsoft.
Note: In addition to quantitative risk analysis, scenario analysis is an important tool for financial modeling. So if you have to communicate the pros and cons of a project to your financial director or CFO, this is a nice way to do that!
Although the scenario analysis is a useful tool, it is usually not enough to make accurate project decisions. One of the reasons for this insufficiency is because while the best (and worst) case scenario analysis shows the overall impact of the combined variables from an entirely optimistic or pessimistic perspective; it does not show the individual effect that each of the variables has on the outcome. Also, this analysis only shows the extreme effect of the variables. As mentioned above, it is very unlikely that all the variables would be equally positive or negative. For instance, if the likelihood of each of the variables to occur in their best case is 10%, the probability of the overall best case (where all the six variables are in their best case) to occur is 0.0001% which is approximately zero.
For project managers to determine which variable has the highest effect on a particular outcome, the sensitivity analysis technique can be employed. In plain words, we need to determine how sensitive an outcome is to each of the individual variables.
A mild example of scenario analysis that all project managers are familiar with is the critical path method. After estimating activity durations and creating a schedule baseline, if a project manager notices that a project is about to exceed the schedule, the project manger would focus on activities on the critical path. The reason is simple; any delay to an activity on the critical path would directly impact the overall project schedule. Similarly, this concept can be extended in analyzing other risk parameters. For example, considering the web development example above, if the project manager was able to control only one variable of the entire 5 project variables, which of those variables would that be?
The way to determine the most sensitive variable is to check the impact of each of them, while keeping other variables constant. Using this method, the Worst case scenarios for each of the independent variables would be;
|Project Variable||Worst Case||Total Cost|
|Project Schedule||150 hours||$8400|
|Personnel Costs||$80 / hour||$8900|
From the table above, the personnel costs have the greatest impact on the overall cost of the project and the project team should focus on controlling that cost. You should note here, however, that costs can have an indirect effect on quality and other aspects of the project which must be considered too.
Again, more specialized tools can be used to perform sensitivity analysis of variable on an outcome. An easy tool to do this is the 1 and 2 – Dimensional tables in the “What-if” analysis tool in excel. This tools can be used to compute an outcome while varying a single (or two) variables. A one dimensional table analysis for the subscription project schedule on the total cost is shown below;
The table clearly shows that when the time project schedule alone is varied while the other variables are held constant, the Project costs can range from $4900 to $8400.
For a better representation of the scenario analysis, more advanced tools can be employed. The most common one is TopRank which can be integrated to Microsoft Excel to determine scenarios. An example of a scenario analysis for the NPV example above using Top rank is;
The scenario analysis shows that the Growth in casual customers has the highest effect on the NPV. Again, these tools can be very useful in communication with project stakeholders and team members.
At the beginning of the post, I mentioned that we would explore three techniques, but since this article is already getting too long, I would break off the article here and we would discuss the last technique in a subsequent article. I would encourage you to try out these tools and techniques on projects of your own (whether real projects or fictional ones) so that you can get familiar with them. Trust me; they would come in handy at some point.
Thanks again for reading and I look forward to continuing this series soon.