Do you have the challenge that the PMO behaves more as a supportive function for individual projects, rather than an entity that looks at the big picture (the portfolio level)?
I have been working with several portfolio management cases, assessments, audits and consultancy assignments in different roles over the past 10 years. One thing always surprises me: when collecting information from different projects within the portfolio, almost all companies use variables that are either very difficult or even impossible to compare. This is related to the information that companies collect e.g., for risk management.
There are two issues here. One is that they are not comparable to each other and the other is that they don’t scale up on the company level. Let me explain.
We have two projects within the company and we are using traditional risk metrics, where we use probability and impact (a.k.a. severity in some models) with the same scale: low, medium and high. Sometimes we see this expanded to a scale with five stages, but it does not bring any improvement to the challenge.
If the projects are roughly similar, the model works on the portfolio level, but the problem becomes more challenging if the projects are different. The main problem is the estimation of impact. Each project looks at it from their own perspective and what is high in a small project may not even be on the list when looking at the same thing from the perspective of the larger project’s point of view. On the project level this is not an issue, but on the portfolio level it is. The risk information collected from projects is not comparable.
The other challenge when comparing the impact in relation to the size of the risk in an individual project. In most of the projects, the high risk is on the project level but not on the portfolio level.
My advice for creating a scale for impact is to use it as a monetary scale. That scales much better and is comparable between projects. On the project level, this model reflects a better estimate of the real cost of the risk. Suddenly, there could be two risks that both used to be high. Now, with the new scale, it can be seen that they are totally different. Maybe even some old middle impact is bigger than the old high. As an example, there was a case in which a key member of the project team was considered as a high risk in a project. Even though it was estimated that the probability was low, the risk revealed itself. There was an accident and suddenly the person was not available for the project. The project itself was worth a few million euros and the actual cost of replacing the key member of the project required us to work overtime with a replacement consultant. In the end, the additional cost was roughly €30k. Of course, this was an individual case and things could be totally different in other cases; but was the impact from project point of view high? I doubt it.
When looking at the portfolio level, you have much better comparable information on risks and grounds for decision making are much better. You can now, for example, collect only those risks that are above the threshold set on the portfolio level.
For probability I strongly suggest using basic mathematical terms that are familiar in probability calculations – percentage. It is far easier to calculate than the low, medium and high.
With this small change in your project model, you can make huge difference at portfolio level with the quality of the data you collect from the projects. Many of these ideas could be used for the prioritization of data as well.
If you want to create an even bigger effect, you could request to look for negative risks in your risk management model. These are better known as opportunities. The same logic applies.
For better results – the value driven model
Author: Tuomo Koskenvaara