Project portfolio evaluation
For a long time I looked at an empty Product Management blog with a call to write something, and decided to write something. But the topic is closer to project management. I believe, product management is more about NPD and others like this, but we will be interested in the yield and risk of the project portfolio. Isn't the maximization of income at a given level of risk the goal of portfolio management? (A rhetorical question).
Immediately I say, talking about dogs, stars, milk cows and so on will not work.
Update: an example . The question is - what would you give up, having one and a half thousand for the next 3 years?
The theme for real projects is poorly developed, despite the abundance of software solutions in this area (each of these solutions is in a lot of ways). There is a PMI standard for managing project portfolios, but I, after reviewing the Internet and re-reading a bunch of things, did not find a solution that would fully support it.
What is included in portfolio management? Something, but not all, is:
If you search in books, you can find various portfolio management tools.
For example, the most interesting ones in my opinion are:
These are all qualitative methods of working with project portfolios. Some of the above can be found for example in Microsoft Project Server 2010 . It was the reading of this article that by the way prompted me to write this post.
Why do I call these methods qualitative? Because they do not say anything about the profitability of the portfolio, or about its risk directly. Yes, you can balance a portfolio with reference to strategic goals. But this, in my opinion, is not all, as long as we do not know how much the portfolio will bring to us with what probability.
From here on, I will continue to talk about investment projects, meaning the financial result (the difference of all revenues and costs for a project) as the main indicator of efficiency. If NPV is more interesting (suddenly), then all the arguments below can be repeated for it.
What is the risk of an investment project? This is the possibility of such a coincidence that its financial result will be less than K (I will take K = 0 below).
')
Measures of risk are different. For example, probability. For example, the probability * amount of losses. Finally, the damage itself. All these measures we consider.
To work with risks, we need to structure uncertainty, and this is best done according to future scenarios. For example, consider three sales cases: optimistic (level A), realistic (level B), and pessimistic (level C).
Take the “probabilities” from the PERT method (in fact, these are weights, but the expectation implies probabilities under this - and there is some feint in our ears. If you want, calculate the probabilities in some other way, for example, by asking yourself - wrong in sales in N, M, L times? "): 1/6 for A, 4/6 for B, 1/6 for C.
Weigh the financial results of the scenarios of probabilities, get an estimate of the financial result. Secondary - this is the expectation of the project.
See in which cases the financial result is less than zero (in the pessimistic, apparently). Here, with the likelihood of this scenario, we have damage equal to the negative financial result.
If there are more than one scenarios with a negative financial result, then
What is for the portfolio? For independent projects we are free to sum up the financial results. For addicts, we are obliged (for simplicity, not to bother with correlation) to combine dependent projects into one.
Further, under the accepted conditions, it is clear that the financial expectations for the portfolio are equal to the sum of the financial expectations for the projects.
We remember that for independent random variables, the variances are summed up, which means we can calculate the variance for each project and sum it up.
The variance for the project is considered as follows: add up the squares of differences of the average financial result and the financial result according to the scenario, weighted by the probabilities of the scenarios.
Add up the variances on the projects, extract the square root - we have the standard deviation.
There is such Chebyshev inequality - says that the deviation in k standard deviations and more is realized with a probability of no more than 1 / k ^ 2. If it is easy to imagine that the distribution is symmetric (if PERT was used, it is so), then this probability must also be divided into two - we are only interested in the probability of deviation to zero. k is here equal to the quotient of dividing the estimated (average) financial result by the standard deviation.
Thus we get the probability of loss. Those. level of risk.
It is clear that in order to compare portfolios, it is necessary to make calculations for portfolios.
Nevertheless, we now know how to find out the expected return of a portfolio in money, and the corresponding probability of obtaining a financial result is greater than zero. And then it remains only to act!
PS The method is mine (done simply by analogy with a portfolio of securities), if you suddenly find a mistake - I will be very grateful.
PPS I hope it was interesting and useful :)
Immediately I say, talking about dogs, stars, milk cows and so on will not work.
Update: an example . The question is - what would you give up, having one and a half thousand for the next 3 years?
A bit about portfolio management
The theme for real projects is poorly developed, despite the abundance of software solutions in this area (each of these solutions is in a lot of ways). There is a PMI standard for managing project portfolios, but I, after reviewing the Internet and re-reading a bunch of things, did not find a solution that would fully support it.
What is included in portfolio management? Something, but not all, is:
- Collecting information about projects and writing out their economic indicators, resource needs, compliance with strategic goals, quality benefits
- Grading projects by criteria (criteria are combined into indicators), their weighting and calculation of the project rating (by weighting according to the weights of the criteria of the ratings). An example of a criterion can be net profit (less than zero, a score of 0, from zero to 5 million rubles, for example, a score of 5, and if higher, then a score of 10)
- Setting priorities. You can compare pairwise projects (for example, someone has a higher rating, or one of the indicators - he has a higher priority) by writing them in a table (columns and lines are projects. If the project in the row is more priority than the project in the column, a single is put. everything is summed up, and it turns priority)
- Analysis of portfolio balance. Various bubble charts are drawn based on the available data (both on indicators and on other data), looking for “clusters”, “empty spaces” and others - everything that can provide qualitative information on the portfolio. At this stage it can be understood that there are not enough projects in the portfolio.
- Align projects in time and resources in accordance with their ratings. Top priorities go ahead, if possible, and allocate key resources to them. Some projects may be rejected.
- Portfolio Authorization
- Portfolio Monitoring and Changes
If you search in books, you can find various portfolio management tools.
For example, the most interesting ones in my opinion are:
- Common threads. This refers to the common components of the project (not necessarily the software components, although they too). For example, you can search for common threads in the area of ​​customer interaction.
- Crushing the project. By this is meant a partition into such pieces, each of which gives a measurable benefit (best of all in money), and after each piece you can decide to change (close) the project
- Testing on scripts. Several business development scenarios are written out (for example, narrowing, retention, and expansion). You can see how each project will behave. Those that remain in all scenarios are probably the most significant.
These are all qualitative methods of working with project portfolios. Some of the above can be found for example in Microsoft Project Server 2010 . It was the reading of this article that by the way prompted me to write this post.
Why do I call these methods qualitative? Because they do not say anything about the profitability of the portfolio, or about its risk directly. Yes, you can balance a portfolio with reference to strategic goals. But this, in my opinion, is not all, as long as we do not know how much the portfolio will bring to us with what probability.
A little bit about the risks of projects
From here on, I will continue to talk about investment projects, meaning the financial result (the difference of all revenues and costs for a project) as the main indicator of efficiency. If NPV is more interesting (suddenly), then all the arguments below can be repeated for it.
What is the risk of an investment project? This is the possibility of such a coincidence that its financial result will be less than K (I will take K = 0 below).
')
Measures of risk are different. For example, probability. For example, the probability * amount of losses. Finally, the damage itself. All these measures we consider.
To work with risks, we need to structure uncertainty, and this is best done according to future scenarios. For example, consider three sales cases: optimistic (level A), realistic (level B), and pessimistic (level C).
Take the “probabilities” from the PERT method (in fact, these are weights, but the expectation implies probabilities under this - and there is some feint in our ears. If you want, calculate the probabilities in some other way, for example, by asking yourself - wrong in sales in N, M, L times? "): 1/6 for A, 4/6 for B, 1/6 for C.
Weigh the financial results of the scenarios of probabilities, get an estimate of the financial result. Secondary - this is the expectation of the project.
See in which cases the financial result is less than zero (in the pessimistic, apparently). Here, with the likelihood of this scenario, we have damage equal to the negative financial result.
If there are more than one scenarios with a negative financial result, then
- probability is equal to the sum of the probabilities
- the damage is equal to the weighted average damage by the probabilities of negative scenarios divided by the sum of the probabilities (the risk measured by the second measure is equal to the sum of the risks by the scenarios for the entire project).
Calculations
What is for the portfolio? For independent projects we are free to sum up the financial results. For addicts, we are obliged (for simplicity, not to bother with correlation) to combine dependent projects into one.
Further, under the accepted conditions, it is clear that the financial expectations for the portfolio are equal to the sum of the financial expectations for the projects.
We remember that for independent random variables, the variances are summed up, which means we can calculate the variance for each project and sum it up.
The variance for the project is considered as follows: add up the squares of differences of the average financial result and the financial result according to the scenario, weighted by the probabilities of the scenarios.
Add up the variances on the projects, extract the square root - we have the standard deviation.
There is such Chebyshev inequality - says that the deviation in k standard deviations and more is realized with a probability of no more than 1 / k ^ 2. If it is easy to imagine that the distribution is symmetric (if PERT was used, it is so), then this probability must also be divided into two - we are only interested in the probability of deviation to zero. k is here equal to the quotient of dividing the estimated (average) financial result by the standard deviation.
Thus we get the probability of loss. Those. level of risk.
Conclusion
It is clear that in order to compare portfolios, it is necessary to make calculations for portfolios.
Nevertheless, we now know how to find out the expected return of a portfolio in money, and the corresponding probability of obtaining a financial result is greater than zero. And then it remains only to act!
PS The method is mine (done simply by analogy with a portfolio of securities), if you suddenly find a mistake - I will be very grateful.
PPS I hope it was interesting and useful :)
Source: https://habr.com/ru/post/130016/
All Articles