The Level of Math in Risk Management March 28, 2015Posted by Edwin Ritter in Project Management.
Tags: decision matrix, decision process, decision theory, known unknown, risk, risk management, uncertainty, unknown unknown
add a comment
Managing risk and selecting the best mitigation choice can be hard. There are many factors that will influence what choice is selected. Using data to manage risks and assess choices is always good. Having a process to use that data to evaluate choices is better. Having good data to assess risks and a process to determine choices is the point of this post.
Some background first. Decision theory is a complicated subject and is used in many different fields in business and leisure (i.e. – games). There are 4 basic elements in decision theory: acts, events, outcomes, and payoffs. A formal definition is “the mathematical study of strategies for optimal decision-making between options involving different risks or expectations of gain or loss depending on the outcome.” An informal definition of decision theory could be “identifying the values, uncertainties and other issues relevant in a given decision, its rationality, and the resulting optimal decision.”
Got that? Good. Sorting out between what is known and unknown can assist in which risk mitigation choice is best. It is always preferable to be in the magic quadrant and deal with knowns rather than unknowns. Consider these groups in a matrix:
- Known Known – circumstances or outcomes that are known to be possible, it is known that they will be realized with some probability.
- Known Unknown – circumstances or outcomes that are known to be possible, but it is unknown whether or not they will be realized. This is known as a Risk.
- Unknown Known – circumstances or outcome a modeler intentionally refuse to acknowledge that he/she knows.
- Unknown Unknown – circumstances or outcomes that were not conceived of by an observer at a given point in time. This is known as an uncertainty.
In short, those things we know we don’t know are risks and those things we don’t know what we don’t know are uncertainties. In the magic quadrant, both are below the line.
Still with me? The uncertainties, the unknown unknowns, are the hardest pieces to deal with. With decision theory, assigning values of probabilities helps determine which choice is best. As an example, consider the following image. Getting the data to assign the options and organizing into a concise arrangement is where the math comes in.
Not all risks require this level of math and analysis to determine the best mitigation choice. The level of math should be appropriate to the complexity of the project and the potential risks involved. It is worth noting that the difference between an issue and a risk is :
– an issue is something that has occurred which impacts* the project.
– a risk is something that may or may not occur that can impact the project.
* impact can be positive or negative.
I prefer to stay above the line and deal with knowns and those risks that are relatively small on my projects. It makes for simple math and for projects where my preferences matter, I find those are the best projects to manage.
Glad I did not lose you and you got this far. How do you manage risks? What is the level of math required for risks? Simple? Complex? Hybrid? Comments invited.
Keep getting Smarter November 28, 2014Posted by Edwin Ritter in Grab Bag.
Tags: closed loop, closed loop approach, communications, decision process, feedback, posterior probability, probability
add a comment
A fundamental challenge in every job is communications. Getting your audience in sync with your message is fundamental to communicating effectively. In previous posts, I have talked about using a closed loop approach, the value in feedback and how numbers improve decision making.
Segue to how the concept of feedback or closed loop is used with mathematics. Specifically with probability and ratios. There is a method to better understand probability outcomes that includes current and prior knowledge. Formally known as Bayes’ Rule (aka, Bayes’ Theorem or Bayes’ Law), it provides a way to numerically express how a degree of belief should rationally change to account for evidence of that belief. It gets complicated but basically, it involves prior and posterior probability.
The posterior probability is the Bayesian inference. The value in using Bayes’ Rule is starting with a prediction, getting results and then improving the prediction based on those results. What I find interesting is it provides a way to get smarter about something in the future and uses a ‘closed loop’ to do it. The intent is to keep getting smarter based on what you know and what you learn.
Improved insight or knowledge may not require using Bayes’s Rule, and whatever form is used for communications, it is important to accurately state what you know and reserve the right to be smarter in the future.
Making Decisions with Data March 25, 2014Posted by Edwin Ritter in Miscellaneous, Project Management.
Tags: data evaluation, decision matrix, decision process, Kepner-Tregoe, root cause, six sigma
add a comment
How familiar is this phrase “We are going to be data driven”? I have heard this a few times in my career. Great concept and useful to manage a business if you are serious and use it consistently. However, when crunch time comes, and everyone’s nerves are worn thin, how many managers stick with the data vs. using their instinct to make a decision?
No one intends to make a poor or bad decision. Assumptions can be wrong; risks occur that are not foreseen. It happens.
The concept is obvious of course, but, one of the most effective ways to make a decision is to use solid data. What is not obvious is the process to define, collect and evaluate data to make an informed choice. Other real world considerations like time, money and deadlines may circumvent staying true to a data driven process.
All things being equal and when there is adequate time, the process I most prefer uses selected weighting on a set of criteria. The process is commonly known as root cause analysis as a decision making method. Most refer to this process as Kepner-Tregoe analysis. It is named after the two people who invented the concept and today, their company is a multi-national consulting company. This method is one widget in the Six Sigma toolkit and is considered part of ITIL practices for problem management.
An overview of process includes :
- State the issue, problem, and decision to be made.
- Explain the use of the decision matrix technique to participants.
- Draft a matrix … with candidate choices positioned as rows and criteria as columns.
- Weigh the criteria, if required (e.g., 1-5 weight).
- Rate each choice within each decision/selection criteria (e.g, 1-5 score – do not rank here).
- Multiply the rating by its relative weight to determine weighted score.
- Total the scores.
- Review results and evaluate, using common sense and good judgment.
- Reach consensus.
Once complete, you have criteria and weighting configured in the decision matrix. When evaluating choices, the score helps narrow the discussion to the best choice(s). The discussions on reviewing the results can lead to animated discussions. Ultimately, the best choice comes down confidence in what the numbers tell you. I like to include a tie-breaker or ‘other’ category in the matrix and give it a small weighting of 5 to 10%. That allows a way to include intangibles discovered during the evaluation. Depending on the score for that facet, it can illuminate the best choice and help the team decide between two otherwise equal choices.
This evaluation process can be used for a range of situations where decisions must be made. I have used this for vendor selection, candidate interviews and for strategy roadmaps. In the end, having data can confirm your choice and give confidence. Using this framework also minimizes biases and leads to an improved appreciation of choices you would not have considered otherwise. Having data is always good; having a process to make a choice with that data is even better.
What process do you use to decide?