Deciding in Light of Uncertainty and Risk 05 April, 2011

Decision making can be difficult because there are often many options to chose from, and varying levels of risk are built in to these options. Many times, when people talk of managing risk, they simply mean transferring the risk elsewhere. Decision and risk are necessary issues that must be dealt with, simply because business is a competitive environment. Conservatism can work for a period of time, but it opens the door for others to leapfrog a business that is unwilling to take some risks or make decisions involving uncertainty.

Risk and uncertainty often are the result of incomplete information, ambiguous information, or time constraints. Sometimes, they are a result of people interpreting the same information differently. A representative sample may be necessary from which to base a decision.

Another reason that this subject warrants attention is because common solutions for dealing with risk and uncertainty have become flawed. History is not an indicator of black swan events. Although using historical data to analyze current situations can be useful, it should probably be used as a last resort in many scenarios. The following set of generic and specific frameworks will help identify what a correct decision should be, regardless of the uncertainty level.

Decision Making Steps
  • Recognize the need for a decision
  • Generate alternatives
  • Assess alternatives
  • Choose among alternatives
  • Implement the chosen alternative
  • Learn from feedback

You may be able to narrow alternatives by removing those that are not:
  • Legal
  • Ethical
  • Economical
  • Practical

Risk Matrix

A risk matrix simply interpolates and visualizes what risk may be associated with familiarity of products/technology, as well as what risk may be associated
with familiarity of markets.

Reality Check
Simply asking the following questions can provide a person with a reality/sanity check and as an initial feasibility test. An entire Harvard Business Review article was written on this:
  • Is it real?
  • Can we win?
  • Is it worth doing?

Cynefin Framework
A decision is categorized into one of four levels of order/organization, and then dealt with accordingly:
  • Simple - Clear cause and effects -> Use best practices
  • Complicated - Multiple right answers exist -> Use expertise
  • Complex - No visibly right answer -> Look for patterns
  • Chaos - No right answer, no pattern -> Establish order

Uncertainty Framework
Three academics (Courtney, Kirkland, and Viguerie) established a model for identifying levels of uncertainty and dealing with them.
  • Level 1 - Basic uncertainty -> Learn the required information
  • Level 2 - A few possible outcomes exist -> Use a decision tree
  • Level 3 - No discrete number of outcomes -> Use scenario planning
  • Level 4 - Even variables are discrete or unknown -> Use analogies to simplify

Decision Styles
Subordinates are all on a spectrum of needed supervision. When others cannot or will not decide, you decide for them. Otherwise, responsive, intellectual, and participative decision styles should be used.

Decision Trees
If a few requirements are met in a given scenario, it is possible and useful to use decision trees to narrow down and grasp options. The following steps outline how to create and use a decision tree:

Prerequisite - You must know all the alternatives, be able to assign probabilities to them, and have clear objectives.
  1. State your decision
  2. Draw branches for any intermediate results that could occur
  3. Draw branches any decisions that should be made at that stage
  4. Repeat steps 1 & 2 until no more intermediate results or decision points exist
  5. End each final branch with a outcome result ($ for example)
  6. Multiply probabilities and outcome values where it makes sense
  7. Based on all probabilities and final objectives, choose the best outcome

Example Decision Tree

Avoiding Decision Making Pitfalls
  • Don't form an immovable hypothesis from the piece first information you get
  • Don't justify decisions simply from historical data
  • Don't use the status quo as a benchmark for success
  • Get an outside point of view
  • Phrase the problem differently to see other sides
  • Be mindful of over emphasis by individual sources
  • Watch out for assumption padding on multiple levels
One of the most important things that can be done to improve decision making abilities is to receive quick and clear feedback after a decision has been made and results are available.

No comments:

Post a Comment