ONE GOOD IDEA
Avoid ‘analysis paralysis’ with this action model for risk
by Jeffrey A. Robinson
Risk and uncertainty complicate decision making and problem solving. Making decisions is difficult enough when problems are clearly defined. But when there is uncertainty associated with potential courses of actions or outcomes, or when there is high potential cost of failure, the process of determining the best course of action becomes far more problematic.
Unfortunately, many decision makers do not properly distinguish uncertainty from risk. These two very different factors are often confused with each other or deemed to be nearly the same. When this occurs, there’s a tendency to spend too much time studying the problem, leading to a phenomenon often called ‘analysis paralysis.’
To start, risk is a relatively linear factor. It is easiest to equate risk with dollar loss—that is, how much money might be lost if a particular event occurs. The greater the potential financial impact, the greater the risk.
Uncertainty, on the other hand, is non-linear. When the probability of an event is very low, uncertainty also is very low—it’s almost certain the event won’t occur. When the probability of an event is high, it is almost certain to occur. Uncertainty peaks when the probability of an event is 50%, when the outcome of two events is equal and the result could go either way.
If you examine the relationship between uncertainty and risk, you can see there are four situations in which different strategies of action apply, as shown in Figure 1:
1. If risk (the cost of an event) is low and uncertainty is low (you know the likely outcome with high confidence), you simply act. For example, a $100,000 piece of equipment could break down because a 5-cent washer failed. What do you do? You simply replace the washer. No analysis is needed; you just do it.
2. On the other hand, if you have high uncertainty and low risk, your best actions are different. In this case, the $100,000 piece of equipment broke because of a failed washer, but the washer was so completely destroyed that you don’t know what size or type of washer it was. The risk of replacement is low (5 cents), but uncertainty is high because you don’t know the type of washer to be replaced.
In this case, the proper course of action is to deliberately seek failure. You go out and buy many washers of assorted sizes and try them out, one at a time, until you find the right size. After all, they only cost 5 cents each, and the faster you eliminate washers of the wrong size, the faster you find one that fits. While detailed analysis might get you to the same end point with low risk, there’s no cost of failure, so empirical testing of outcomes is faster and easier.
3. In the third case, you have high risk and low uncertainty. In this situation, the $100,000 piece of equipment broke, and you know the specific part that must be replaced (low uncertainty), but it’s the most expensive component in the machine (high risk), costing $50,000. In this case, the appropriate action is a quantitative judgment.
If the equipment is near the end of its life and doesn’t even produce $50,000 of revenue in a year, it wouldn’t be worth replacing the broken component. Conversely, if the machine produces $1 million in revenue a week, you want that component replaced as quickly as possible. Not much analysis must be done. A simple financial cost-benefit calculation would direct you to the correct decision.
4. The final scenario involves high uncertainty and high risk. You don’t know what caused the failure of the machine, and the cost of repairs could range from 5 cents to $50,000. This is when you must analyze the problem. Learn more about the failure, the problem, the costs and possible solutions before you select an appropriate course of action. Indeed, the purpose of the analysis is to reduce the uncertainty so you can make a judgment or reduce the risk so you can test possible outcomes.
Because uncertainty and risk are often confused with each other, analysis is often performed in the last three of the situations, leading to ‘analysis paralysis’—analyzing when you should be acting. The inability to differentiate between uncertainty and risk can lead to unnecessary analysis and preclude appropriate actions. To avoid this, decision makers must understand how uncertainty and risk differ and how their various combinations suggest specific courses of action.
Jeffrey A. Robinson is vice president of technology for Accelerated Quality Improvement LLC in Newtown, PA. He has a doctorate in information systems from Nova Southeastern University in Fort Lauderdale, FL. Robinson is a member of ASQ and an ASQ-certified Six Sigma Black Belt.