Determining chance of failure

Finding the most complete possible list of chance of failures is not an easy task. The internet, literature studies, experience with similar systems and interviews or sessions with experts are tools to help map the chance of failure as adequately as possible. The ‘Risk factors per quality characteristic’ checklist specifies chance of fault possibilities for each characteristic.

In an interview or session, the participants estimate the chance of fault and frequency of use per object part. Generally speaking the (intended) users of the product are the people most capable of estimating the use frequency and the technicians those best able to estimate the chance of fault. Asking the participants to answer a number of specific questions per object part is an effective way to gain insight into the chance of failure.

1. In what way(s) could the object part fail?
2. What are the possible causes of such a failure?
3. What is your estimate of the chance that the object part will fail?
4. On which processes will the failure have a negative impact?
5. How often are these processes executed?
   

The starting point is that all chance of failure indications start at ‘low’; the participants must argue why the chance of failure is higher for specific requirements.

In the example below, the classes High, Medium and Low are used to classify the chance of failure. As an addition we use the classification ‘very low’ when an item, whether it is a requirement or a subsystem is about to be forgotten.

Characteristic: Functionality

Object part	Chance of failure	Arguments
Total system X	Low	The chance that all subsystems will fail is extremely small.
Subsystem 1	High	This subsystem is accessed hundreds of times every day and is built with technology new to the company.
Subsystem 2	Low	Is built with familiar and reliable technology.

^{Example of a chance of failure table for the characteristic of functionality.}