Appendix B: Description of Decision Types
The following is an excerpt from The Process of Reliability Engineering, a book by Carl S. Carlson and Fred Schenkelberg. Within the book see section 8.4.2 Select potential methods to inform key decisions for more information on using the nature of a decision, the decision type, as an aid in identifying the appropriate reliability method to include in your plan.
As an aid when selecting methods intended to inform key decisions, we have found that we generally face six different types of questions:
- Prevention: What can we do now to avoid failures or improve reliability?
- Comparison: Which design, vendor, or procedure option is better considering reliability?
- Priority: Where should we focus our resources to best improve reliability?
- Resources: Who and when should accomplish a specific task?
- Objective: How to set or what are the reliability and availability performance objectives, goals, or requirements?
- Measurement: What is the reliability performance now or expected to be in the future?
In Appendix C, we list to which decision type the results of the method inform, e.g., Derating (Prevention).
In the categorization of tools within these six types of decisions, there is some overlap. For example, an FMEA study may help identify ways to prevent failures, and it is often used to prioritize the many possible ways to improve reliability. We will list the tool under the type of decision primarily associated with the tool and note the other decisions the tool may inform.
In practice, major decisions may involve many subordinate questions to address to fully inform the decision maker. For example, a decision to make a critical component in-house instead of buying the component from a vendor may appear as a simple comparison type of decision. Yet, to make the comparison, the decision may rely on competing objectives, the ability to measure expected reliability performance, and the capability of each option to implement reliability improvements to prevent field failures.
Take the time to understand each key decision and any subelements, as it is common that more than one reliability method may be necessary to fully inform the decision makers for a single key decision.
Avoiding the occurrence of failures or downtime is a common desired method to improve the reliability of a product or system. To do so relies on using best practices gleaned from learning from past failures (ours and others). From selecting suitable materials for the intended use to establishing processes to quickly diagnose and resolve problems, prevention requires careful observations, capturing lessons and best practices, and rigorous engineering.
Here are a few example questions you may encounter that indicate a prevention-related decision:
- How can we avoid similar problems in the future?
- What is the best practice to design a reliable product?
- What amount of margin is sufficient for this feature?
- How can we anticipate failures before they occur?
- How can we minimize downtime?
There are two ways to make a comparison: by a measurement to an objective or by a measurement to one or more other measurements, e.g., the current reliability estimate compared to the goal or when choosing a vendor based on measured reliability performance.
Here are a few example questions you may encounter that indicate a comparison-related decision:
- Is the manufacturing process capable of achieving the objective?
- Should we accept this lot?
- Which design option provides the best solution?
- How can we optimize this process or design?
- Did the design change make a real improvement?
We rarely have the resources to accomplish all the potential tasks to create or maintain a product or system. Consequently, we often want to focus resources on the specific actions that have the most return for the effort. The ordering or ranking of potential actions may include the highest adverse risk, most common issue, most costly consequence, or some combination of factors.
Here are a few example questions you may encounter that indicate a priority-related decision:
- What improvement will provide the most benefit?
- Have we resolved all identified critical faults?
- What are customers complaining about this month?
The use of people, time, materials, etc. relies on making resource allocation decisions. It also suggests that everyone within the organization understands their roles and responsibility, the plan to achieve an objective, and the specifics to improve the organization’s capabilities and personal development.
Here are a few example questions you may encounter that indicate a resources-related decision:
- Who is going to conduct this analysis or run this test?
- When will the failure analysis be completed?
- Who makes the decision on vendor selection?
- What capital equipment do we need for the lab next year?
- How am I going to improve my skills?
From establishing a reliability vision for a brand to selecting candidate vendors for components an initial decision is “How reliable does this item need to be?” Reliability objectives encompass deciding on a product’s reliability position within the market, system availability, product reliability, supplier requirements, cost of ownership, and more, all of which reflect business and customer expectations. These objectives provide a touchstone for comparison to reliability measurements.
Here are a few example questions you may encounter that indicate an objective-related decision:
- What are the goals, metrics, and objectives?
- How do you break down the system goal to various subsystems and components and suppliers?
- What are the desired cost of ownership, life cycle costs, and warranty target?
The ability to measure reliability and related metrics provides essential information. The measurements provide a basis for trend analysis, comparisons, assessments, and priorities. The ability to make a measurement of reliability often relies on available information. Prior to the creation of the product, reliability estimates may rely on engineering judgment, whereas long after product shipments have begun, one can use field returns and detailed failure analysis to create reasonable reliability estimates.
Here are a few example questions you may encounter that indicate a measurement-related decision:
- Will this product or item survive in the intended use environment?
- How reliable is the current design?
- Are the vendor reliability claims valid?