Confidence from Understanding

Abstract

Chris and Fred discuss what this term ‘confidence’ means … especially when it comes to statistics (and reliability)

Key Points

Join Chris and Fred as they discuss what ‘confidence’ means … especially when we do things like statistical analysis to help us understand the reliability characteristics of a system or product. This comes to us from a question (which is like a lot of other questions) we get about how much testing do we need to do to <guarantee something never happens>? … how many samples do I need to make sure there are no quality control issues? … how long do we need to test to demonstrate reliability?

Topics include:

Confidence is a measure of YOU … not the SYSTEM. Never forget this.
(#1) Confidence can come from STATISTICs. But this is expensive, challenging to get right, and prone to manipulation (you know … keep testing until it passes). The organizations and people who tend to lean on confidence from statistics usually don’t understand how their product, system or service fails. Which means they don’t (didn’t) know how to design out failure. Which is a problem.
(#2) Confidence can come from UNDERSTANDING. This is much better. It means you know what the weak points are, what you need to do to design them out of the system, and model the VITAL FEW that are remaining so they know (with supreme confidence) that their system IS reliable … and just how reliable it is. This means they don’t really need to do statistical testing (… which is something that organizations who rely on statistics simply can’t understand).
Let’s look at an example … quality control issues from a supplier. What happens if you are building a system using components from suppliers (like most producers do). How can you get ‘confidence’ that these parts are of high quality? There are a few options …
1. Establishing a SYMBIOTIC RELATIONSHIP with your suppliers based on transparency and trust where information is shared and failure root causes are fed back into the next generation of componetns.
2. BUILD COMPONENTS YOURSELF so you can control the quality yourself
3. LEARN (through things like FMEAs and FTAs)/Understand failure and provide guidance to suppliers 4. Inspection based on knowing what will fail 5. Screening (HASS) or ‘robust’ testing to accelerate the unveiling of failure root causes and remove ‘problems’ … but you still need to know what the likely weak points are.

Enjoy an episode of Speaking of Reliability. Where you can join friends as they discuss reliability topics. Join us as we discuss topics ranging from design for reliability techniques to field data analysis approaches.