Fred Schenkelberg, a reliability expert, discusses the challenges of conducting reliability tests with small sample sizes. He emphasizes the importance of focusing on critical information and making informed decisions. Schenkelberg shares strategies like utilizing alternative methods such as simulations and expert judgment. The conversation also highlights the impact of environmental factors on testing outcomes and advocates for cross-departmental collaboration to maximize insights. Listeners learn about innovative analytical methods like Bayesian analysis and the bootstrap method.
00:00
forum Ask episode
web_stories AI Snips
view_agenda Chapters
menu_book Books
auto_awesome Transcript
info_circle Episode notes
question_answer ANECDOTE
Scarce Samples at HP
Fred Schenkelberg recalls requesting 100 inkjet printer samples while at HP.
These were costly, and other teams also needed them, highlighting the scarcity of samples.
volunteer_activism ADVICE
Prioritize Samples for Decisions
Focus on the purpose of the sample and the decisions it informs.
Prioritize getting samples for critical decisions rather than routine tests.
volunteer_activism ADVICE
Justify Sample Requests
Tie sample requests to important information needs and decision-making.
Justify the number of samples based on the value of the information they provide.
Get the Snipd Podcast app to discover more snips from this episode
It's rare to have sufficient samples for all the desired reliability work. It's common to have very few samples. So, how does one manage to learn what is necessary to learn from the few samples? How can you maximize the value of reliability work in the scant sample situation?
Let's discuss a few approaches and a process that enable you and your team first to have the right amount of samples. And what can be done when there are too few samples available? The trick is to start with what future decisions need better information and not start with how many samples you have.
There are three strategies to deal with small sample sizes. First, focus on the critical information only samples can provide and how valuable that information is to the team. Second, use other methods to create the needed information without using samples. Third, organize the various testing and measurements of the sample allocated to reliability testing along with samples or units assigned to other sections of the team to extract as much information as possible.
This Accendo Reliability webinar was originally broadcast on 9 January 2024.