CAR & 5Whys

Mike S.

Happy to be Alive
Trusted Information Resource
I was in a hurry, but when I read the OP I considered "the painter" as an organization, not a person. And yes, who pays "the painter" matters here.
 

optomist1

A Sea of Statistics
Super Moderator
Gun Lake, how does your firm track discrepancies, defects, etc.? Do you have alrger framework within which the larger problems can be tracked; i.e. , near term, permanent CA and closeout, added to corporate lessons learned?
 

PCH52

Registered
As 5 why is something fairly new, or unused by me and my small company, i have asked around about its use.
I have been told that its acceptable to not necessarily use all 5 steps to get to the root cause. It maybe that 3 or 4 will do it.
Any suggestions please?
 

Ninja

Looking for Reality
Trusted Information Resource
5 why is pretty old.... 1970's

Functionally...what matters is that you accurately discern the root cause.
Depending on who you're talking to, some will insist on exactly 5.
Sometimes things are so obvious that 2 is more than enough...sometimes 7 is a better choice.
Functionally...what matters is that you accurately discern the root cause... 5-why is a tool, use it until the job is done.
 

Tagin

Trusted Information Resource
As 5 why is something fairly new, or unused by me and my small company, i have asked around about its use.
I have been told that its acceptable to not necessarily use all 5 steps to get to the root cause. It maybe that 3 or 4 will do it.
Any suggestions please?

Agreed. There is nothing special about 5, it is just a guideline. Sometimes, it might less than 5, sometimes it might require more than 5. So, document your process to flexible about the number of iterations.
 

Miner

Forum Moderator
Leader
Admin
As 5 why is something fairly new, or unused by me and my small company, i have asked around about its use.
I have been told that its acceptable to not necessarily use all 5 steps to get to the root cause. It maybe that 3 or 4 will do it.
Any suggestions please?
Typically you will keep asking Why until the answer leaves your span of control. For example, if your problem was you were late for work because you overslept, because your alarm did not go off because you lost power because a substation exploded because it was hit by lightning because there was an electrical storm. At some point along those Whys, you lose the ability to make a corrective action. That is when you stop. You cannot take action on anything from substation backward, but you can purchase an alarm with battery backup (or use a smartphone).
 

John Broomfield

Leader
Super Moderator
The scope of your management system may limit your whys.

Some orgs are able to influence the systems of their customers and suppliers.

But don’t stop just because top management would have to take action.
 

Jim Wynne

Leader
Admin
An artist, asked what was the most difficult thing about painting, answered "Knowing when to stop." It's possible to engage in regression that goes back to the Big Bang, but what needs to be done is to find the cause and eliminate or neutralize it. Two steps. How much investigation finding the cause entails will vary, and sometimes too much is done.
 

Bev D

Heretical Statistician
Leader
Super Moderator
Thanks Jim! as is too often the are the corrective action process suffers from investigations at 3 extremes:
1. Blame the operator
2. Do nothing because it is too infrequent, too hard, impossible, not my fault
3. Go too far

One thing that seems to be missing (besides a true commitment to improving quality) is that there is more than 1 type of action we can take (and even the standard is guilty of this).
feedback control (e.g. SPC) where the output can be seen to be changing and we know what factors to change to bring the output back into control. Think about tool wear...
feed forward control: measure a factor that can’t be easily controlled and adjust another factor to compensate for the first one.
Inspection/screening: can be a bit expensive but if a supplier is still sending you ‘off the shelf’ or ‘catalog’ material that will cause failures yet is within the material specifications you can inspect and screen out the material that doesn’t work In our application.
robust against the cause: there are some causes that cannot be controlled but we can alter the product or process against that variation. Thing ‘rust proofing’...we can’t control the rain cut we can control the material that will rust. Error proofing falls into this category. People - all people - will commit errors. We cannot stop that no matter how hard we train , yell or discipline. But we can error proof our processes.
reduce the variation in the causal factor: this is waht most people think of in correction actions
redesign the process or product to truly remove the causal factor. Usually a last resort as it can be time consuming and expensive but it may be necessary in some cases.
 

Miner

Forum Moderator
Leader
Admin
One thing that seems to be missing (besides a true commitment to improving quality) is that there is more than 1 type of action we can take (and even the standard is guilty of this).
feedback control (e.g. SPC) where the output can be seen to be changing and we know what factors to change to bring the output back into control. Think about tool wear...
feed forward control: measure a factor that can’t be easily controlled and adjust another factor to compensate for the first one.
Inspection/screening: can be a bit expensive but if a supplier is still sending you ‘off the shelf’ or ‘catalog’ material that will cause failures yet is within the material specifications you can inspect and screen out the material that doesn’t work In our application.
robust against the cause: there are some causes that cannot be controlled but we can alter the product or process against that variation. Thing ‘rust proofing’...we can’t control the rain cut we can control the material that will rust. Error proofing falls into this category. People - all people - will commit errors. We cannot stop that no matter how hard we train , yell or discipline. But we can error proof our processes.
reduce the variation in the causal factor: this is waht most people think of in correction actions
redesign the process or product to truly remove the causal factor. Usually a last resort as it can be time consuming and expensive but it may be necessary in some cases.
Right out of Statistical Engineering: An Algorithm for Reducing Variation in Manufacturing Processes. An excellent book.
 
Top Bottom