Evaluating processes
Six Sigma – iSixSigma › Forums › Old Forums › General › Evaluating processes
- This topic has 22 replies, 13 voices, and was last updated 17 years, 1 month ago by
Advisor.
-
AuthorPosts
-
March 9, 2005 at 6:27 pm #38685
My company has been involved in Pareto analysis for quite some time now. It has been accomplished by individual production lines recording defects on an inspection sheet. At the end of every day the data from the sheets is entered into an access database which generates the pareto charts in ppm. We are wanting to change this practice as most generally pareto analysis is for the short term. Does anyone have any ideas as what we could do next and get away from the pareto altogether?
0March 9, 2005 at 10:35 pm #116162I’m puzzled…
1) Why are you using only pareto analysis to evaluate your processes? —there is such a plethora of tools that can additionally used…
2) Why would you want to discontinue using pareto analysis altogether? —it is a useful tool…0March 10, 2005 at 12:38 pm #116204Ed, I was hoping to find another tool suited to our needs. We have lowered our PPM (3.4 ppm or lower) so much that we feel the Pareto is no longer needed. It takes alot of time to keep the current system going. We have also went to a check/work/verify system. This means that every operator now has an inspection sheet (hence even more time consuming). We also have the impression that no one in the organization is using them any more becasue we have had this system for such a long time. What other tools would you suggest? Thanks for the reply.
0March 10, 2005 at 12:43 pm #116207Scott,
What nonsense – you say you are better than 3.4 ppm and no one is using the inspection sheet.
First of all, I don’t believe you are below 3.4 because if you were, data collection and analysis could be done on the back of a business card.
Second, if no one is using the inspection sheet, how do you really know your ppm?0March 10, 2005 at 12:49 pm #116209Stan, the inspection sheets are being used by the operators. The amount of sheets coming in is considerably more than what was turned in, in the past. I didn’t mean that they are not using the inspection sheets, I meant that no one is using the pareto charts anymore. And yes, most of our PPM’s are lower than that according to the data. Our PPM’s are figured by amount of defects found versus amount of parts shipped and it is done automatically once data input is complete.
0March 10, 2005 at 1:41 pm #116214If you are really less than 3.4 your data collection and paretos are no brainers. I could do it in 2 minutes on the back of an envelope.
What are your real issues?0March 10, 2005 at 1:45 pm #116216If you are really less than 3.4 your data collection and paretos are no brainers. I could do it in 2 minutes on the back of an envelope.
What are your real issues?0March 10, 2005 at 2:03 pm #116221I agree, they are no brainers. We have mutiple product lines spanning several different locations. I am refering to the time involved in getting the data into the system. If our ppm is that low, where are we benefitting from pareto’s? Especially since we feel that they are only being utilized by a select few. What other process performance tools could we implement and do away with the pareto analysis altogether in your opinion? Or should we keep? Thanks.
0March 10, 2005 at 4:56 pm #116231
Bob the Black BeltParticipant@Bob-the-Black-BeltInclude @Bob-the-Black-Belt in your post and this person will
be notified via email.Something doesn’t sound right about this subject…It doesn’t seem to make logical sense to me. Might I suggest that maybe your operators aren’t being truthful on these inspection sheets?
0March 10, 2005 at 4:57 pm #116232Is your definition of a defect so liberal that it is just “too darn easy” to meet these goals? Case in point, we worked one project where the goal was to reduce the turn-around-time in a batch processing operation. We had a relatively high average TAT (around 14 days) and an atrocious standard deviation with this process (around 8 days). So we set a goal of getting everything in and out in less than 7 days. After we modified the process we appeared to create a better than 6s process. That was great, but there still was no reason why we shouldnt be getting things in and out in at least 1 day (actually within ~1.5 hours).
Another idea. Are you sure that what you are getting an accurate picture or could people be gaming the system to appear like they are Hot S*** but the reality is they still arent meeting customer expectations?With me only being a novice and knowing what you have said so far, I can only guess (or Question) what is up with all that you are saying0March 10, 2005 at 6:02 pm #116239Let me try and break it down for everyone. We are a company that produces interior automotive parts. There are several production lines that are involved in a check/work/ verify system. Each operator checks the previous operators work, does his or her work, then verifies there work and passes it on to the next operator. Each operator has an inspection sheet where they can record any defects they find. It is a sheet that lists every possible defect and they just use a tick mark. Before this system there was an end of line inspector who had an inspection sheet. So you can see that the amount of inspection sheets coming in is far greater now. More time is now involved in getting this data in the system, which is very little but still time consuming between multiple plants doing there own data input. From all of this data input comes pareto charts, broken down by line and location, and is calculated by amount of defects found versus the amount of parts shipped. The ppms for each line is very low with some lines having no ppm. I realize that there could be a problem with the reporting of the data but that is not my concern. We are wanting to get away from pareto charts altogether for reasons I have already mentioned. We have beat process improvement to death on these individual lines from data through these paretos. But we still want to be able to monitor process performance. All I am wanting to know is what other tools would work given that you now know how we are currently doing it. Is man hour cost vs gain worth it at this point? Please let me know your thoughts.
0March 10, 2005 at 6:48 pm #116244Poka-yoke.
It will cut down the time people are devoting to the inspection process because you have automated it. If possible, that is…0March 10, 2005 at 7:24 pm #116246Ed, that is an excellent idea. But most of the process is manual, labor intensive. Maybe we are stuck with the pareto. How do you poke yoke a manual process that involves human interaction to get the job done? I don’t have any answers myself. Thanks again for your reply.
0March 21, 2005 at 1:53 pm #116600Scott,
Yes, there are many other approaches. We should not be writing perscriptions without diagnosing the problem. Here are some questions that may get the process off to a better start.
The actual ppm can and should be verified through process audits or other sampling. If your data collection is no good, you are wasting effort until you identify what the real problem is. This is the purpose of GR&R at the beginning.
Are the pareto rankings pretty constant for an area/work center? If so, then you can concentrate on preventing those defects. Mistake proofing is best, followed by increasing capability of the process.
Once we get past these issues we can move on. Mark0March 21, 2005 at 3:18 pm #116603Scott,
Congratulations first of all. I have never seen a manual process that delivered 3.4 ppm.
The other comment I would make is that if you are not concerned about the accuracy of the data being turned in by the operators, then the pareto is useless. Stop using it.
However, I suspect that the operators are telling you what you want to hear. It takes a pretty mature organization to tell management about the defects you see coming from one of your peers.
If you are truly as good as you say, then eliminate the operator paretos and focus on customer metrics.
Other interesting metrics would be:
Customer Feedback.
What is your customer complaint rate? How many parts are returned?
What is wrong with the returned parts?
Revenue:
Have you lost any business to a competitor? How much has your business grown with your < 3.4 ppm defect rate?
DEJ0March 21, 2005 at 7:15 pm #116618You say, “process improvement has been beated to death…” then there is little chance for further improvements. Also, you mentioned that the problems with data is of little concern.
I recommend stop using Paretos charts and any other charts and rely on the sites to report their defects on $100 bills.
JC0March 21, 2005 at 8:00 pm #116622I say data is of little concern at this point since process improvements have already been made. Keep in mind that we are dealing with defects such as thread tails for instance. Not that this defect is unimportamt but it certainly isn’t a safety concern defect either. I agree that the other sites need to be able to report something but does it have to be pareto analysis? Isn’t this tool most generally used for the short term??? Are there not any other tools out there that we can start using? That is all I am asking.
0March 21, 2005 at 8:34 pm #116625Scott,
I was making light of your comments. As someone mentioned previously it’s time to look at other measures.
I use Paretos to drilldown on contributing factors. In my case I look at First Call Resolution for contact centers. Once improvements have been made and are sustained for a period of time I use control charts to measure high level performance metrics such as, Call Times, Escalations, and VOC surveys. If a special cause is identified I return to the Pareto once again to pinpoint the contributing factors (eg. site, shift, call drivers). Once a chart loses it’s effectiveness or purpose it’s time to move on to other opportunities.
Before you drop the existing process make sure you identify the new metrics and methods for measuring them.
Good luck
JC0March 21, 2005 at 8:49 pm #116627
SigmaLayerMember@SigmaLayerInclude @SigmaLayer in your post and this person will
be notified via email.Great Detail, excellent posting — thanks JC
-SigmaLayer0March 29, 2005 at 5:07 am #116944Scott,
May I know what’s PPM meaning ? Is that a software program ? Now, I’m trying to analyse the quality problems by using pareto data.
Regards,
Yanuar
0March 29, 2005 at 5:33 am #116947
AnshikaParticipant@AnshikaInclude @Anshika in your post and this person will
be notified via email.Yanur,
It’s not a software PPM means part per million. This term is used to count defect rate in the process.
Regards,
Anshika0March 29, 2005 at 7:46 am #116948
indreshParticipant@indreshInclude @indresh in your post and this person will
be notified via email.hi anshika,
do forward your email id if possible, i think we can share our learning and as well learn more in the process
others most welcome to write in at [email protected]
rgds0March 29, 2005 at 11:32 am #116958
AdvisorParticipant@AdvisorInclude @Advisor in your post and this person will
be notified via email.Hi Indresh ,
If you are not aware of PPM then you should first study the basics first and then talk about sharing of your knowledge.
Regards
Advisor0 -
AuthorPosts
The forum ‘General’ is closed to new topics and replies.