How can I cut 14000 NCRs (Nonconformance Reports) by 75%?

[FLEETWOOD]

I have been reading the threads on this forum for several months, and I think this is all good stuff. We all suffer from the same pain, and it’s good to know that you are not alone in the struggle for quality. I have been involved with quality for over 20 years in my organization, and we have had many managers over this period. Every manager that walked through the door, having his own view of quality. The last guy got us into this mess that I’m trying to clean up.

This company has existed for over 60 years, and being approved to the MIL Q and AQAP’s and the Canadian Z299 series and today ISO 9001, we have a fairly matured QMS. However, to suit demands from other players, over the years, we have tweaked the QMS so badly, that it’s gone out of control. Hence, this brings me to the 14000 NCRs raised in our last fiscal year. My goal is to reduce this by 75%.

We raise NCRs within all elements of operations, production, procurement, test and for customer service issues. Basically the flavor of the month was to make one form that did every thing. This form is also used to generate unplanned hours and demands. It’s a cure all miracle form. The thinking behind this was that “ we need to track trends”. However, we (they) realized, drafting up an NCR to flag that a 50 cents fuse is blown and we need to replace it, is probably not very cost effective approach. And reviewing 14000 ncrs to track trends with limited resources is not easy.

I would like to use the NCR systems that I have existing and work with it. I can’t change the culture. That’s one thing that I learnt over the years. However, I would like to know, if any other people have the same issues. Any comments would be great. I’m so deep into it that really, I need a view from the outside. Are we loosing control?

Regards.

 

[Wes Bucey]

I have been reading the threads on this forum for several months, and I think this is all good stuff. We all suffer from the same pain, and it’s good to know that you are not alone in the struggle for quality. I have been involved with quality for over 20 years in my organization, and we have had many managers over this period. Every manager that walked through the door, having his own view of quality. The last guy got us into this mess that I’m trying to clean up.

This company has existed for over 60 years, and being approved to the MIL Q and AQAP’s and the Canadian Z299 series and today ISO 9001, we have a fairly matured QMS. However, to suit demands from other players, over the years, we have tweaked the QMS so badly, that it’s gone out of control. Hence, this brings me to the 14000 NCRs raised in our last fiscal year. My goal is to reduce this by 75%.

We raise NCRs within all elements of operations, production, procurement, test and for customer service issues. Basically the flavor of the month was to make one form that did every thing. This form is also used to generate unplanned hours and demands. It’s a cure all miracle form. The thinking behind this was that “ we need to track trends”. However, we (they) realized, drafting up an NCR to flag that a 50 cents fuse is blown and we need to replace it, is probably not very cost effective approach. And reviewing 14000 ncrs to track trends with limited resources is not easy.

I would like to use the NCR systems that I have existing and work with it. I can’t change the culture. That’s one thing that I learnt over the years. However, I would like to know, if any other people have the same issues. Any comments would be great. I’m so deep into it that really, I need a view from the outside. Are we loosing control?

Regards.

Welcome to the Cove, Gerry!
We're always happy to convert a lurker to a registered Cover.

Forgive me for drooling, but I would LOVE to get my hooks into your situation. Check my thread (broken link removed)

I sympathize with your situation. Other organizations have also found themselves 'paralyzed' with bureaucratic red tape - certainly, this is just another version of $25 purchase orders ($25 in overhead to produce a purchase order) to buy a 50 cent pack of index cards.

I doubt they gave you much authority to go with your task. You need to scout around for someone with enough 'juice' to authorize some serious process changes once you create your wish list.

Often, the best tack is to do a cost accounting of the soft costs of opening, investigating, and closing a NCR with the current method, compared against the dollar value of the process or product affected.

Typically, what happens is the "smart" organization empowers folks to make cost estimates and short-circuit (evaluate and close) the NCR process for items below a "trigger" dollar value, aggregating the results in a weekly or quarterly report for trending/tracking purposes. If similar items keep popping up, the trends will show opportunities for preventive action. If there are no trends, at least there will be a cost reduction in the impact of the NCR by eliminating a lot of soft cost formerly incurred in having the NCR travel through a bureaucratic maze.

Keep us informed on what you are able to do.

 

[Claes Gefvenberg]

Welcome to the Cove, Fleetwood

We all suffer from the same pain, and it’s good to know that you are not alone in the struggle for quality.

Indeed. This is a very good network to be part of, and once in a while we all need to talk things over with someone in the same situation.

this brings me to the 14000 NCRs raised in our last fiscal year. My goal is to reduce this by 75%.

14000 of them? :mg: I definitely see your point.

And reviewing 14000 ncrs to track trends with limited resources is not easy.

Um... I see that you have a knack for understatements This is an interesting point: What resources do you have at your disposal?

I would like to use the NCR systems that I have existing and work with it. I can’t change the culture. That’s one thing that I learnt over the years.

Yes... Well, you can probably bend it a little bit but a complete change takes a lot of effort and support from the right people, so I'll take your word for it.

I’m so deep into it that really, I need a view from the outside. Are we loosing control?

Loosing control? Judging from your tale you may already have lost it a bit. I guess the deciding factor would be the extent to which you are able to deal with this mountain of NCR's. Anyway, even if your grip is loose, you are obviously trying to claw yourself back in control.

Good first post, btw.

/Claes

 

[D.Scott]

Hi Gerry, welcome to the Cove.

I don't think you've lost control. If anything you are trying to over control. It would seem to me that over the years you have lost sight of the objectives in the NCR system. My first suggestion would be to analyze what the 14000 NCRs represent. Obviously some would be a higher priority than others and I would Pareto them to determine just what sort of NCRs you are dealing with. You made the obvious comment that the cost effectiveness of an NCR for a blown fuse is questionable and that is exactly where I would start. The idea of tracking trends is extremely useful but with 14000 NCRs, it is unlikely that trends are really being tracked. Find out, by analysis, what "categorys" your NCRs fall into. Once this is done, involve the departments/functions they relate to an determine what is actually being done with the NCRs and the data provided by them. You will probably find out you are creating NCRs that "die on the tree". Some department/functions may use the data but in such a way to leave them outside the current NCR system - maybe a simple reporting format will do. Once you know what you have, think about what you really want to accomplish.

14000 seems like a lot but remember, Rome wasn't built in a day. You have to approach this one small step at a time. Explore why you are collecting the data before you try fixing all the problems. I have no doubt a 75% reduction is a reasonable goal but don't try to achieve it overnight.

Good luck.

Dave

 

[Jim Wynne]

Often, the best tack is to do a cost accounting of the soft costs of opening, investigating, and closing a NCR with the current method,
compared against the dollar value of the process or product affected.

Wes has nailed it here. I have personal experience in this situation, only the number of NCRs wasn't as great. I inherited a situation once where the number was ~2000 per year, which was way too many. Most of them had to do with receiving inspection issues related to poor engineering specifications, which was evidenced by the fact that nearly 75% of the NCRs received use-as-is dispositions. The problem was that there was no followup; drawings never got changed to reflect reality (the UAI dispositions were almost always appropriate) so many of the same issues got written up repeatedly. The rule was, if the material doesn't meet specifications, an NCR gets generated--no exceptions.

I decided, as Wes has suggested, that I needed to speak the language of upper management and translate the NCRs into dollars and cents. Long story short, I enlisted the help of a guy in accounting and developed a ballpark figure for the cost of documenting and dispositioning. I then took a representative stack of UAI-dispositioned NCRs--300 or so--and bound them together with a few rubber bands and brought the stack to a staff meeting which was presided over by the operations VP. I put the stack in the middle of the conference room table and waited for someone to ask me what it was. Fortunately, it was the VP who asked the question. "That's $75,000," I replied, "and there's more where that came from." The VP was very attentive after that.

 

[Rachel]

From someone just popping back in...

Part of the issue here - when you climb up to that high of a number, how often are the identified issues mere repeats of ones identified months before?

We are in a similar situation right now with our CAR database - there are 25 open issues (the highest amount since I took over the system) and it's stressing me out. Several of these relate to late shipments of customer orders. Since *really* digging into it, we have realized that one of our injection moulding (IM) machines is tapped - it is dedicated to moulding parts for internal use. We have too many parts that we mould and use internally for subassemblies, and we simply can't keep up with make-to-ship demand - the time required for mould changeovers, the number of changeovers required in a make-to-ship scheme - we just don't have enough machine hours to make production. So, we increased the mins on a lot of these internally-used parts - in this way, we are working on a system of replenishment. Goes against lean, but we are shooting ourselves in the foot with lean on this approach...without carrying stock, we are setting ourselves up for trouble. For one product in particular, there are *seven* separate IM parts that are required...without carrying enough in stock, there's bound to be delays.

Anyway, what does this have to do with you?...in coming to this conclusion on one CAR, we outlined a corrective action of increasing the mins on the parts for this one product in particular. Our preventive action was to look at the rest of our item codes and come up with a list of similar IM items affecting other products.

Guess what? Of the items that were listed - the products in which they are used were the subject of three additional CARs. Bam - there's our corrective actions!

Isolate and fix the root cause of one of these, and - guaranteed - you will eliminate several others. If I were you, I would start there - look for repetitive or recurring issues.

As for the fifty-cent fuse - the CAR database should not be used as a to-do list. If this is the way that people have been using it, well, maybe some re-training is in order. I would think a maintenance tag would be far cheaper.

Hope this helps,
-R.

 

[FLEETWOOD]

Thank you so much for your support.

Our company is currently in the cost cutting phase, so the approach that I am looking at is to hit them were it hurts....cost. I started looking at a burden cost for raising NCR. For example - open database, write NCR, send, person gets NCR, reviews it, sent back to QA and so on. We are not looking at contents at this point, just a burden rate to handle NCRs. Can anybody contribute to this? Has anybody done this exercise before?

 

[Rachel]

Thank you so much for your support.

Our company is currently in the cost cutting phase, so the approach that I am looking at is to hit them were it hurts....cost. I started looking at a burden cost for raising NCR. For example - open database, write NCR, send, person gets NCR, reviews it, sent back to QA and so on. We are not looking at contents at this point, just a burden rate to handle NCRs. Can anybody contribute to this? Has anybody done this exercise before?

Sounds like you're sort-of talking about cost of quality - which is a very difficult number to ascertain. In terms of the cost of initiating a nonconformance, I think it's a good idea - where you want to be careful is when you start moving on to the "cost of the CAR" in its entirety - the cost of the fix, and the effort hours it entails. When we open a CAR, an administrative fee is automatically incurred...what that fee is for you depends on what your calculations tell you. When someone initiates a CAR, there's an automatic series of steps that I have to follow: assign an owner, distribute the report, organize a meeting, and sit on people for answers. For us, let's say that is, oh, $250. Now, that's not just my salary (I wish!) - it's also the money that the company has lost because I'm focused on other efforts...that's the cost of the time that it takes everyone to get up to speed on the issue, and the money that we've lost as a resut of *their* efforts being diverted.

I don't know that it's so easy to put a hard number on that - but it can be useful, at least, as a frame of reference. When we calculate COQ, we have set costs for effort hours incurred, and those costs vary as you climb up the ladder (i.e., an exec's time is worth more than a lineworker's time). Though it's not perfect, we can at least apply those costs consistently in order to determine where our *relative* expenses are in the system. Material stocking costs (i.e., how much is it costing us to hold the NCP in our warehouse?) are done similarly - it's always the same percentage of the material cost. Direct costs (i.e., additional freight, the cost of a new machine) are entered as they are.

Hope this helps,
-R.

 

[FLEETWOOD]

step one - burden rate

We decided to gather the cost of raising a NCR, some of you call it the "soft" cost, we call it the "burden rate". This is the cost of just writing and analyzing the NCR, the touch labour, if you wish to call it that. The cost was determined to be $150/NCR.

So we are ready to sit down and review our process based on this figure. We plan not to raise NCR on issues that are estimated below this $$ value. However, we still would like to gather the problems so that we could do some trend analysis? We had in mind a type of log book, however, in a large place, this would end up getting lost or not used. Is there any other creative and effective way of doing this? Any ideas?

gp

 

[jmp4429]

We decided to gather the cost of raising a NCR, some of you call it the "soft" cost, we call it the "burden rate". This is the cost of just writing and analyzing the NCR, the touch labour, if you wish to call it that. The cost was determined to be $150/NCR.

Ack! That's $2.1M per year on NCR's!!!!! If you can, would you mind telling us how many of your 14,000 would be estimated below that value?

Maybe I'm way off base here, but I think to track the NCR's with a dollar value below the threshold, you need to separate them by type of NCR. For example, a blown fuse should be tracked by maintenance (which could help with TPM) while a misassembled part should be tracked by engineering/quality. Another way to do it (bearing in mind I have no idea about your particular operation) would be to track issues by workcell or department, and have the engineer for that area track it ie:

Product A
4/15/05 - Fuse blew on crimping machine, caused 15 min downtime.
4/18/05 - Operator used wrong grommet on assembly.
4/29/05 - Fuse blew on crimping machine, caused 15 min downtime.
5/2/05 - Operator used wrong size screw on assembly
5/17/05 - Operator used wrong grommet on assembly.
5/22/05 - Fuse blew on crimping machine....

Then you could see trends better, and track problems in a way that makes them useful. "Hey, something more than the fuse is wrong with the crimping machine, why does it keep blowing fuses. And we really need some better errorproofing to prevent using the wrong grommet."

Does this make sense with your current situation?