QMS Improvement - Where To Start?

[Golfman25]

TM is smoking stuff that hasn't been analyzed yet!

Scenario..........I'm a contracted truck driver, you get all your stuff made, inspected, loaded and I'm out the gate a day early.....On the way to your customer I take a phone call from my employers saying "That increase you asked for, it ain't happening"............Regardless of where I am, I park the truck, get on a Greyhound, and go home, not telling anyone where it is. A month after your product is due for delivery the truck is found. Was the OTD metric met? How good are your planning, and metrics now?

So what's your point? One instance doesn't establish any type of trend.

 

[Joe Cruse]

I also am with other responses on here about collecting paperwork for the sake of collections. Don't start asking from data from EVERY point in your processes, all at once. If you know the processes fairly well by now (if you don't, get at it), it shouldn't take too long to at least get a hint of where at least a few bottle-necks are located. Start getting process measurables/data there and do your own trending to see if those spots ARE part of the problem. Keep working your way through the system until you have some good, measurable proof that certain areas are the biggest contributors to killing your company's OTD. Then come up with potential Action items, including maybe some suggestions from people in those problem areas, then go to TM.

Also have to agree on language used with shop floor AND TM. Be careful not to throw "quality buzzwords" around when going through all of this.

BTW, now that someone in TM actually has been shown they are not complying with customer requirements in delivery of purchased materials (and the whole company is also aware of this NC), how is this being handled, keeping in mind your next surveillance audit for ISO 9001:2015? To conform to the standard, you can't just say "well, we'll all try harder" and document that as a proper response. As a customer or a registrar auditor, I would ding you.

 

[Joe Cruse]

Yeah, let's throw TM under the buss. The place is doomed. Maybe they do need to "try harder."

Look, I have been there and done this. We had less than 75% OTD to our largest customer. It was a complete cluster blank. They where looking to get rid of us. What OP needs is a structured approach to solving the issues, one bite at a time. They don't need "lean training." They need a little common sense and a unified goal. The goal has been set -- improve OTD. Now get into the details and go do it.

It is actually pretty simple. Track the shipments every day. What went out and was it on time or late. If it was late, determine why. Possible answers: capacity issues, not enough lead time, machine down, incoming material was late, tool needed repair, sat on someone's desk too long (i.e.; disorganized office), etc. Do it for a month, then pick the top 2 or 3 issues and make an improvement. Rinse and repeat. If you go to TM and say "the top reason we are late on jobs is this, and here are some solutions" they will work with you. If they don't, then you know.

All I'm saying is, from TM over what looks to be a serious problem (it's costing them business and revenue from at least one customer and TM appears to be concerned at that), all our OP has been given for correction is "try harder". Okay, in what way? That is an extremely general directive on correcting the situation, and OP is giving us the idea that TM doesn't appear to want to deep dive into details, as you have said.

I agree completely with you on the approach to solve this issue. It really is pretty simple, but it is also very hard if the TM attitude that the OP is conveying is spot on.

 

[Kurt Smith]

I am creating the new "On Time" metrics by comparing the due dates to the completion dates on the completed work travelers. With our standard expected delivery time subtracted. It isn't perfect, but its clear our problem isn't currently a failure of the shipping services, but us not finishing the parts in time. I may push to rename the metric before our next audit if needed.

Currently the dashboard can sort by time range and customer. I am also recording the date work started. Ideally, I will be able to sort late orders in a way that shows how many were due to production delays and how many were logistical issues that led to the work not having enough lead time to begin with.

From there I can focus in on the sources of delay. If most of the late work was started with plenty of lead time but wasn't completed in time, we can look into why. What are the bottlenecks. Were there too many jobs that needed the same workers at the same time? Was there a machine down? Were people on vacation?

If the late work was done quickly but had insufficient lead time we can look into scheduling and material ordering. Or a dozen other options.

I understand the fear of overburdening everyone with too much data collection. But it's almost as bad to wander around blindly.
Avoiding buzzwords is good advice, but I still need some way to communicate that the company has reached a point they can't keep winging their way through problems. That they will need some way to control the flow of work and gage if new orders stand a chance of completing in time before committing to the customers.

 

[Golfman25]

I am creating the new "On Time" metrics by comparing the due dates to the completion dates on the completed work travelers. With our standard expected delivery time subtracted. It isn't perfect, but its clear our problem isn't currently a failure of the shipping services, but us not finishing the parts in time. I may push to rename the metric before our next audit if needed.

Currently the dashboard can sort by time range and customer. I am also recording the date work started. Ideally, I will be able to sort late orders in a way that shows how many were due to production delays and how many were logistical issues that led to the work not having enough lead time to begin with.

From there I can focus in on the sources of delay. If most of the late work was started with plenty of lead time but wasn't completed in time, we can look into why. What are the bottlenecks. Were there too many jobs that needed the same workers at the same time? Was there a machine down? Were people on vacation?

If the late work was done quickly but had insufficient lead time we can look into scheduling and material ordering. Or a dozen other options.

I understand the fear of overburdening everyone with too much data collection. But it's almost as bad to wander around blindly.
Avoiding buzzwords is good advice, but I still need some way to communicate that the company has reached a point they can't keep winging their way through problems. That they will need some way to control the flow of work and gage if new orders stand a chance of completing in time before committing to the customers.

You're a job shop. 90% of your business is "winging it."

 

[Kurt Smith]

You're a job shop. 90% of your business is "winging it."

Yeah, but the goal isn't to always be a job shop. TM wants to grow and get larger automotive contracts...

 

[Bev D]

wishin, hoping and praying is a great song. It’s a lousy business strategy.

And you can implement TPS in a job shop. I did it. Twice.

And I’ve found that it isn’t so much on time completion (delivery minus the logistics between you and your customer) is mostly effected by cycle time. So it’s good that you are recording the start date as well. The real time you need will be hidden in the aggregate data: idle or wait time. Do you get start or completion times for each operation on the traveler? I also suggest categorizing build by complexity - no more than 3 categories probably - simple, moderate and complex should do it…

 

[Kurt Smith]

I was imagining something similar. Assigning every P/N a complexity level 1-3. Incorporate that into a production schedule so we can approximate the workload.

We do also record each operation on the traveler. Operators already record the time/date they complete the task. It seemed too daunting to include as I am working through the backlog of completed jobs. But if the problem turns out to be mostly production it could be well worth it to look into wait times between workstations.

And I don't think this place needs a full conversion to TPS or some in depth Six Sigma implementation. I'm not expecting to record our DPMO or anything like that. But some red flags that fly when we are approaching capacity would go a long way. A standing rule to alert management when material starts to run low. Anything to give us more reaction time when problems come up.

 

[Bev D]

And I don't think this place needs a full conversion to TPS or some in depth Six Sigma implementation. I'm not expecting to record our DPMO or anything like that.

Maybe full TPS (the manufacturing engineering stuff not the human soft stuff) is possible in the future if you really want to grow.
That DPMO stuff is voodoo crap. And unless you have a lot of defects you don’t need robust problem solving or FMEA. You may need some MSA tho. But again you have an uphill battle if your TM stays stupid

 

[Golfman25]

I was imagining something similar. Assigning every P/N a complexity level 1-3. Incorporate that into a production schedule so we can approximate the workload.

We do also record each operation on the traveler. Operators already record the time/date they complete the task. It seemed too daunting to include as I am working through the backlog of completed jobs. But if the problem turns out to be mostly production it could be well worth it to look into wait times between workstations.

And I don't think this place needs a full conversion to TPS or some in depth Six Sigma implementation. I'm not expecting to record our DPMO or anything like that. But some red flags that fly when we are approaching capacity would go a long way. A standing rule to alert management when material starts to run low. Anything to give us more reaction time when problems come up.

The problem is how much of your work is standard repeat items vs. 1 off projects? Repeats will be easier to standardize. 1 offs are killers. Remember, the hardest shop to manage is the job shop.

Your problems will be mostly production related -- your highest reason for lates will be capacity constraints. It's the nature of the beast. You can only run so much thru each work center. Your 1, 2, 3 complexity level will work, but you'll need a good idea of how many 3s you can do per week, month, etc. Keep it simple.

And please, don't proceed with the idea the TM is "stupid." Those guys probably put their house on the line to get the company going. They may be ignorant and/or resistant to change, or even worn out, but they ain't stupid. If you understand where they are coming from you'll have a lot more success. Good luck.