I’m now five months into my journey of transforming a high-production shop into a kaizen facility. My long-term goal is to accomplish a complete, “real” transformation – not because it has now become the sexy thing to do in a lot of industries, but because it’s the smartest thing to do.
Most people, including people I deal with on a daily basis, don’t even truly know what being a lean operation means, let alone a kaizen organization. I could divide the shop up with tape on the floor and throw up some visuals on the wall and fool a lot of people, but why bother? It’s not for show, it’s for results.
Five months into this journey, I’m nowhere close to where I thought I would be. I was ignorant to how long this transformation was going to take.
In my head, I thought it would take about two months to get the buy-in from the team and begin to establish flow. In the third month, I would start introducing standard work and the very beginnings of problem solving. By month six, we would be “rockin’” – by no means even close to perfection, but a 180-degree turn from where we started. At the very least, I thought we would be getting vehicles out the door on time and with quality, making our customers happy. Let me be the first to admit, however, that I didn’t get us there.
We’re basically back to the beginning and fighting non-stop fires in order to come close to on-time, quality work. What I’m not used to is the non-stop amount of work literally blowing through the shop doors. It doesn’t scare me; it intrigues me. I get asked all the time, how can you be so calm? It’s because my mind is going a mile a minute on how to fix all this stuff, and there’s no time for panic. Also, there’s no downtime and no time for training.
For example, I have two full-time repair plan techs, with one mediocre writer in between them and eight to 14 vehicles a day scheduled for repair plans. That doesn’t count the unscheduled drops that typically average around 10 vehicles on any given day. So, to save the math, two techs are expected to do from eight to 24 quality 100-percent teardowns a day. There are rules in place and quality standards to follow, but when you as a tech or I as a manager look at 24 repair plans in the eye, what happens?
Firefighting and the creation of a countermeasure road for stabilization. Unless you have a team of people and plan on doing a kaizen event, you cannot pick this point to do your training. The repair plan is the most important area of the process; if you screw up here, you send a bomb downstream that will pick the worst possible place to detonate.
This is one of my main focuses right now. By the time this article comes out, I’ll have a different repair plan team in place, one that I am 100 percent sure will buy in – if not for the right reason, then to please the boss. At this point in time, I’m OK with that because that light will eventually go off in their heads. Quality standards are already in place, and I’m now working on standard work for the repair plan team. They will be my first guinea pigs!
I’m fully aware that the repair plan overload would be conceived as a scheduling problem, which I don’t have my hands into yet. However, even if I did, I couldn’t fire off the way to immediately fix it. It would take a lot of work to figure out how to manage this shop’s workload.
I was taught that you scheduled to your throughout goal/demand, which had to be the correct mix of work and correct number of vehicles, too. You didn’t want to schedule eight heavy jobs, nor did you want eight speed jobs coming into your system in one day.
You had two ways of dealing with an unscheduled drop: you either blocked capacity into your schedule to allow these jobs, or they went into the next available schedule slot that could handle it, which may have been two or three days later. My current shop adds a new twist to this. There is capacity built in for unscheduled drops daily, and Fridays are shut off completely to take care of the rest of the unscheduled drops. However, that still isn’t enough. At any given time, we’re scheduled out at least six weeks and get an average of 10 unscheduled drops a day. Where do you put those, seven weeks out? Do you move out the scheduled customers who have already waited six weeks to accommodate the non-drives? I don’t think that would go over well with the owner or insurance partner. This all leads to incredible waste, inflated work-in-process (WIP) numbers and days where you need to deliver 19 cars. Basically, it’s insanity. Fixing it is not going to be easy, but if I don’t, the downstream operations will always feel the pain.
I explained in my first article in the December 2013 issue of BodyShop Business that this was my first shot at implementing lean. I’m not going to feed you a bunch of BS and tell you that I did A, B and C, and we’re now the greatest shop out there. I’ve been keeping a daily journal on what I do, what I tried, what failed and what worked to put into an informational article that hopefully will allow you to avoid my failures and pick up on some of the things I did right.
The first and best thing I can possibly tell you is to make it your own. Don’t try to imitate something you’ve seen, read or have already done elsewhere. If you do that, chances are you’ll fail badly. You can take bits and pieces, but ultimately your best asset is going to be your knowledge.
Take the knowledge you already have but make it work in your new surroundings; it’s going to have to be a hybrid. You cannot duplicate someone else’s model – unless you’re an MSO, in which case your model will work across all of your locations. But taking Shop A’s model and trying to adopt it at Shop B isn’t going to work. There are way too many variables that keep this from being successful.
That was my first failure: trying to model my shop too closely to the shop from which I learned everything. My thinking was, it worked there and I have a very good understanding of it, so why can’t I just put it into play over here with a few tweaks I always wanted to make anyway? Looking back at it now, I realize I was way wrong. The culture was wrong, the building was different, the volume was three times that of the shop I came from and I was all alone with no kaizen team to work with. What the heck was I thinking?
And the failures didn’t end there; this was just the most blatant one. Of all the things I implemented in the first article, only two stuck, and one of those two already had to be revamped.
The heijunka board worked like it should have and provided the important information we needed. But it failed. People didn’t understand the importance of it, and the shop layout didn’t allow it to be as visible as it should have been to everybody. The design didn’t yell out where the trouble was, because they didn’t have the knowledge of where to look. Plus, I was the only one who could update it. We’re currently working on a new, simpler version that should do the job just fine; I’ll let you know if it works out!
Repair Plan Priority Rack
The repair plan priority rack withstood the test of time, and it’s still an integral tool used throughout the day.
The production start rack failed and turned into a storage bin where all the files are kept but not where they belong. It was designed to let everyone know what the next logical job to start was by indicating part status and target delivery dates. Without me staying on top of it, it fell by the wayside. Once again, the team lacked an understanding of how important this system was.
The quality standards are still posted on the wall in each area, but haven’t been looked at since. The quality check-off has been completely abandoned, and papers were showing up at the front desk with one signature on them versus the six they should have had. Ownership feels it’s a waste of paper and ink to continue right now. The quality control sheet is still in play, but doesn’t work as it should. How do I know? Vehicles are still making it to detail and getting bounced back for quality issues.
Vehicle Check-In Sheet
The vehicle check-in sheet is still in play and working for the most part. I had to revamp it so it would make more sense to the estimators doing the check-ins. The first draft was way too complicated for them to follow and was geared toward setting customer expectations. They didn’t fully understand it, which led to them just not doing it. The form is simplified now and is vastly improving communications between the estimator/customer and the shop floor.
So, there you go. No sugarcoating, no rainbows and unicorns – just the real deal. These are the many failures that directly point back to me for not properly coaching the team on the importance of these elements and sustaining them.
Changing the Culture
I’ve spent more time on reading/researching culture over the last five months than anything else. I walked into a bad situation, one of the worst cultures I’ve ever been a part of, to the point I didn’t understand why these people got out of bed in the morning. A couple weeks in, it seemed to be coming around, and people were actually having some fun and getting along. Then, for no apparent reason, the bottom fell out and the culture went back to rock bottom.
I can’t count how many conversations we had about the culture of the shop. Every day it changed – it was either “get rid of them all and start fresh” or “it’s a management problem.” I was teetering on the “let’s start fresh” angle, but decided that maybe it was me and that I needed to be a better leader.
I know that no matter what I do, nothing will be sustainable without a winning culture. But I also need data. I have no idea why the culture fluctuates so much, so I need to know what’s on the teams’ minds. I’m in the process of creating an anonymous employee survey with carefully thought-out questions and discover what we as leaders can improve upon to make it better for them. I’ll share the results with you when it’s completed.
The Real Work Begins
The main element of kaizen is the problem-solving work, and I wanted to try to pick the lowest hanging fruit. One of the first opportunities to gain buy-in that jumped out at me was the frustration of the paint team.
There was no organization; techs were just throwing cars and parts at them all day long. There were six available bays in the paint department, with no less than 10 vehicles jammed into the paint line. It was a classic case of overproduction. I stopped the pushing of vehicles into the paint area and saw an immediate response from the paint team. I explained to the body and paint techs that we were trying to establish a pull system and flow and what that meant, and also what overproduction meant. I got them to understand that if the downstream operation couldn’t accept any more vehicles, we would have to move resources to that area to help out.
In between running production and trying to educate the team (in small doses) on where I came from and what I was trying to do, I was trying to figure out my next logical step. I was starting from scratch without the ability to shut down production and implement the system as a whole.
I decided the first step would be to adjust the layout of the shop to support flow. At that point, I saw three key opportunities to change: there were a lot of smaller jobs fed into the system that caused bad decision-making, too many bays causing overproduction and reassembly was located outside the flow line. Vehicles were leaving the paint booth and going outside and back into the building to be reassembled. Also, there was no visual to indicate to a tech that there was a vehicle ready for reassembly. There was a large area outside the two paint booths that could easily fit three vehicles but wasn’t being utilized. I knew that the time a vehicle spent in reassembly was too much and that moving it hastily in front of the booth would blow the whole shop up, but I had no data to back that up.
Once again, I went to the gemba. The feedback I was getting was that putting reassembly in front of the booths would cause a bottleneck, but it shouldn’t because that would be overproduction. So, I researched it. I had the owner buy me a few stopwatches, and I timed numerous vehicles from various stages. I had to tread lightly doing this, because the techs automatically think you’re tracking their performance. I had to explain in great detail the purpose for which I was collecting the data.
I created spreadsheets detailing job size versus the time it took to execute it and tried to quantify it. I measured vehicles in three different phases: reassembly only, booth cycle only and from the start of priming to coming out of the booth. At that point in time, the data was pointing to the reassembly jobs being too large to produce by the time another vehicle was ready for them. In order to boost booth efficiency and speed, all the panels were being painted and cut in at the same time, which led to basically a shell coming out of the booth. Not only was this a large build, there were so many alignment issues either from inexperienced techs putting vehicles together or a defect being passed from upstream that they didn’t have a fighting chance to flow vehicles at the rate we needed.
On paper, I redesigned the shop floor to be able to flow work through the system and stop overproduction. You had your typical areas: repair plan, heavy body, light body, prime, prep, paint and reassembly. In this specific shop, I saw the need to add more because the shop had a large volume of tow-ins and heavy work, along with a lot of speed work. I added: 1) a three-bay speed lane that would have all work (except paint) done in that bay, 2) two “heavy only” repair plan bays and 3) two heavy build-down bays. The heavy build-down bays would have all weld-on panels completed, cut in and built down (interior, etc.) in that area so that once the vehicle hit light body, it could easily flow through the rest of the system. The flow line looked good on paper; all we needed now was implementation and execution of the plan.
We knew we had a lot of problems to fix, but we were shooting in the dark. We had a huge on-time delivery issue and ideas on what was causing it, but no data to support it and know what to attack first.
I started tracking and trending different things: WIP, on-time percentage, cycle time, throughput and severity. Also, tech performance: hours per day and productivity. And finally, a master constraint list.
I created a spreadsheet to track all problems and categorize them into buckets to see what we needed to fix first. When I started tracking all this information, on-time delivery was at 52 percent and the data pointed directly to paint quality 28 times. Breaking the data down even further: 10 color match issues, three times damage painted over, 12 times panel was run so bad or had solvent pop it had to be repainted, and three times they forgot to paint something on the work order. We found our first area to attack with solid data. I’m currently working on an A3 and PDCA (Plan, Do, Check, Adjust) in the paint area and will post our results next month.
The Good Stuff
It’s not all doom and gloom. The team has shown glimpses of greatness from time to time and, through all my failures, we’ve experienced some success that continues to trend upwards:
- Productivity went from 117 percent to 191 percent.
- Hours per day went from 135.7 to 199.6.
- We were also able to take on-time delivery from 52 percent to 79 percent in the last 30 days. There’s still some work to do here, but that’s good progress in a month.
You have to always find a positive to build from. If the team is constantly beat down, the culture will change for the worse. Having transparency to the true numbers of what’s really going on and celebrating small victories with your team goes a long way toward building up the culture.
Next month, we’ll continue through my journey of successes, failures and everything else that I think may be able to help you.
Tim Komoroski is currently a lean facilitator/operations leader for a collision center in Pittsburgh, Pa. He has spent 18 years in the auto repair industry, the last five learning and studying kaizen/lean from an industry leader. Email him at [email protected].