In the Spring of 2007, Fit Couture gave up on its long-stated strategy of migrating its production off site. We really never wanted to be in the manufacturing business, but for a variety of reasons ultimately decided that our product strategy and size required it. After a series of false starts and a hard look at the numbers, we decided outsourcing was impractical and undesirable.
Given that we were committing to continuing to invest in our own manufacturing, we wanted to be sure we were choosing the best manufacturing process that we could find. Prior to his retirement, my father was a lean manufacturing consultant for IBM. IBM’s practice revolved around Eliyahu Goldratt’s Theory of Constraints. Years ago, I read Goldratt’s The Goal: A Process of Ongoing Improvement and found it interesting, though not overly applicable to what I was doing at the time.
In addition to my father recommending a lean approach to manufacturing, Kathleen has long been a strong proponent of lean processes, publishing many articles on this blog about it. I knew that the basic premise of lean – reducing waste, increasing agility, and increasing efficiency were things that we strove for. Prior to our lean implementation, we were a typical (though small) sewing operations. Our two operators sat down at their sewing machines and worked through their batch of garments – fully completing each operation for all garments in their bundle before moving on to the next operation.
We did stress cross training – and both operators could sew any of our styles. We also developed a “library” of operations, trying to duplicate as much as possible across each garment. In general, the release of a new style relied on operations that our sewers were already familiar with and they could look at a sample and work out how to sew a new garment in minutes.
Some of our machines had labor saving devices on them (e.g. computer control lockstitches, chain cutters, etc), but some did not. Because of all the material handling required to move bundles around the sewing room, we ended up requiring a fairly lengthy clean/press/pack cycle at the end each bundle.
Three years ago we first saw the Toyota Sewing System (TSS) in practice at the SPESA show in Miami. America’s 21st had brought an actual working sewing mod, operators included, to demonstrate. I don’t recall the name of the company that owned the mod, but they made wrist guards. We watched as the wrist guards got made one at a time, including cutting and packing right before our eyes by three operators who were standing up at their machines. Not only were they standing up, but they were moving quickly and with a minimum of extraneous motions.
At the time, Amy and I had just gotten our first set of industrial machines for our little home-based factory. We didn’t have any money to buy the tables to set them upright nor had we hired anyone at the time. Still, we wanted to experiment with the style. Amy arranged all of the machines in a ring in our dining room and I set out to learn how to sew so that we could do a sort of sit-down mod. We learned two things from that, first – it was a pretty cool system, though sitting down made it really hard to work and second – I’m not a sewing machine operator.
Since we were planning on eventually outsourcing all of our production, we abandoned TSS and went the bundling route.
This Spring, we saw another demonstration TSS mod at the SPESA show. This time, it was MotionWear, a company in our space that specializes in dance wear. They use similar fabric and machinery to what we use and getting to see a real-life mod in operation made it crystal clear exactly how much more efficient their operators were than ours. We spent hours at the America’s 21st booth quizzing everyone that we could find to talk to. We spent time talking to both the America’s 21st representatives about how the system worked as well as some folks from MotionWear about their experiences.
We decided right there on the floor that if we could make it fit into the budget, we were implementing the system this year – before our busy season started (November – April is our busiest time of year).
When we returned to Houston from Miami, I called Fernando Vasquez at America’s 21st to start talking about an implementation for us. (Fair warning – I found it somewhat difficult to get a hold of someone from them for a couple of weeks. They can be slow to respond to new customers though I will say that there response once we got started with them has always been prompt and in many cases, proactive.) Eventually, Len Egan called me back. Len founded America’s 21st and is their engineer. We talked about what Fit Couture does, and I laid out our budget for this and asked him straight up if they could fully implement the system with what we had available. He indicated that he thought they could. I won’t go into details on the cost breakdown, but we had allocated about $50,000 overall.
The first step of the process was Len coming down and looking at our operation and setup for a couple of days. He timed the operations on several garments and worked with Amy to understand the specific manufacturing steps involved with each garment we made. He talked to me at length about our current volumes and expectations for the future. It took about two days to do everything he needed for our fifty of so styles. After gathering all his data, he left and started working on designing a TSS implementation for us. About a month later, he returned to Houston to deliver the results. A large binder held all the details of their proposal. Included were the financial impact of our current approach, the return on investment and payback period (about a year – though often these are theoretical in nature, and I would expect the true payback for us to take 2-3 years) we could expect, the details of how to sew each of our garments on a TSS mod, the machinery they recommended we purchase, a proposed bonus plan for our employees, and the overall cost of everything.
There is a charge for the engineering proposal. It is in the $5000 range and compensates Len for the significant amount of work he puts into the proposals. (I suspect that its also used to weed out companies that are just tire kicking.)
America’s 21st gets paid two ways. First are the consulting services to get TSS implemented and second is the machinery needed to convert. In our case, we purchased all of the special upright tables, pedals, and motors from them as well as several new machines. My understanding with TSS is that often times you need more equipment than you do with a bundled system. Because of the way the workers construct garments, you often have to duplicate machines to make sure there aren’t conflicts. In our case, we had three or more machines that we needed to duplicate, but we also wanted to upgrade some of the machines that we were already using. Strictly speaking, we could have implemented without the upgrades, but felt it was better to get it out of the way all at once.
About a month or so after we agreed to implement, Ruben Rivera drove a trailer full of our gear from South Carolina to Houston and began the implementation. Ruben’s role was to unpack everything, set up the new equipment and start converting our existing equipment to stand-up. We didn’t want to shut our plant down while he did his work, so he ended up having to come in over the weekend to tackle machines that we needed until the very last minute.
At the same time Ruben arrived to start setting up, Ann Egan arrived. Ann is Len’s wife, though their marriage is much more recent than their history at America’s 21st. Ann’s role was to train our people on how to operate as a team. Her experience has been that without a good foundation in the fundamentals of teamwork, teams often struggle with TSS and find it difficult to reach their potential. Amy and I also thought it would be a good segue from our old system to the new.
A couple of days after Ruben and Ann arrived, Fernando Vasquez came in. Fernando’s job was to help Ruben with any last minute hardware issues and then to train our operators on the system. Fernando planned to be in town for an entire week to work with our staff.
Bright and early on Monday morning, our operators came in to see their workspace converted from sit-down sewing to a stand-up module, with machines organized in a ring in our sewing room. Fernando started working them through the approach that they would use to make garments.
The general idea of TSS is that you want to have only one garment in production for each worker in the mod. So, for us, two operators equals two garments being worked on at a time. Only after a garment is completely finished is another started. When the last person in a line completes a garment, they move back until they reach the station that the worker preceding them is at. They take the garment in progress, “bumping” the previous employee to the start of the line where she picks up a brand new garment and gets started. (In a mod with more than two people, the bumping cascades. Each person bumps the person before them into the first person in the group starts a new garment.)
Fernando had warned us that the first day was going to be messy and he was right. I think Fernando was more worried about Amy’s and my reactions than anything else. I’d be lying if I said there wasn’t some frustration evident from everyone involved – our operators got frustrated with Fernando, we got frustrated with both of them, etc. Both of our operators had made it clear that if we expected them to use this new system, they expected us to pay them more. We’d let them know there was a bonus program coming, but hadn’t talked about its details. Prior to TSS, we’d paid a flat hourly salary. We were going to maintain that and add a bonus based on their efficiency. (Len Egan had provided us with the expected times for each garment, allowing us to calculate efficiency.)
Each day of the week got a little better until we felt reasonably confident that Amy and I understood how the system ran and could run it ourselves. On Friday of that first week, I sat down individually with each of our operators and told them that we appreciated their efforts and thought they were doing great (which they were.) I also outlined our expectation that they would continue to work hard learning the system and that we expected to see them make progress towards reaching 100% efficiency. Amy and I expected it to take several weeks to get to 100% and I stressed that we wanted to see progress and incremental improvement. My goal was to make it clear that there was no “I’ll just do the minimum and collect my hourly rate” case. We expected the team to eventually average 100% a week.
The higher the team’s efficiency, the lower our cost of production per unit. Many companies choose not to pay a bonus to their employees for efficiency, but we decided that if our cost of production dropped, it was fair to share the savings with our folks. In our first week on our own, the team hit an average efficiency of 80.5% and last week they hit 92.5%. Their bonus starts at 75% efficiency and increases the more efficient they become. So, they are already starting to see an impact in their paychecks and we’re already starting to production efficiency improvements.
For us, that’s a mixed blessing. Because we aren’t in our busy season quite yet, we’re still building inventory faster than we’re selling it off (which is why the payback time frame is theoretical). That should turn in another 4-6 weeks and by the height of our busy season, we’ll probably actually be under-capacity, but will have plenty of inventory to be able to draw against. Assuming our growth rates hold up, by next summer we’ll be a bit overcapacity, but ready to install a second TSS mod to handle our busy season (and then shutter it for the slow season).
And finally, some random thoughts about our experience to-date:
- Check the work of consultants carefully. We found a few errors in the calculations that Len made. Most were minor, but one in particular had a huge impact. Fortunately, it was easy to spot and easy to correct. In the end, its still your business, so you’ll want to keep a close eye on things.
- America’s 21st really impressed me with their focus on getting us the right gear. They took away quite a bit of equipment that we’d planned to purchase because once they got on site and saw things in operation, it was clear we didn’t need it. Someone willing to do the right thing like that was refreshing.
- On the other hand, while the mod was up and running when they left, not everything was fully functional. One of our new machines is still sitting on the floor of our warehouse, a second wasn’t upgraded at all as spec’ed out and is sitting useless on the mod, and a third didn’t receive the upgrades we were expecting and is still manually operated. I expected some glitches during the install, but I didn’t expect to have quite so much not implemented. Ruben is supposed to be coming back next week to finish up, so they aren’t going to leave us hanging. He’ll also be doing some additional work for us beyond the original plan.
- Getting to 85% was relatively easy. Both our operators are comfortable with the system and don’t need any hand holding to run it. But, getting that last 15% to 100% efficiency is requiring them to hustle in a way they didn’t before. They are having to economize their movements, plan a few steps ahead, and just generally move a bit quicker. It’s not for nothing though, the bonus difference between 85 and 100% will be significant to their take-home earnings.
I don’t want to go into too much more about the specifics of what we spent and other financial information. We’re hosting a “lean party” at our Houston factory on November 2nd for people interested in seeing TSS in operation. I’ll be happy to share cost and return information in more detail in that setting.
The TSS orientation at Mike’s facility for Nov 2nd, is full. If interested, email me (kf) to be placed on the waiting list in case someone drops out. Mike has mentioned he’d be amenable to adding another session were there sufficient interest.