Our Sites

Learning to see waste and become a more focused sheet metal shop

Metal fabricators cannot eliminate waste they cannot see

Rolls of sheet steel stored on racking on factory floor

PKM1 / Royalty-free / Getty Images

Made in Japan. The phrase meant cheap, disposable, and low quality in the 1950s and 60s. Today, it often implies a product is made with exceptional quality. Much of Japan’s transformation is attributed to individuals like Taiichio Ohno, who in 1943 joined Toyota, where he became known as the father of the Toyota Production System and, eventually, a legend in the realm of lean manufacturing. Legend has it that he’d coach department managers to develop “eyes, ears, and a nose” for the seven types of waste: transportation, inventory, motion, waiting, overproduction, overprocessing, and defects. To this end, he would take each department manager to a central location in his department, draw a circle on the floor with chalk, and order the manager to stand within it and observe.

Ohno would then leave for an hour or more, and upon his return, ask the manager (still standing in the circle), “What have you seen?” The nervous manager might say something like, “I saw people working.” Ohno would shake his head and say, “You are not seeing! Keep watching.” Another hour or so would expire. This process would repeat itself until the manager’s reply satisfied Ohno. “Those parts over there have not moved in the last hour. People are busy, but they are just handling and rehandling components.” Ohno probably responded, “Now you are seeing it. Well done! What are we going to do about it?”

Viewed through the lens of 2023 cultural norms, the technique might be considered condescending or demeaning, but you cannot argue with the results. Learning to see waste is step one. You cannot eliminate what you cannot see.

After serving as a manufacturing consultant for over 30 years, working with over 350 companies and facilitating over 1,000 kaizen project teams, I can tell you that there is waste in every process. It’s shocking to most managers when, after direct observation, they realize their teams are operating within a process having significant nonvalue-adding activity, often between 25% and 30%. If they could eliminate the wasted motion, unnecessary waiting, repetitive stacking and unstacking, counting, moving, damage from handling parts unnecessarily, miscommunication between departments, wasted space covered by inventory, and countless other factors, they theoretically could manage their businesses with labor costs that are 25% to 30% lower. More positively speaking, they could produce and sell 25% to 30% more product with the same staffing levels.

The Cellular Sheet Metal Shop

This is not just a dream scenario. I grew up in a precision sheet metal job shop and experienced a Toyota-like transformation firsthand. In the early 1980s, our largest customer, Tektronix, demanded that we deliver weekly batches of 150 mainframe computer cabinets instead of the 600 per month lot sizes we were used to. They had heard about Toyota’s JIT approach, and Tektronix demanded all their suppliers adopt it.

Our assembly department quickly adapted to the change. Our fabrication department continued to punch, form, and insert hardware for 600 units in a batch (to minimize setup time). We assumed that we had met the customer’s expectations. But shortly thereafter, we were informed that the goal was not just to assemble cabinets 150 at a time, but to fabricate and assemble 150 each week. “And oh, by the way, at some point we will want you to fabricate and deliver 30 per day!”

How could we possibly be expected to set up machines for dozens of complex parts 20 times each month? After all, we had nearly 100 other customers to satisfy. At the time, we had three turret punches, five press brakes, and half a dozen hardware machines.

We peeled off one turret, two press brakes, and a hardware machine, then added a spot welder and a small assembly area. These mild steel parts also needed to be plated, so we had to coordinate with our plating vendor to turn around 30 units daily (handled each day by their second shift) so that we could assemble each day’s fabricated parts beginning the very next morning. It was as if the parts never left our shop. When there was no need for outside processing, we tied fabrication and assembly processes directly together.

Careful value stream mapping (VSM) allowed us to determine the best sequence in which to punch and form the components. Step two was to determine a manufacturing rhythm or takt—that is, the available time divided by demand. Knowing our takt time would allow us to determine how many people, how many pieces of equipment, and how much space would be required.

Our team had 36 hours available to produce each week (accounting for break times and planned downtimes). We needed to produce 150 units per week, which, according to takt time, required us to fabricate and assemble one mainframe computer cabinet every 14.4 minutes (36 hours per week × 60 minutes per hour divided by 150 units = 14.4 minutes per unit). If there were 25 minutes of forming in each unit, then (25/14.4 = 1.74, round up to 2) we would need two press brakes and two press brake operators.

Manufacturing workers

FIGURE 1. The team at the fab shop Gary Conner worked for during the early 1980s. After going cellular, the company went from 70 to 270 people within just a few years.

We also balanced cycle times between operations. We did this by reviewing operator and machine cycle time data gathered during VSM. This helped us establish standard work, or the assignments each person on the team must perform to maintain takt time and smooth workflow. If the total average cycle time to process a unit is 100 minutes, and the takt time is 14.4 minutes, then we need 6.9 team members (100/14.4 = 6.9). We try to avoid loading people more than 85%, so we rounded this up to 8.

The next step was to divide the work as equally as possible, in this case 100 minutes of work divided by 8 people (100/8 = 12.5). Each person should be assigned 12.5 minutes’ worth of work during each 14.4-minute takt time. This allowed time to physically and mentally recover, clean the work area, or assist a teammate.

If we experienced minor imbalances, or if one process couldn’t be managed within the takt time, we had to be flexible. We rotated break and lunch times and staggered start times. In the worst-case scenario, we added a few extra hours of production to the constraint process—something we viewed only as a stop-gap measure. If a bottleneck persisted, we considered rebalancing the standard work, adding staff, or, as a last resort, temporarily farming out some work to another team.

We had a saying in our sheet metal shop: “It does not matter how many parts we punch and form; it matters how many assemblies we get on the truck!” In a batch manufacturing mode, punching and forming operators (driven by departmental goals) may have had little or no concern for whether the assembly team had the parts they needed to complete their work. Out of sight, out of mind.

Our team chose to reengineer the turret punch programs to capitalize on a standard turret while still allowing us to punch in kits. We developed a staged tooling strategy that allowed us to bend different components across the same press brake setup. This did require a few one-time tooling purchases and lengthened some setup times, but the average setup time dropped significantly and had a huge financial impact. Moreover, the arrangement allowed the hardware, spot welding, and assembly operations to immediately process the parts without waiting for all 150 of one component to be formed, only to wait more for the mating components (150 of each) to catch up.

Understandably, business owners want to optimize their most expensive machines, and our shop was no exception. One turret punch generally kept two or three press brakes busy. Our customer demand required two press brakes at about 90% capacity, but it left the turret punch idle about 40% of the time.

The production manager and owner would frequently visit the cell’s team to investigate why they were not hearing the rhythmic and constant thump, thump, thump of the punches. We had to explain that we simply did not need to punch parts ahead just to build inventory that no one could use. And besides, we had no space to put the parts in our tiny work area. Our cell was designed for efficiency, not to be a warehouse. Our mantra was, “If we pick it up, we finish it.” (To quote one of Ohno’s favorite sayings, “Sleeping parts make no money!”)

Line balancing was critical, and since some components were hardware intensive while others required little or no hardware, we built specialized carts with heavy-duty casters. These allowed the punch and brake operators to move the hardware machines closer to the primary operation, where they could install hardware on some parts. And for more hardware-intensive components, we assigned a dedicated hardware installer. When dealing with the reality of job shop work, flexibility was key.

While focusing on one-piece flow did not immediately make sense to us, one-kit flow did seem more reasonable. But even one-kit flow did not always make economic sense, so we explored a 25-kit flow, 10-kit flow, and five-kit flow, pushing the envelope until we exposed the threshold of pain, then backing off to a reasonable kit quantity.

Eventually, the issue of machine optimization resurfaced as a target-rich opportunity. We chose to standardize our turret punch tool list. By doing so, we were able to maximize the number of parts we could process without a setup. Utilizing our excess capacity, we programmed and punched parts for other teams. This helped to solve the turret punch overall equipment effectiveness (OEE) issue, and in the process relieved a good deal of stress for our management team.

Lean manufacturing graphic

FIGURE 2. This layout, typical for a sheet metal job shop, promotes excess movement and local, versus global, efficiency.

At times, our largest customer hit a sales slump or went through a design change, resulting in orders drying up. We worried and wondered if our team and cellular layout would be disassembled and absorbed back into the rest of the departmental layout of the larger shop. But we became so adaptable at digesting work of any type that within just one year our team of eight people (which represented about 10% of the 70-person shop) was producing 40% of the revenue. Once management saw that figure, it was only a short time later that the rest of the shop was divided into smaller work teams that mirrored our cell.

Our company became one of the leaders in adopting and adapting the Toyota Production System to a make-to-order environment. No one in our region could really compete with us, and we went from a 70-person shop to a 270-person shop in just a few years (see Figure 1). Eventually adding lasers, more turrets, and associated machinery, we broke into multiple cells across multiple shifts. Ultimately, we had close to a dozen customer-specific teams, each responsible for a discrete number of parts or assemblies.

Even after our company broke into smaller teams and focused factories, we continued to practice an Americanized version of “standing in the circle.” We never assumed we eliminated all the waste in a process. Gemba walks (regular visits to the work area) and seeking input from work teams (kata) remained essential.

Connecting Fabrication With Assembly

Divided into functional departments, the traditional shop layout optimizes machine utilization (see Figure 2). It results in lots of motion, transportation, waiting, and tends to foster overproduction and the resulting inventory. It also creates a significant disconnect between fabrication and assembly. The fabrication department lead person is generally measured on the “output” from their department.

Figure 3 presents an alternative, requiring the movement of only a few smaller machines. Notice the line-side assembly area, much smaller than the one shown in Figure 2. The area can permit a handoff directly from fabrication personnel to the assembly staff.

Having a right-sized assembly area embedded in the value stream improves teamwork, communication, cross-training, and feedback. It also makes it easier for the team to share work and balance the line. Since the value stream alignment means there are a discrete number of parts to assemble, we can streamline the number of hand tools and space required.

Note that where there are no repetitive jobs and each work order is unique, or when parts must leave the building for outside processing, an immediate handoff between fabrication and assembly might not be possible. That said, a cell can certainly facilitate those efficient handoffs for jobs without those requirements.

Having fab and assembly capability close together within the cell presents an entirely new set of considerations. If components being assembled are not all available at the same time, then assembly grinds to a halt. Again, kitting is part of the solution (see Figure 4). Punching and forming components in the sequence that the assembly team needs them demands clear communication. Components requiring subassembly need to be processed first. A few minutes spent reviewing and strategizing the assembly sequence generally helps determine the correct fabrication sequence.

Working with smaller lot sizes and kits, people responsible for downstream processing (such as welding, spot welding, or assembly) discover fit and function issues immediately. In batch manufacturing, such errors might not show up for days. Efficient feedback from downstream processes quickly improves the quality of output, reducing rework and further contributing to financial gains.

Communication remains critical, because let’s be real—eventually there will be a tooling conflict, machine-specific requirement outside the cell, or some other reality that interrupts the flow. Again, after a quick but careful review of the day’s work orders, the cell’s team can identify potential roadblocks, brainstorm intervention techniques, and coordinate with other teams if needed. For instance, when die clearance and tooling incompatibility issues arise, having a small amount of material as work in process (WIP), or having a Kanban (pull) replenishment system, can go far in reducing the waste of waiting for parts.

Metal fabricated cart

FIGURE 4. A kit of parts is staged on a custom-built cart. If all parts aren’t available at the right time, assembly can grind to a halt—hence the need for constant communication and problem-solving.

Benefits of the Focused Factory

In a functional, departmentally aligned job shop, operators might process hundreds or even thousands of unique parts, possibly not seeing the same part repeated for months. On the other hand, cellular teams working as a focused factory might be accountable only for a few dozen components from a handful of customers. This builds a real sense of ownership and the satisfaction of producing a customer’s product from beginning to end, rather than performing a single operation only to see their hard work palletized and hauled away to another department. They can feel no real sense of completion.

Focused factories are very much like a factory within a factory. Teams are responsible and accountable for a defined set of parts. These might require similar process steps, or they might be of the same material type or are perhaps for a specific customer.

Because they gain familiarity with a specific set of parts, teams have a better chance of working out the bugs. They become familiar with the nuances of a particular material, part size, or other such characteristic associated with their assigned set of components. They perfect setups, identify quality expectations, develop shortcuts, and discover and document the potential pitfalls. They standardize best practices and establish standard operating procedures.

So, how do you get started? My advice is, start small, but start soon. Select a small value stream that can act as a model line where you can test the process and educate your teams in the many disciplines of world-class manufacturing. Not every lean tool applies to a job shop, but dozens apply to any company, regardless of size, location, customer base, or product type. These include 5S, VSM, total productive maintenance (TPM), lot size reduction, cross-training, cellular layout, pull systems, visual performance metrics, and more.

We have an obligation to our businesses, customers, families, communities, and ourselves to continually grow in our ability to see and eliminate waste. While Ohno’s methods might have been viewed as harsh or overbearing, he certainly proved that by developing an awareness of waste (in all its variations) and eliminating as much as we can, we set ourselves apart from the competition. So, go stand in the circle!

Lean manufacturing graphic

FIGURE 3. This layout alternative reorganizes the fab shop to incorporate a fabrication and assembly cell. The layout includes a smaller press brake and hardware department for parts that don’t yet flow completely through the cells. Transitioning to cellular manufacturing doesn’t happen overnight.

About the Author

Gary Conner

Independent Consultant

Gary Conner is a retired lean manufacturing consultant for Oregon Manufacturing Extension Partnership (OMEP).