Microfactories, Automation, Talent, Investment
Deconstructing the Reality of Made-in-the-USA Manufacturing
By Jordan Speer
When you’re talking about automating the apparel industry, and where we’re headed in the future, it’s helpful to understand where the industry has been, why it changed, and what’s sparking the changes that it will need to accommodate going forward.
A little industry background
If you go back many decades, you’ll find a thriving textile and apparel industry here in the United States; in 1950, the industry was bustling with nearly 2.5 million employees, and 23 years later in 1973, that figure was about the same, (although it represented an increase in apparel jobs and a decrease in textile jobs).
By 2005, that figure had declined to 1.3 million workers, and by 2017, to 341,000. In the 25-year period from 1991 to 2016, apparel had the largest decrease among all U.S. manufacturing industries with a decrease of 85 percent, and over that period, employment in textile mills decreased 76 percent.
Why the enormous declines? There were multiple shifts taking place, but the two primary culprits were technology automation and offshore sourcing. As processes in textile and apparel production were increasingly automated — think improved weaving and knitting machinery, laser cutting of fabrics replacing manual cutting, or sewing machinery that automated specific tasks such as pocket welting or button holing — those developments in automation were already eliminating jobs. But it was the exodus of manufacturing offshore that really put the nail in the coffin of large-scale textile and apparel manufacturing in this country. It also curtailed much of the research and development into automation — with labor overseas so cheap, why bother?
There were many reasons for the exodus offshore, including cheap labor and free trade agreements such as NAFTA and CAFTA that allowed U.S. apparel companies to manufacture offshore but then import those garments without duties, if they used certain amounts of U.S. fabrics and other inputs. For companies that did not initially move to source offshore, most were eventually put out of business or forced to follow suit, both because of the inability to compete with companies who paid so much less for labor, and because of the decreasing capability of both finding materials and manufacturing sources here in the States, as mills and factories closed up shop.
(It’s instructive to note, too, that this ‘chasing the low-cost needle’ had already occurred here in the United States, with the bulk of the apparel and textile industry moving in the mid-20th century from the Northeast, where it initially thrived, to the Southeast, where labor was much cheaper and unions did not have a great stronghold.)
Some of the earliest pioneers of offshore sourcing went to the ‘Asian tigers’ (Hong Kong, Korea, Taiwan); then, in the 90s, the Central American region became the focus of U.S. offshore manufacturing, but ultimately Asia, and especially China, became the primary sourcing destination; labor was more plentiful and cheap and there were more opportunities for “full package” manufacturing. For a while, quotas that had been put in place decades earlier under the World Trade Organization (WTO) restricted the amount of sourcing that could be conducted in China and other nations, but with the elimination of the last set of quotas on Jan. 1, 2005, the end of decades of quantity restrictions were lifted, and apparel sourcing in Asia exploded.
By 2009, just 2.2 percent of apparel sold in the United States was also manufactured here. Since that year, when the industry bottomed out in terms of apparel jobs leaving the country, apparel manufacturing has returned, to the tune of 3.0 percent. That figure may not sound significant but, as industry veteran Ed Gribben noted when he shared those statistics in an Apparel webinar last year entitled, “Made in USA: An Industry Talent & Skills Gap Assessment,” it represents quite a number of jobs, and billions of dollars of Made-in-the-USA goods.
Made-in-the-USA: What does it mean to say the apparel industry is experiencing a resurgence?
So here we are in 2019, and there’s a lot of buzz about textile and apparel manufacturing returning to the United States. The internet has forever changed how consumers shop, and where their expectations are set. Today’s shoppers are not restricted by store hours, have instantaneous connection to a virtually endless number of retail options and to trends in real-time. This has led them to expect access to an ever-changing lineup of new products, delivered to them with lightning speed, at the location of their choosing. On top of this, consumers now expect to be catered to on a one-on-one basis, with marketing messages and offers targeted toward their preferences, and a range of fulfillment options to suit their needs. They also expect opportunities to customize the actual product itself, by size and by style. Indeed, 74 percent of Millennials and Gen Z are interested in buying products that are customized to their taste or made specifically for them, says Toni Lublin, communications and partnership manager for Lectra North America.
It’s against this background that much lively discussion took place at the SPESA Advancements in Manufacturing Technologies Conference, which was co-located with the TechTextil North America show in late February, in Raleigh, N.C. The all-day conversation focused on where the U.S. industry is headed and what needs to be done to grow U.S. apparel manufacturing. Topics ranged from the state of microfactories and automation to the need for better workforce training and the improvement of the overall image of apparel and textile manufacturing as a career. The discussion also ranged into areas of machine functionality and the unique properties of fabrics and apparel that present distinctive problems to this industry.
Building microfactories to cater to on-demand shopping
In the digital age that we’ve entered, the ideal scenario, across industries (and leaving aside discussion for the moment of job losses) is to have one set of data, about any product, that flows seamlessly and automatically across the supply chain. This is entirely possible with the technology we have today, said Kirby Best, president and chairman, PAAT International. Even so, we are far from achieving it, for a number of reasons, including the fact that some technology and equipment manufacturers use their own proprietary codes and languages that do not necessarily make integration easy. And then there is the work of doing it, which on the scale of flow we are talking about is still many years out, but on a smaller scale, there are instances where you can see this at work.
When you have data that flows easily from system to system, that data can direct robotics and machinery to automatically run operations — if you have robotics and machinery that can run the specific operations you want to do. If you can manufacture a product from start to finish without human intervention, you have fully automated your supply chain.
Today, you’ll find some version of this type of scenario in microfactories, which are small-dimension factories that take up a fraction of the space of a typical factory, are highly automated, use less energy and can quickly produce a product from start to finish. Microfactories in apparel — which to date do not carry a product from end to end, more on that in a moment — are ideal for on-demand orders, small-batch orders and customized manufacturing, and because of their small scale, they can be set up more easily and at lower cost, allowing for a localized strategy. Imagine a scenario where a brand sets up microfactories around the country, with consumer and also retailer orders directed to the closest location, where product is produced on demand, in a few hours, and shipped out the same day to the purchaser. Today, many brands are having conversations about setting up microfactories in towns across the country, to service local need, said Michael Rabin, president of Morgan Tecnica America Inc. “I know of one company that is already doing this, manufacturing baseball and football jerseys in locations around the country that cater to the specific to the needs of the local community,” he said.
Beyond their speed and on-demand production, or perhaps more accurately, intrinsic to them, microfactories can address one of the greatest challenges of apparel retailing — inventory management — by eliminating the need to produce well in advance of selling, which is how apparel has traditionally been produced. That method inevitably leads to overstock (and thus markdowns) or out-of-stocks (and thus unhappy customers and lost sales).
There’s a need for quick, U.S.-based manufacturing, because there’s a demand for it. “What we hear from our customers is that they need to reduce time to market and inventories while responding to the see-now-buy-now attitude of customers,” said Lenny Marano, vice president, product management and marketing automation systems, Gerber Technology.
Make no mistake: microfactories aren’t just for local production. Rabin notes that in China, you’ll find scads of traditional large-size factories that have set up microfactories within their larger structures to handle on-demand orders from the United States and elsewhere, with faster turnaround time.
Automation, whether microfactory or not, also addresses a challenge of Made-in-the-USA manufacturing, which is that labor is both expensive and unavailable. Higher-tech jobs driven by automation in the industry require new skills that much of the current workforce doesn’t have (and those people who do possess those skills often go into ‘sexier’ industries). On the flip side, apparel manufacturers cannot find people to do traditional sewing operations that are still necessary, because there is simply no interest. “No one wants to be a Juki girl,” says Marano.
In the digital apparel revolution the sewing is the obstacle
Which brings us round to the actual sewing. There’s just one problem with automating apparel production, whether in a large-scale operation or a microfactory: fabric moves. Unlike cars or microchips, apparel assembly requires sewing one piece of flowing, draping, stretching material to another, with significant variation in drape and texture and stretch and thickness and hand from fabric to fabric, and with significant variation in how those materials are held and tensioned and moved in the sewing of a garment, and with significant variation in the type of stitch used to put two pieces of fabric together. (If you’ve never thought about this, go to your closet, flip a few garments inside out, and take a look at the different types of stitching you’ll find.) In other words, there is a big difference in terms of handling when it comes to attaching a door to a car body vs. sewing the side seam of a cotton/poly t-shirt, and there is a big difference in terms of the variability of handling when it comes to sewing the side seam of a cotton/poly t-shirt vs. fell seaming a pair of denim jeans. “Sewing is all very product specific,” says Rabin. “You’re not going to handle a pair of jeans the same way you handle a tent, or workwear; how we pick it up and move it around is different.”
The upshot? To date, the entire process of sewing a garment generally speaking cannot be left to robots, and this is the “no. 1 problem” when it comes to automating the end-to-end apparel supply chain, says Yoram Burg, regional sales director, U.S. and Canada, EFI Optitex. There is some progress in this area. Sewbo Inc., created by Jonathan Zornow, offers a solution that utilizes a polymer called polyvinyl alcohol to temporarily stiffen fabrics so that they can be manipulated and sewn by robots; additionally, Softwear Automation’s “Sewbots” — which spun out of research at Georgia Tech’s Advanced Technology Development Center that was funded by a grant from The Department of Defense — can make a t-shirt from start to finish in about four minutes.
Still, the work of sewing clothing remains today a job for humans, although adding automation where possible and sensible is an industry goal. “We have infrastructure and technology to get seamlessly through the cut room. How do you automate to get from the cutter to production? How do we get apparel out the door in a more automated way?” asked Marano. “Little by little, bit by bit,” using high tech components, using cobotics (collaborative robots) that work alongside humans, some of which employ cameras and visioning techniques that allow robots to “see” and direct them where to place things, and so forth, certain operations are being automated. But figuring out how to automate a sewing machine and then engineering it is only worthwhile if you’re going to do an operation a few hundred thousand times or more, reminds Frank Henderson, president and CEO of Henderson Sewing Machine Co., because automating a process can really only be accomplished for the same operation on the same type of garment.
Even though the sewing process is only semi-automated, huge strides have been made in the digitalization of sewing machines. While the traditional emphasis focused on mechanical capabilities (how do you get your machines to run longer and faster with better quality), recently, sewing machines have moved into the electromechanical age, says Mariano Amezcua, president, DAP America Inc., where machines can be networked together, and where data about every step of the process (machine downtime, number of stitches, amount of use) allows apparel manufacturers to better monitor the machines and the operators, identifying shortcomings in real-time and making adjustments before problems happen, whether that’s preemptively fixing a sewing machine (which can be done remotely) before it breaks down, or adjusting a workload based on operator performance. “But you can’t use a new machine with new technology and run it the same way as you ran an old machine,” he says. “That would be like buying a car and then sticking your foot out the door and moving it like a skateboard.” Likewise, older sewing machinery cannot be digitalized.
We’re also short on talent and investment
So, while the process from design file to cutting can be fully automated, the apparel microfactory, as it were, still requires humans when it gets to the sewing parts, and not a lot of humans want these jobs, which is a challenge the industry has been struggling with for years now. And it’s not just the sewing machine operators. There is only one — that’s one — mechanic in the entire state of Texas who can fix sewing machines, said Michael McDonald, president, SPESA. “No one wants to be a Juki girl, but no one wants to be a mechanic either,” he said.
Most of the old guard — people who worked in sewing factories in their heyday — have died, retired or are aging out of the workforce, and the younger generations are seeking work in industries they believe are more high tech. But this mindset, as many in the discussion noted, is less about reality than perception; the apparel industry is increasingly full of high-tech excitement and new ideas. Consider, said Gribben, companies such as Rent the Runway, Stantt and ThirdLove, which are revolutionizing the apparel industry. Part of the challenge in rejuvenating the industry lies in changing its image, which will require a coordinated effort not only at the industry level but also at the school level.
“There are multiple issues,” said McDonald. “Training is one of them. Union Special used to do this. … [But then,] people wanted to go into fashion, or chemical engineering. Now you’re seeing a shift. People are seeing that manufacturing still exists, but now you can’t just be a mechanic. You have to be a software engineer too. [Awareness about this industry] can’t just happen at university level. It has to happen at high school level. There are a lot of components to this,” he said.
Cindy Istook, professor of textile and apparel design at NC State University, concurred, and suggested getting the message out even earlier. “We need kids in high school and as early as middle school to see that this is an advanced, highly technical industry, and we also need the parents to understand that this is not ‘sweatshop’ culture.”
The other thing the industry will need is more of it. “We talk about bringing manufacturing back to the U.S., but people are afraid. We lost this industry 20 years ago; we’re going to need people to start investing,” said Rabin.
“It’s hard to turn a steamship. Five years ago, the early adopters of microfactory on-demand manufacturing were small and agile, but over past two years, we’ve seen an acceleration of on-demand cells co-existing with regular factory floor,” says Marano.
“[The year] 2008 was the ‘end of the world,’ if you remember. Now it looks like a small hiccup,” says Burg. “If we compare where we are with microfactories to other industries, you’ll see we’re so behind. As retailers, brands, manufacturers — how much are we willing to invest in technology? Are we going to just react to changes in customer demand, or are we going to take an active approach?”
The changes may be coming slowly, but they’re promising. As Gribben noted: “I’m very optimistic about our industry going forward.”