To make a new kind of shoe, adidas had to change everything

A pair of adidas Made for London trainers sit on a powder-blue tray table the size of an A3 sheet. The shoes don't look like much - but then they wouldn't, because they haven't been made yet. The scant ingredients include two kinds of spooled thread, three cups of white plastic beads and a couple of rolls of green tape. It's like coming across a Longhorn cow, then hearing that, this time tomorrow, it will be a Chesterfield sofa.

"In Asia today, it takes between 90 and 60 days to turn these materials into a product," says Gerd Manz, adidas's vice president of technology innovation. "Today, if we're ambitious, we can go from here to final product within… Uli?"

Manz, a tanned, square-jawed 45-year-old who wears his black adidas T-shirt tightly tucked into his blue jeans, looks over at Ulrich Steindorf, the sportswear giant's gangly senior director of manufacturing.

"Days, I would say," Steindorf replies. "There are some production related settings…"

"Within a day," says Manz firmly.

"Within a day," repeats Steindorf, "you can make a shoe."

As Manz speaks, an orange robotic arm the size of a small digger lifts into the air with a rhythmic hissing. Nearby, a digital laser cutter whirrs into place, using cameras to identify its target. It's a hot day and the air is stuffy, but inside the 4,600-square-metre warehouse all is calm. A few workers in black adidas polo shirts stroll across the polished plastic floor, pausing to tap at rubber-cased tablets.

This factory is very different from the vast, cramped workshops of Asia, where workers as young as 15 hand assemble 97 per cent of the 360 million shoes adidas produces each year. But the most remarkable thing is its location: we're in Bavaria, just an hour's drive from the small German village where adidas founder Adolf "Adi" Dassler began making sports shoes after his return from World War One.

Adidas is building a similar site in Atlanta, Georgia, and "seriously considering" further sites, including London. When Atlanta starts production at the end of 2017, the two factories will churn out a million pairs of shoes a year. What's more, thanks to robotic flexibility, there's the possibility that every single one could be tailored to taste and fit.

"It's the future of shoemaking," says Manz. Adidas calls it the Speedfactory.

Over the past 40 years, the majority of shoe production has migrated east in search of cheap labour. For the Speedfactory, it has returned home.

Summer, 2014. In Rio's Maracanã Stadium, MarioGötze raced through in his adidas kit to slot an adidas ball past the adidas glove of Argentina's Sergio Romero and win the FIFA World Cup for Germany, adidas's national team. Sporting success, however, concealed a sombre outlook for the brand with three stripes. Away from the football field, adidas was losing.

Most concerning were the firm's struggles in the crucial US market, which accounts for more than 40 per cent of the world's sportswear sales. Lagging behind cooler, more confident Nike, as adidas had for decades, was bad enough. Now, after several years of underwhelming shoes and unlucky endorsements, it fell back even further, slipping below Under Armour for the first time, as annual sales slumped 23 per cent to $1.1 billion (£859m). (By comparison, Nike, who made Götze's World Cup-winning boots, had sales of $8.9 billion in the same period; as one headline put it, "Under Armour is beating adidas in the race to eat Nike's dust".) If adidas was going to compete, something had to change.

"Our two main competitors are constantly reinventing the model," the firm's new North American head, Mark King, admitted to [i]The Wall Street Journal[/i]. "We just have to move faster."

Inside adidas, speed was on everyone's lips. "If you want to be new you have to be first," says creative director Paul Gaudio. "We were a bit more measured. We contemplated and pondered the perfect path." Gaudio, a tattooed, silver-haired Pittsburgher who has worked for adidas on and off since 1991, took charge of the firm's design in September 2014 as part of an all-encompassing "reset", and immediately pushed his 650-person team to "get things out there". The mantra was made official in March 2015, when adidas CEO Herbert Hainer launched a five-year strategy promising to "transform the adidas Group into the first true fast sports company". But for all the symbolic changes - sports clocks in meeting rooms to stress that time was always ticking - one question remained unanswered: how could you speed up a business as sprawling and divided as adidas?

Nike: Nike has partnered with supply-chain experts Flex to double the speed of its production, a project it calls "Manufacturing Revolution". It's also teamed up with DreamWorks on animation software for shoe designers.

Under Armour: In June 2016, the US firm revealed "Lighthouse", a 3,200-square-metre test lab filled with 3D printers, body scanners and automated assembly robots. The brand's new Architech range of shoes have 3D-printed midsoles.

Puma: Adidas's historic rival has launched a pilot project to create a robotic "intelligent" warehouse. The shoe brand is working with German firm Magazino, whose Toru robots can pick and stack without human involvement.

The problem was the same system that had turned sportswear into a £300 billion industry: outsourcing. The shift was total: when Nike was founded in 1964, just four per cent of US footwear was imported; by 2014, it was 98 per cent. But in order to take advantage of cheap South Asian labour, companies were forced to split themselves in half. Fifty thousand in-house staff handled design, development, marketing and sales; a million independently contracted labourers did the sweaty work of manufacturing. The two sides of the business talked, but in cordial, formulaic ways, like divorced parents trying to arrange childcare.

The most visible downside of the split were the conditions in the factories. But firms also faced the niggling, day-to-day difficulties that came from losing control over the manufacture of their own product: quality control; six-week waits for ocean shipments; piles of marked-down goods because forecasts were over optimistic. "They call it the supply chain, but a better name would be the sprawl chain," says ManMohan Sodhi, professor of operations and supply chain management at Cass Business School in London. "You keep chopping operations into more granular pieces, so you shave pennies off, but now you're less flexible." Individual shoe sizes could only be made in lots of 20,000. Factor in design and the typical shoe took 18 months to produce.

Like so much else, this system was upended by the internet. Now shoppers could see what the rest of the world was sporting on social media, they weren't so interested in the shoes retailers had ordered wholesale nine months earlier. Fashions flared and died online in the time it took to get a prototype back from China. The world was leaving brands like adidas behind.

"When we started our data initiative back in 2014, everybody was carried away by, 'Let's do trend analytics and predict which colour is hot,'" recalls Michael Voegele, adidas's chief information officer. "And I said, 'Oh, that's fantastic guys. We'll do that. And then we'll all wait until we deliver the product two years from now.'"

Adidas went after delays in its supply chain, cutting off drag, like a swimmer shaving their legs. It was a painful adjustment. "We were killing ourselves to get days out of the existing process by being more disciplined," says James Carnes, vice president of brand strategy. "Reducing waste time, ordering times. Getting days and weeks out of the buffer."

Carnes, a slight, black-clad American with a carefully razored blond haircut, joined adidas as a designer in 1995, but in the company's search for speed he found himself moved into strategy. "There were two parts," he says. "One was just speed: how do we get faster? And the other part of it was, what's the big revolution? How do we completely change the model we have? What would be a completely different alternative?"

That was when Manz stepped forward with a radical proposal.

To understand shoe factories, it helps to think of a kitchen. Cooking turns a raw material (ingredients) into a product (food). There are two ways of doing this: change the ingredients' form (chopping, crushing, mixing) or change their chemistry (heating, cooling, combining with other ingredients). Chefs trying to come up with something new have to do one or the other. With the Speedfactory, Gerd Manz did both.

Manz joined adidas's Futures Group, the division responsible for new technology, in 1997. The first product he worked on was the Climacool breathable running shoe. An earnest, low-profile German with a degree in Leather Technology - his LinkedIn page has two listings: experience (adidas) and interests (adidas) - Manz might seem an unlikely radical. But, after ten years at adidas, he was increasingly frustrated with the group's focus on designing what he thought of as "gadgets".

Changes in other industries made those limitations especially grating. Silicon Valley firms such as Amazon and Apple weren't just releasing products, they were collecting data so they knew how to improve. Fast-fashion firms such as Zara had built factories in Europe to get clothes from sketch to store in as little as two weeks. But when Manz looked around him, there was nothing remotely comparable in the pipeline.

Read more: These lightweight adidas shoes are made from spider silk grown in a lab

"We thought, goddammit, we're cooking in the same kitchen with the same ingredients as everybody else," he recalls. "All of our competitors have the same board with the same ten spices on it, and because we all go to the same suppliers and get the same catalogue, the only thing that changes is what we pick. So all the soups that come out of this kitchen taste the same, no matter what the brand."

Manz started to look for new ways of manufacturing. His opportunity arrived in 2009, when a representative from the giant German chemical producer BASF got in touch about a lightweight plastic with tiny bubbles of gas trapped inside, like a miniature tennis ball. The material was highly elastic - in tests, it sprang back to its original state much faster than equivalent substances - but BASF didn't know how it could be used. The representative showed Manz the plastic in its raw form of small white beads. "All we had from them were these Tic Tacs on the table," Manz says. "OK, how do you make a shoe out of this?"

The answer - arrived at after several years of tweaking, testing and designing - was to press 2,500 of the beads into a mould, creating a super-springy cushioning for the middle of the shoe. The resulting midsole, with its distinctive rice-cracker look, was the biggest trainer innovation since air cushioning. "It was so unique," Gaudio recalls. "You could instantly feel the difference." Adidas introduced it in February 2013 under the name Boost.

Inside the Speedfactory, Manz demonstrates how the Made for London's Boost sole is made. In a metre-tall glass case, a machine blows hot steam on to a mould. Then a robot with a vacuum pump on the end of its arm sucks the sole up and, moving with terrifying speed, deposits it in an exit slot. Manz picks one up. "They're still wet and warm," he giggles boyishly, holding it out. It's slick and bobbly and as warm as a human hand.

Boost made Manz look afresh at the way adidas made shoes. So did another technology he'd been working on: the robotic knitting method called Primeknit. Crucially, both processes not only produced new materials, but also combined previously separate operations. Whereas sports shoes were usually made from lots of separate pieces, Primeknit knitted the entire upper in one seamless whole. Manufacture and assembly were simultaneous.

At the same time, Manz experimented with "spider silk" manufacturer AMSilk's artificially produced fibres, which could be tuned to contain different properties on demand. He teamed up with Carbon, a Google- and General Electric-funded startup with a new version of 3D printing that used ultraviolet light to fix the shape of a design after it had been pulled, Excalibur-like, out of a vat of liquid resin. 3D printing had always been too slow, expensive and heavy to use in shoes. Now it seemed like it could finally be ready. (The technology is not currently in use at the Speedfactory, but adidas says it will make 100,000 pairs of shoes with 3D-printed midsoles by the end of 2018.)

These technologies all had something in common: they were flexible. Existing mass production only functioned if it produced large batches of identical products. (This, in essence, is why restaurants have menus.) The new systems, which sat under the collective term of additive manufacturing, could create 20,000 shoes and each one would be different. To Manz, it felt like a start of a seismic shift.

"We started developing some of the processes and we thought, well, actually you could put them together and they would make a pretty damn good factory," he says. The cost per shoe would be higher, but the savings in time and materials could compensate. (Nike claims that Flyknit reduces waste by 80 per cent compared to cutting and sewing.) It would be hard, but it would be different. Above all, it would be fast.

The moment of decision came in 2015 at a meeting with two of adidas's most senior executives. The first phase of the Speedfactory was a laboratory-style test of the different machines. To build it, Manz wanted €2 million (£1.8m) (an adidas source put that number closer to €10 million). "It was pretty steep," recalls Carnes, who watched on, nervously, as Manz made his pitch. "It was a big ask for a project they had no idea would work. And they agreed in the meeting. That was probably the coolest part." The money, however, came with conditions. Mindful of their own demand for speed, the board insisted the Speedfactory start production within two years.

Once, adidas might have hesitated, as it had over Primeknit, a technology it worked on for years, but released so slowly Nike ended up suing over breach of copyright. (A court eventually decided the two firms had come up with it at the same time.) Now, things were different. Adidas had Gaudio, it had speed - and, crucially, it had Kanye West.

The new era of adidas dawned on May 17, 2015, at the Billboard Music Awards in Las Vegas, when adidas collaborator West strode on stage to perform wearing a pair of pristine white Ultra Boosts. One viral image captured the moment: a shot of the rapper mid-leap, shoes flashing in the dry ice. Instantly, the trainers began to sell out online. "That was the moment where Boost left the pure realm of sports and crossed over into culture," says Gaudio. Kanye made that Boost famous. After years of irrelevance, adidas was cool again. Eight months later, the Speedfactory broke ground.

The Speedfactory is built on an old potato field on the outskirts of the medieval city of Ansbach. After kilometres of placid wheat fields and chalet-style houses, visitors pull up to a warehouse on an anonymous industrial park. Above the front door, instead of the Adidas logo, clumsy grey cursive spells out the words, “Oechsler Motion.”

Oechsler Motion started as a button manufacturer in the mid-nineteenth century. Today, it makes car parts across the world – one of the famous mid-sized firms that powers the German economy. Other local contributors (and automotive suppliers) include robot makers Bielomatik and Kuka, and industrial giant Siemens. With the Speedfactory, Adidas hasn't stopped outsourcing. It's just brought it closer to home.

At every stage, the Speedfactory is both more and less than it seems. Take automation: there are plenty of robots, sure, but also a good deal of old-fashioned manual labour. Next to the patching conveyor belt, a line of middle-aged women stitch uppers at sewing machines. After that, a man steams the parts by hand.

"To be honest, it's difficult to find those kind of workers now because this industry barely exists in Europe," says Steindorf, who oversaw the hiring. Adverts for experienced stitchers turned up women who'd worked in shoe production before it moved to Asia 30 years previously. (To refresh their skills, Oechsler brought in a cobbler who still practised traditional shoemaking.) But other positions couldn’t be filled in the same way, for the simple reason that the work involved wasn’t done anywhere else in the world. Workers had to learn their jobs from scratch.

The hope that technology might reinvigorate skilled manual labour is another reason people get excited about the Speedfactory. But while its jobs are better paid and have greater autonomy than their Asian equivalents, there are far fewer of them, making, at present, far fewer shoes. By 2018, the two Speedfactories will employ around 300 workers to produce one million shoes.

This is the politically unacceptable truth about reshoring, as the return of manufacturing from Asia is known: it doesn't bring back many jobs. Parkdale Mills, the largest buyer of raw cotton in the US, shut down in the 90s, killed by competition from China. Thanks to new textile machines, it reopened in South Carolina in 2010, producing 1.1 tonnes of yarn a week - but with 140 staff compared to more than 2,000 previously. Raspberry Pi's move from China to Wales created 30 jobs. Foxconn, which makes Apple products, promised 13,000 new jobs for Wisconsin, but only with $3 billion in government subsidies - $230,700 per employee.

In any case, production isn't moving that quickly. A recent Morgan Stanley report concluded that in five years' time, 90 per cent of sports shoes will still be made in Asia using traditional methods. Earlier this year, new adidas CEO Kasper Rorsted told [i]The Financial Times[/i] that large-scale reshoring was "a complete illusion". "And that goes for the entire industry, not just for adidas," he added.

Seen up close, the Made for London isn't nearly as customised as its name suggests. Adidas did consult five local runners on its design, but they were "influencers" rather than representative Londoners. (Together, the group, which included television presenter AJ Odudu and celebrity personal trainer Bradley Simmonds, had more than 250,000 followers on Instagram alone.) Some of their ideas made it into the final product: the green stripes on the upper; the choice of fabric; the reflective laces, which the Londoners insisted on for morning and evening running. But as the shoe's basic outline had already been determined, making key design decisions wasn't an option.

The reason was simple: starting from scratch is hard. Visions of a sci-fi future where our shoes are all 3D printed to our exact physical specifications sounds cool - but adidas knows that in reality, it's not what most customers want. In 2000, it launched mi adidas, a website where customers could change the look of certain shoes before buying. People loved the idea, but in practice, they didn't use it. "Everybody has found the same thing," says Carlos Cordon, professor of strategy and supply-chain management at business school INSEAD. "People are willing to pay a bit more [for customisation], but not much more, and not that many."

Cordon's research shows that Nike's rival ID website accounts for less than two per cent of the American company's sales. And it makes sense. Customisation takes effort: most of the time, the generic option is just easier. (This is the other reason why restaurants have menus.) "I'm a designer and I wouldn't use our customisation, or anyone's," says Carnes. "It's overwhelming. Just give me what I already want."

What makes the Made for London significant is that, for the first time, each shoe is digitally unique - "named", as Manz puts it. They contain an NFC chip near the tongue; inside, each sole is sprayed with a QR code near the heel. The machines that make it are mapped and tracked with sensors so they can stream data about the operations.

Digitising production brings many benefits. One is quality control. Another is planning: if Manz's team want to prioritise an incoming order of Stan Smiths, or change the plastic on a shoe upper, they can simulate the process using the Speedfactory's "digital twin". "It's a factory in the web," says Manz. "Traditionally, we would just set [the production line] up and see what happens," says Steindorf. "With the digital twin you can simulate consequences."

But the benefit adidas anticipates most keenly is data. This is the vision of a product that sends back data about how it is being used: a shoe that behaves like an iPhone. With that, adidas could finally peer inside the black box of its own product. Do people wear it? If so, what do they like? Most importantly, from a commercial point of view: what do they want next? "If you think about this right now," says Carnes, "already our shopping experience is giving algorithmic recommendations, but they're still offering you something that was created two years ago. What we're creating with Speedfactory is now you can actually get a recommendation for something that doesn't exist." He describes how adidas might be able to "take all of the visual information that you've customised already [online] and deliver a shoe. Eventually, a pair of pants. Without having to go into the detail of creating it and picking and choosing and all that stuff."

That vision won't be enabled by the Made for London's basic NFC chip and branded app. But in a small way it's already happening. In Russia, where adidas has more than 1,200 stores, it is using location and search data to manage stock. At present, store managers estimate which shoes and sizes to order based on past experience, the way retailers have always done. With the new model, Carnes explains, "You can correlate what's been ordered for a retailer with what we're selling online in that area with other things, like who's searching for which models." Location data might show there are lots of basketball players in the area; knowing this, the store can order larger sizes, fresher styles. Adidas can, for once, move at the pace of the world around it.

The true reason for the Speedfactory is just that: speed. In today's world, where you're more likely to order a pair of Yeezys from the back of an Uber than in a store, the gap between wanting something and expecting to have it has shrunk down to almost nothing. Companies working on the old 18-month supply chain are already behind. The last 40 years have seen commercial empires built on the cost savings produced by economies of scale. In this new world, what matters are economies of speed.

This is where the reshoring narrative goes wrong: for companies, what matters isn't location, but distance. The closer a factory is to its consumer, the faster the product can be delivered - and despite being more expensive to make, today's customer will pay more to have it sooner. (This, simplified, is the logic behind Amazon Prime.) And about those reshoring jobs: according to Morgan Stanley, 45 to 50 per cent of athletic footwear is sold in Asia. Local manufacturing means being local to those shoppers too.

The Speedfactory - and those like it - are not so much about the return of jobs or the rise of automation. They are about manufacturing catching up to the speed of the internet.

It's Friday afternoon in the Speedfactory, and the week's last batch of Made for Londons is almost finished. The soles are moulded, the uppers knitted, patched and stitched. Now the parts just need to be put together.

In a factory in China, this would be an unpleasant task. To assemble shoes, workers there use glue. Wearing masks to protect themselves from the fumes, they apply several layers, then press the parts together and wait for it to dry. In the Speedfactory, that process is eliminated - because the Made for London is not glued, but welded by lasers.

Seen in action, the welder resembles a high-tech blacksmith's forge. In a glass-fronted cabinet the size of a fridge, a yellow light surrounds a pair of upside-down trainers. The light fades, then glows again, brighter and red with heat. As the parts are effectively melted together, the smell of burning rubber fills the air. Then a burly, bearded man in a sweaty T-shirt opens the doors, takes out the shoes and sets them on a ledge. The team gather around to admire their creation. "In China, it's an imprecise process," says Carnes. "This makes sure everything is lined up before they touch."

Manz picks up one of the shoes. "It's consolidating now."

"It's actually pretty stable at this point already," Steindorf says.

"Very clean," says Manz admiringly.

He puts it down. The trio turn away. But the Made for London isn't finished just yet. Quietly, the bearded man takes the brand new shoe and does something only a human can do, something no robot is yet close to mastering. He turns it on its side and, with practised swiftness, threads in laces.

This article was originally published by WIRED UK