Adidas Speedfactory: Impossible is Nothing with Automation
With fast fashion, consumer demand and preferences change daily. So far the sports shoe industry, an $80Bn market, has not kept up.
Can the Sports Shoe Industry Keep Pace?
The timeline of a sneaker, from the initial sketch, prototyping, testing, material ordering, production, and finally shipping takes 18 months . 75% of sneakers, however, are on sale for less than a year . A sold-out sneaker can take 2 to 3 months to be restocked, by which point demand may disappear . Simply put, the sports shoe industry has not kept pace with fast fashion and changing consumer demand.
In addition, rising prosperity in Asia, where sneaker production has typically been outsourced, has resulted in higher labor costs. These factors demand a revolution in sneaker production and supply chain management. Adidas, like many other German manufacturers, has brought production back home as part of a wave of Industry 4.0, a push by German companies to remain innovative and competitive with automated “smart” factories and shortened supply chains .
The Race is On
Adidas is building two Speedfactories, one in Ansbach, Germany, and one in Atlanta, Georgia. Adidas’ Speedfactory is one of the new “smart factories” being built in Germany. The shoe design process is digital, with shoes styled on a computer and then tested for fit and performance. In parallel, another computer will test the production process to make it as efficient as possible. The factory will make the majority of the materials in house, using raw goods such as plastics and fibers. Once the design is complete, the machines will use computerized knitting, robotic cutting, additive manufacturing, and 3D printing to quickly assemble the shoes directly from the computer’s design. This minimizes the setup times and switching costs that typically come with producing multiple shoe designs, which can save hours, if not days, in labor time. The new production system will allow for speed and customization, with the design process taking 6 months, one third of the typical time, and production taking days with the hope of eventually reducing it to hours .
Because the Speedfactory creates the materials and controls the complete process, it will use data-driven production to “minimize waste, remove glue completely from the product, increase energy efficiency, and significantly reduce the carbon footprint of all product elements” . The data and process can be visually shared with the consumer, increasing transparency and consumer involvement with the design and production.
The factory will not be fully automated, with certain precision jobs still being completed by humans. The Speedfactory will create 160 production jobs, versus the typical thousands in a factory in Asia . Though current output per factory is small (500,000 shoes out 301MM sold in 2016), Adidas hopes that by bringing the shoe closer to consumers, it can increase speed of its supply chain and flexibility in production . The factories can also produce batches as small as 500 shoes (versus the usual 50,000 – 100,000 pairs), enabling Adidas to better control its inventory and react to consumer demand . If these factories are successful in Germany and America, Adidas will roll them out to more geographies .
Using its German Speedfactory, Adidas is launching a limited line of 6 sneakers made for 6 cities (Paris, Los Angeles, Shanghai, Tokyo, London, and New York). Adidas traveled to these cities to collect data on the habits of local runners (running routine, sensors for their strides, gait analysis using motion capture) . As a Speedfactory design, these shoes leverage data on fit, function, biometrics, personal preferences for materials and colors, and usage requirements based on geography and weather. Using simulations and computer testing, Adidas iterated and improved on previous designs based on user data. Accordingly, Adidas used the data to create small batches of city-specific shoes. The London shoe, for example, is more reflective and rugged, because Londoners love to run to work, often in the dark with lower visibility and in rainy conditions .
The Finish Line
While this new line is generating buzz, Adidas needs to determine whether its Speedfactories are actually scalable, the company’s timeline for expansion, and the changes required in its supply chain and delivery to match the factories’ new speed.
Adidas also needs to make sure it keeps pace with its top competitor, Nike, who is experimenting with computerized knitting to make high-tech fabrics like Flyknit and automated production that can assemble the upper part of a show 20x faster than a human employee .
As these sports companies race to automation, they need to remember and re-think the human component in forward-thinking design and technology changes. Will our future be fully automated?
Word Count: 774 Words
 “Adidas’s high-tech factory brings production back to Germany,” Economist, January 14, 2017, [https://www.economist.com/news/business/21714394-making-trainers-robots-and-3d-printers-adidass-high-tech-factory-brings-production-back], accessed November 2017.
 Christopher Alessi & Chase Gummer, “Germany Bets on ‘Smart Factories’ to Keep Its Manufacturing Edge,” Wall Street Journal, [https://www.wsj.com/articles/germany-bets-on-smart-factories-to-keep-its-manufacturing-edge-1414355745], accessed November 2017.
 “Enter the Speedfactory”, Mashable, [http://mashable.com/2017/11/03/adidas-speedfactory-future-tech/#LtDxMg8g0ZqH], accessed November 2017.
 David Meyer, “Why Adidas is Turning to Robots in Germany and the U.S.,” Fortune, [http://fortune.com/2016/05/25/adidas-robot-speedfactories/], accessed November 2017.
 Jörn Poltz, “Adidas to Return Mass Shoe Production to Germany in 2017,” Reuters, May 25, 2016, [https://www.reuters.com/article/us-adidas-manufacturing/adidas-to-return-mass-shoe-production-to-germany-in-2017-idUSKCN0YG0X6], accessed November 2017.
 Stephanie Pandolph, “Adidas uses Speedfactory to localize shoe designs”, Business Insider, [http://www.businessinsider.com/adidas-uses-speedfactory-to-localize-shoe-designs-2017-10], accessed November 2017.
 Jake Woolf, “Adidas Is Changing the Way It Makes Sneakers—Again,” GQ, October 5, 2017, [https://www.gq.com/story/adidas-made-for-london-speedfactory], accessed November 2017.
 Richard Weiss, “Adidas Brings the Fast Shoe Revolution One Step Closer,” Bloomberg, [https://www.bloomberg.com/news/articles/2017-10-05/adidas-brings-the-fast-shoe-revolution-one-step-closer], accessed November 2017.
Student comments on Adidas Speedfactory: Impossible is Nothing with Automation
Like you, I’m also curious how the Speedfactory concept will impact other parts of the supply chain (e.g. shipping) and just how scalable this production model is.
Some aspects of the Speedfactory seem feasible and worthwhile to replicate. I’d think that digitizing the design process, for example, should be relatively easy to do at any Adidas factory. Similarly, constructing raw materials in-house can be an effective method for reducing waste and thereby costs (a la Asahi).
On the other hand, flexibility in a complex production facility usually implies a trade-off with unit costs. In that light, I wonder if Speedfactories should ever dominate Adidas’ production capacity. For the near- and medium-term future, I predict the Speedfactory will retain a niche role in Adidas’ supply chain, satisfying small production runs or augmenting production capability for under-supplied SKUs.
The scalability concern you bring up seems very relevant to me. What stood out to me was that Adidas decided to do this itself, by owning the factory rather than doing a pilot in collaboration with one of its manufacturing partners. Going forward, if they do expand this concept, will they put more investments in manufacturing and decide to own their own plants? Traditionally, apparel and shoe brands have shied away from owning their production facilities to the large capital investment as well as management bandwidth that entails. It also seems like, in order for this model to be successful, real time sales data need to be shared at the factory level (in order for them to project demand and get raw materials inhouse early). Therefore, if the final goal is to eventually outsource this model of manufacturing, will the company be willing to share such sensitive data? This could be a big barrier to scalability of this concept.
Loved this article! It’s very interesting to see that even athletic shoes are following the trend of fast fashion and are grappling with matching their supply chain to customer demands.
The question I have regarding this operational change is the impact on the rest of Adidas’s supply chain. If Adidas is working to shorten their lead times and move closer to a just in time delivery model, then its customers must also be willing to take less inventory, thus trusting that Adidas can respond quickly enough to any changes in demand. Is this a reasonable expectation? I tend to think not. Stores feel the impact of short supply first and are likely unwilling to take risks with their inventory. In this case, I would expect to see a bullwhip effect traveling back to the producer. For automation of its facilities to have the impact desired, I would recommend Adidas first work to bring the other components of its supply chain up to speed. For instance, its raw material suppliers should expect more frequent orders and the potential for more urgent shipping. Adidas’s distributors and retailers will also need to be comfortable with holding less WIP and requesting orders based only on demand, not inventory management.
The second question I have is, what is Adidas doing to better manage the demand side of the equation. While the ability to produce smaller batches helps Adidas meet customer demands, I think Adidas should also do more to work with retailers and data aggregators to better identify leading indicators of demand shifts. Like in the Barilla case, it profoundly confuses me that the producer would not have sufficiently accurate consumer demand data to avoid supply crunches like the ones that necessitated automation of Adidas’s factories in the first place.