The fashion industry has long been a crucible of technological innovation. The Jacquard loom created to mechanize the production of textiles in the early 1800s became a key milestone in the development of early computing. Punched cards were employed to choreograph the generation of complex embroidered patterns in a process resembling a form of physical rasterisation:
Not long afterwards, the sewing machine was born and with it an early instance of the visceral fear of technological unemployment:
The first functional sewing machine was invented by the French tailor, Barthelemy Thimonnier, in 1830. Thimonnier's machine used only one thread and a hooked needle that made the same chain stitch used with embroidery. The inventor was almost killed by an enraged group of French tailors who burnt down his garment factory because they feared unemployment as a result of his sewing machine invention.
Today fashion is worth $1.5 trillion annually globally. The industry is once more on the cusp of a revolution as advances in automation, software and robotics accelerate new approaches. Production processes and methods are becoming progressively more efficient while offering consumers greater choice than ever.
The fashion supply chain can be broken down into four broad verticals namely Design, Manufacture, Inventory and Distribution. Each are being disrupted by different technologies.
Design
The role of the fashion designer is being changed by Artificial Intelligence (AI) both in terms of the creative process itself and how intelligence around fashion trends is discovered. As individual designers become increasingly overwhelmed by the demands put upon them through an explosion in data, AI can be used to support them in making sense of it all. Project Muze, a 2016 collaboration between Google and Zalando exemplifies the state of the art. A deep learning algorithm was developed by the project team to create personalised designs based on specific user interests. These were aligned with style preferences understood by the network which in turn were generated from information gathered from a panel of industry experts:
the team used Google’s open source platform TensorFlow to train a neural network – an algorithm modelled on the human brain – with 50,000 data sets based on the style preferences of more than 600 fashionistas. Over time, the code learned to predict what kind of fashion people might like and creates designs that match them.
Unfortunately the results weren’t quite as compelling:
More recently a Chinese AI proposition called DeepVogue from Shenlan technology won an award at an International Fashion Design innovation competition using deep learning to produce original designs drawn from images, themes and keywords imported by human designers. Technologies like DeepVogue once perfected and scaled could render the human element in design obsolete. Perhaps the reception from those it threatens to displace will echo that of the French tailors two hundred years ago to the sewing machine:
Less controversially, AI is now capable of providing recommendations and intelligent predictions about trends to fashion brands and designers by processing information provided by customers, experts and other sources. The Reimagine Retail collaboration in 2018 in New York between IBM, Tommy Hilfiger and the Fashion Institute of Technology (FIT) is a good example. The intent of the project was to explore the use of AI “including computer vision, natural language understanding, and deep learning techniques specifically trained with fashion data” in assisting brands with trend analysis.
“AI can assist design teams by enhancing and reducing overall lead times, and expand their creative discovery by analyzing and remembering insights from thousands of images and videos using computer vision. These designers can also more easily find how they can integrate trending colors, key patterns, and styles … It is the critical difference between systems that enhance and scale human expertise, rather than those that attempt to replicate all of human intelligence. The AI inspired us and breathed new life into the design process”
Manufacture
The fashion industry traditionally followed a cycle built based around seasonality. That paradigm has now been supplanted by fast fashion which has fuelled an instant gratification culture:
Fast fashion can be defined as cheap, trendy clothing, that samples ideas from the catwalk or celebrity culture and turns them into garments in high street stores at breakneck speed to meet consumer demand.
The ability for brands to respond with just in time (JIT) manufacture has become an important business driver. Spanish chain Zara are among the market leaders. As soon as they detect a new trend they can pivot their supply chain to bring it to market faster than their competitors.
The aim of [the'] company is to give society what it wants as quickly as possible. Zara therefore changes its collections every two weeks. To put this strategy into perspective, look no further than Mango or Gap: these competitors renew their collections about six times a year. Although most fashion experts consider Zara’s model to be unsustainable—due to its supposed high cost—the company has managed to grow steadily year after year, leaving its competitors in the dust.
This ability to rapidly innovate and produce new clothes has mean the costs of starting a new fashion brand have plummeted. It is now possible for startup labels to leverage small batch manufacturing as a service in New York through propositions like Maker’s Row.
3D printing enables brands create fast fashion personalised products just in time and provides them with new avenues for customisation. Sneakers have been a particular target for this technology. Adidas’s Carbon 3D-printed soles for instance are personalised using biometric data. Gucci Art Lab in Italy is also trialling a more agile, multidisciplinary approach. Billed as a “futuristic centre of industrial craftsmanship and experimentation, a laboratory for leather goods and shoes” this colourful establishment fuses elements of rapid prototyping with tradition.
The manufacturing process itself is also being disrupted by automation and robotics mechanising the process of making clothes. Cutting has proved easier to automate than sewing to date. However Atlanta-based Softwear Automation are seeking to change that. Their “sewbots” have robot arms, grippers and “micromanipulators” that guide threads with precision. Cameras and computer vision software can track individual threads. In February 2019 Softwear announced a sewbots-as-a-service proposition very much in tune with a growing appetite for reshoring in the US:
a rental lease service to allow manufacturers, brands, and retailers to source and manufacture here in the US at a lower cost than outsourcing and with greater predictability and quality. While we understand the benefits of “Made in America”, the focus of this program is to offer US textile manufacturing more control, greater margin, faster turn times and less inventory.
The spectre of technological unemployment again looms large in the background. This time its at the other end of the economic spectrum in the developing world where sewing is an important source of income for many:
Fast fashion assisted by technology has boosted productivity and output. However these gains have come at great cost and have been fiercely criticised:
In “Fashionopolis,” Dana Thomas, a veteran style writer, convincingly connects our fast-fashion wardrobes to global economic and climate patterns and crises, rooting the current state of the fashion biosphere as a whole — production methods, labor practices and environmental impacts — in the history of the garment industry.
The industry is currently a major contributor to our collective impact on the environment largely as a result of the rise of fast fashion. According to the World Bank:
The fashion industry is responsible for 10 % of annual global carbon emissions, more than all international flights and maritime shipping combined. At this pace, the fashion industry’s greenhouse gas emissions will surge more than 50 % by 2030.
There is growing scrutiny on the long-term sustainability of the industry raising questions for the likes of Zara on the ethics of producing so much new clothing:
One reason fashion is such a blight on the planet is that we’re simply consuming too much of it. The apparel industry produces more than 150 billion clothes every year for only 7 billion humans. Over the last 15 years, clothing production has doubled, and yet the average number of times a garment is worn before it is discarded has decreased by 36%. Many clothes are worn just seven to 10 times before they’re thrown out.
Sustainability concerns resonate especially strongly with the Generation Z cohort with evidence that they are willing to pay more for sustainable products. Brands ranging from Everlane, Levis and Selfridges with their Project Earth eco-friendly clothing service have all sought to emphasise their ethical credentials. Resale is another big growth area. Providing technology that enables people to buy and sell unwanted or used clothes genuinely reduces the impact of fashion manufacture on the environment so should be universally welcomed. In the pre-internet era this exchange was the preserve of thrift shops and market stalls in locations like Portobello Road and Camden in London. That tradition has been increasingly disrupted by online resale. Depop are a stand out in this space, a fashion marketplace targeting Generation Z and Millennials. Through a unique mixture of effective branding, eCommerce and a strong mobile-first presence they have been able to capture what appeals to those cohorts:
Depop is meme-driven, it’s about rave culture, it’s nostalgic and ’90s. Depop shoppers are more interested in what their peers are wearing than what’s happening in the fashion industry. And they’re most likely to shop clothes by designers who are from — or influenced by — the ’90s and the 2000s.
The proposition is increasingly compelling to mainstream brands too:
Fashion brands are increasingly taking to resale platforms, both to compensate for the current lack of physical sales channels and to minimize losses on excess inventory. In doing so, they’re also getting in on the sustainability conversation at a time when brands’ processes are under a microscope.
A key aspect of depop’s success is the ability for buyers and sellers to directly message each other. This memorialises the kind of colourful haggling banter that was common in the old physical world of the flea markets. The resultant memes highlight a new kind of democratisation of fashion. Sites like @depopdrama celebrate with fantastic examples of real exchanges. Resale is likely to segment in further niches in time. The Vestaire Collective based in Paris for instance have already specialised their focus on high end clothing resale. The market is set to grow five fold between 2019 and 2024.
AI is also being used to monitor supply chain risk. Interos leverage Natural Language Processing (NLP) to scan a variety of sources of information to build up their risk scores for participants.
Inventory
Technologies like RFID will make clothing inventory more traceable. Goods produced using JIT manufacturing methods using RFID tags can be tracked and distributed using sophisticated inventory management tools. Brands like Burberry and Zara are already applying RFID labelling in their shops and on the manufacturing floor respectively. Zara’s parent group Inditex uses RFID signals to trigger stock replenishment in store. At the high end French luxury outfit Moncler has been labelling all its products with RFID since 2016 in an effort to combat extensive counterfeiting operations in China.
RFID could also serve as a on-ramp for blockchain technology to audit fashion supply chains. By giving a product a unique token on a distributed ledger, it is possible to build a complete unmodifiable digital history for it. Blockchain startups like Provenance in the UK have already piloted these approaches typically in the context of high-end items.
Distribution
Many fashion brands over the last decade have pivoted to a direct to consumer (D2C) distribution model aided and supported by online technology. Digital-first outfits like Boohoo and PrettyLittleThing create their own markets aimed at young women through effective use of social media. Everlane seek to differentiate themselves from the competition by providing transparency around supply chain and pricing. Brand consistency is a key part of the messaging. So too is leveraging SEO with great value attached to ranking high in Google search results. Instagram, TikTok and Facebook are key social platforms used by these brands for online marketing and targeted advertising. Pinterest is frequently used to curate interesting looks. Customer interaction and behaviour is used to inform decisions on what to design next. Once their new designs are unveiled, brands cultivate a See Now Buy Now sense of urgency across the different touch points. Actively encouraging consumers to buy what they see right away raises some ethical concerns.
The rise of eCommerce has led to lots of dire predictions that it will accelerate the decline of brick- and-mortar stores for retail distribution. Covid has only intensified that existential pressure on physical outlets like shopping malls. While there is undoubtedly a trend to downsize physical store space, shops will remain a key part of the D2C distribution proposition for many fashion brands as demonstrated by Levis:
many traditional retailers [are] pivoting to a direct-to-consumer model to better serve their customers. Levi's is one example. The company is pulling back on its wholesale business as DTC sales grow. With the growth will come about 100 new company-operated stores.
Subscription boxes and pop-up shops offer new channels for distribution. Bulletin go further and offer brands seeking floorspace a WeWork style model with flexible options to rent space in existing stores.
As the internet has disrupted traditional models of ownership, another innovation in Distribution has emerged in the form of Clothing-as-a-service (CaaS) providers. These companies support users who want to pay for usership rather than ownership. Key to their rise in prominence is the growing consumer desire for variety and sustainability at an affordable price. CaaS propositions also address a key problem with online shopping for new clothes, namely returns. As much as 40% of online purchases are returned many of them due to fit and finish issues. Rent the Runway is one of the biggest CaaS players. Their service allows members to rent designer fashion items either on a per-piece basis or on subscription.
Brands have to balance emerging trends in the digital world with a need to execute in the real world. This requires them to revisit how they address omni-channel experiences. Users increasingly expect to shop across multiple touch points in a way that is convenient for them. Physical stores can help build and strengthen this connection. Augmented and Virtual Reality (AR/VR) technologies help further blur the boundary between online and in-store. Obsess offer retail brands the opportunity to “digitize your retail store and make it shoppable online”.
AI-based digital assistant stylists in the form of chatbots can provide feedback on outfit choices and suggest alternatives based on their understanding of user preferences and history as well as current inventory on offer. Digital assistants like this can also be personalised. For instance Levis use the mode.ai platform to support their Virtual Stylist in helping users find the right pair of jeans. Uniqlo have magic mirrors in their stores that allows customers to see how clothing they try on will look in different colours. Over twenty years ago in his celebrated book Visions, Michio Kaku predicted precisely this would happen by 2020. As visual search and recommendations systems improve, users have the opportunity to send photos of items they like to chatbots and receive recommendations. Product discovery propositions like Syte and Snap Tech offer a visual search components as a plug-in to existing retailer mobile applications allowing customers to copy a particular look as a service.
The Future
The need to support an omni-channel presence across a growing range of digital and physical touch points will continue to be a major driver for retail brand innovation moving forward. Incumbents may find they need to undertake significant digital transformations to ensure their technology platforms are capable of supporting the rapid construction of new responsive digital propositions spanning multiple device formats and platforms concurrently. In technology terms, this typically requires moving to more standardised software architectures based on microservices at the backend and React micro-frontends on the client side. There are other requirements besides. Online brands have vastly more daily digital contact with customers. It is vital they use automation to support key functions like customer support and content moderation as part of scale growth. Doing so ensures they avoid the cost and complexity of scaling the corresponding manual processes and organisation to support business needs. This form of investment in operational excellence is particularly vital to get right for the marketplaces of the future.
Important though they are, these challenges are really just forms of sustaining innovation. Looking beyond the evolution of the core digital platform, there are other tech trends to be aware of that are more in the domain of disruptive innovation. These include novel materials like Bolt Threads’ spider silk and Modern Meadow’s lab-grown leather, wearable technology that can be updated like Google’s Jacquard Threads, self-repairing materials as well as the creation and trading of virtual goods and 3D body scanning. The latter offers the prospect of comprehensively solving the fitting problem.
Bringing together the elements outlined in this article we have a new way of mapping the fashion value chain shown below. The line here separates currently viable innovations from more nascent propositions:
More important than these technological advances is the necessity of cultural change. The tech giants of the current era have found that their corporate cultures are too often incompatible with fixing the societal problems they are causing. The leading fashion brands of the future regardless of business model will need to pay attention and be much more responsible and engaged around sustainability, diversity, equality and social responsibility given their demographic. They may end up leading an exciting shift away from consumption to creation by helping to place the means of fashion production at the disposal of participants. Can they show an alternative way of harnessing technology for social good? Prepare to stand by for more creative destruction in the future of fashion.