Tackling Fashion Industry’s Waste Problem

As mentioned in Fashion Retail Safari, New Retail is about new processes, new analytical capabilities, innovation, technology, new store formats, …, that respond to digital generation trends such as experience, sustainability and authenticity. We live in very volatile times – weather extremes, political distruptions (e.g. Brexit) and public health threats (e.g. COVID-19) – and today, more than ever, companies should embrace a business model based on digitalization and flexibility, as described in Fashion Technology Platforms or in my latest post about Fashion and Coronavirus (1/2).

Winter 2017 Stella McCartney campaign

Fashion Industry Waste has become a major concern for both consumers and governments which have increasingly begun pressuring brands to tackle this issue. Hopefully, Demand Planning with the help of AI, On-demand Production or Microfactories can help the sector to stop overproduction.

Excess production in the garment industry

The French Government is working on legislation to ban fashion brands from throwing away or destroying unsold clothes by 2022, announced Edouard Philippe on the 4th of June 2019. According to the government, about 800 millions euros of unsold goods are thrown away each year.

“Too many companies feel OK with just throwing away or destroying the shoes or the clothing that haven’t been sold,” French Deputy Ecology Minister Brune Poirson said at the recently concluded Copenhagen Fashion Summit. “You can’t do this anymore. It’s shocking.”

The problem of excess product is endemic in the garment industry. It has led to constant discounting, dumping of unsold clothes to lower-income countries and in the worst case, stock destruction i.e. burning of perfectly wearable clothes. According to the Pulse Of The Fashion Industry Report 2019, clothing production is expected to hit 102 million tons by 2030. That’s a whopping 64.5% rise from 62 million tons in 2017.

While France has taken the lead on tackling the challenge of overproduction and waste, one can expect other countries to proceed with similar legislative measures. That translates into a paradigm shift for the fashion brands. They will have no choice but to change the way they operate and address the question ‘how much to produce’ with vigour.

“They will have no choice but to change the way they operate and address the question ‘how much to produce’ with vigour.”

Fashion waste and artificial intelligence

Personalized Production and Technology to fight clothing waste

Indeed the ask is challenging as brands need to respond to trends and changing customer demands in a fiercely competitive landscape. However, companies now have an opportunity of using technology-based tools to find answers that are backed by precision. Such tools also bring more transparency to the global supply chain and enable personalized production. By 2025, personalized luxury goods will grow by 130% thanks to demand from millennials (source: CB Insights).

This new ecosystem of fashion production and stock management is more connected. Artificial intelligence (AI) and advanced analytics help with breaking down of the silos of the traditional organisation and bring together the design, production, sales and marketing teams to work towards improving end-to-end efficiencies, supply chain flexibility and higher sell-throughs.

Fashion waste and artificial intelligence_2

Trend Forecasting and Demand Planning

The product journey begins with trends that are integral to the nature of fashion. But due to the pace of social-media driven trends, companies often miss weak signals that could lead to faulty trend forecasts and product assortment plans. It is estimated that AI can reduce forecasting errors by up to 50% as well as reduce total inventory by 20-50% as per the 2017 McKinsey study, ‘Smartening Up With AI’. Heuritech is a personalized market intelligence platform that helps with accurate trend prediction and sales forecasts to enable stock planning. Its AI mines Instagram for trends that are emerging amongst the target audience. And that data is correlated with the brand’s past sales to optimise production and demand planning. The Heuritech tools are applicable to large as well as small-batch production i.e. planning a one-year production timeline or one that is much shorter.

Fashion waste and artificial intelligence_3.jpg


For example, if data suggests that the best-selling plaid coat from last year is slowing down, the level of demand for that trend in the coming season is calculated and fed into the brand’s product assortment for the future.

For luxury products, once a product is brought to market, the dynamic Heuritech dashboard tracks Instagram to measure the desirability of new launches. That information helps a brand adapt its stock production and offer relevant product quantity at the store. For brands that produce 6-8 collections annually, yet in small numbers, Heuritech can predict an upcoming successful trend and calculate the quantity to produce.

On-Demand Production

Meanwhile, to respond to changing customer preferences, the on-demand production model is gaining traction. It marks a change from the traditional paradigm where products were ‘pushed’ into the market based on predictions made by design and merchandise teams. If the products are rather ‘pulled’ into the market based on actual demand, the outcome is smaller inventories and more operational flexibility.

Nineteenth Amendment is a retail platform cum production management company that specializes in on-demand manufacturing in the USA. Its marketplace showcases work by independent, emerging designers as pre-order items. No stock inventory is held and only when a customer places the order on the site does the manufacturing process begin, going all the way from fabric sourcing to final shipping. The item is produced in the US by a factory partner of the Nineteenth Amendment network and hence the turnaround time is less than six weeks. Also the factories ask for no or low SKU order.

Unmade is another SaaS solution that helps brands offer customisation to the consumer on their online orders before the actual production takes place. Unmade’s technology integrates individual short-run orders into the brand’s existing production line.

Fashion waste and artificial intelligence_4.png

The on-demand model brings agility but production costs tend to be generally higher. It also depends on whether the supply chain infrastructure is geared to handle shorter  turnaround cycles. However, big players like Zara and Asos are on-board with this model and use real-time, customer-centric data to be flexible with their inventory.

Source: https://algorithms-tour.stitchfix.com/#inventory-management


When fashion brands produce in large quantities, the unit cost comes down. In the fast-fashion model, typically high-volume-low-cost orders are churned out with a longer lead time at a traditional factory that carries a significant resource footprint. In contrast, the microfactory is a new business model that empowers brands to be fast, nimble and sustainable with their process of making.

The genesis of microfactory lies in the digital textile revolution where developments like Computer-Aided-Design and Manufacturing software (CAD/CAM), digital patterning, laser cutting and digital textile printing have made it possible to produce less but produce better and with a high degree of personalisation. Here’s an example of a sportswear giant trialling the concept. In 2017, Adidas set up a pop-up shop in Berlin offering a customised sweater suited to one’s size and taste in just a couple of hours. To achieve that result, a body scanner had been linked up with an augmented reality pattern creation software and knitting machine.

Large players are looking at microfactories to test prototypes and for customised high-speed production capability. Such facilities could also be shared by small designers to produce their garments locally and with low minimums.

Fashion waste and artificial intelligence_5

Change won’t happen overnight but the tech-driven capabilities will enhance the speed of transformation and increase the return on investment. With data analysis available at a city-level and hyper-local supply chains closer to reality, a brand can potentially offer micro-targeted products. The bottom line is that the entire industry, from mass to luxury, have to collectively invest in a more progressive and sustainable way of producing apparel and footwear to reduce waste.

5 responses to “Tackling Fashion Industry’s Waste Problem”

  1. Very interesting ideas and I really appreciate what you’re doing here as a citizen of the world. If all of us could think like this, the world would certainly be a much better place 🙂

  2. […] inventory as set out by Rolnick et al in Tackling Climate Change with Machine Learning and Alfonso Segura Tackling Fashion Industry’s Waste Problem. Furthermore, the pandemic serves as an opportunity for us to consider the needs of our society and […]

  3. […] resulting in unsold inventory Rollnick et al in tackling climate change with machine learning and Alfonso Segura to tackle the fashion industry’s waste problem. In addition, the pandemic serves as an opportunity to consider the needs of our society and the […]

  4. […] resulting in unsold inventory Rollnick et al in tackling climate change with machine learning and Alfonso Segura to tackle the fashion industry’s waste problem. In addition, the pandemic serves as an opportunity to consider the needs of our society and the […]

Leave a Reply