Algorithms for Fashion. How Machine Learning Makes the Apparel Industry More Sustainable
In the fashion industry, return rates – sometimes referred to as “the plague of eCommerce” – are skyrocketing. Besides placing a burden on the companies’ budgets and logistics, this behavior is also detrimental to the environment. However, innovative software programs based on Artificial Intelligence and Machine Learning have the potential to reduce return rates, leading to a more sustainable industry.
The fashion industry is frequently accused of unsustainable practices, with its high carbon emissions, consumerism, high water consumption, toxic effluents and materials, child labor, and breach of human rights being just a few examples. While predictions regarding the industry’s capacity to meet decarbonization goals look quite somber, the industry has launched numerous initiatives to promote a sustainable agenda. From new business models and innovative materials to new production processes, green distribution, and packaging alternatives, these efforts are showing great promise, even if they are still in their infancy.
Isabelle Ohnemus is the founder and CEO of EyeFitU, a Swiss startup at the confluence between fashion and technology. Highly concerned by the increasing return rates in retail and e-commerce, she and her team developed a Size-as-a-Service AI software that aims to address retailers’ trillion-dollar problem while simultaneously helping them reduce their carbon footprint..
Returns – a major threat to the sustainability of fashion industry
E-commerce is expected to increase its market penetration from 46.6 percent in 2020 to 60.32 percent in 2024. While between eight and ten percent of products purchased in stores are returned, the numbers rise to a whopping 40 percent for online purchases.
Ohnemus points out that consumers behave very differently in the e-commerce environment than in brick-and-mortar stores. Clients buy clothes to wear on special occasions and then return them, purchase items only to try them on, without any intention of keeping them, or order multiple sizes and colors only to pick the one that fits best.
“Most of the time, companies are not even able to tell you what happens with the items you just returned. In most cases, they end up in a landfill even if they are in perfect condition because the processing costs are too high for the company. And that’s after they have been shipped over multiple countries or even world regions.” – Isabelle Ohnemus, founder and CEO, EyeFitU
The consequences are noticeable on all fronts: Retailers have to deal with increased shipping and fulfillment costs, disturbed supply chains, and logistical complications, and require more storage space as well as human resources to handle the returns.
Example of a retailer’s returns operations, adapted after Cullinane, Browne, Wang, and Karlsson (2019)
Shipping parcels between postal offices, hubs, consolidation centers, sorting departments, and repackaging locations takes a toll on the environment. In addition, large quantities of items that aren’t included again in the retail process add up to the millions of tonnes of textiles already in landfills.
Machine Learning algorithms could help the apparel industry
Since one of the main reasons for returning clothes is that they are the wrong size or fit, Ohnemus partnered up with AI and machine-learning specialists to develop software capable of helping customers find the best suitable garment for their body type and preferences. The app can be used both in brick-and-mortar stores and in online environments and offers in-store contactless sizing, as well as QR code and smart mirror integration.
With one of the largest databases of body measurements per region, the software uses machine learning, user-generated data, and multi-parameter algorithms to determine the correct clothing size for a specific person, taking into consideration various body shapes and sizes. After its implementation, clients reported a decrease in returns of up to 55 percent, with significant increases in conversion rates and average order value.
Final thoughts on sustainability
However, like many other industries, fashion is also prone to greenwashing. Customers and stakeholders are overwhelmed with misleading messages regarding a company’s environmental performance, being led to believe that the firm really cares for the environment. Given the vast proliferation of such deceptive self-promotion practices in the last decades, it is becoming increasingly difficult for customers to differentiate between true claims and mere marketing messages.
“I don’t think that any company today buys software for the environmental impact. They might love the sustainability story and the PR impact, but actually what sells the product are its KPIs – decrease in returns, increase in basket size and conversion rates.”
Even though the reality behind corporate claims of environmentalism may sometimes be disappointing, the scale and pace of current innovation and technological developments can bring about a significant shift towards new, more sustainable business models that are geared not only toward profit, but also to the needs of the planet and its inhabitants. Access to the right data, gathered through automated processes and processed by powerful analytical tools, might be the first step towards truly sustainable brands.
However, as Ohnemus points out, behavioral change has to come from both sides – from companies and consumers. Technology can support this change by providing consumers with tailor-made solutions to ensure their online fashion purchases satisfy their needs. Reducing returns is a three-way win-win-win situation: Good for the industry, good for consumers, good for the climate and environment.
Find out more on the topic in our free reports dedicated to sustainability.