What the World’s Biggest Brands Aren’t Doing for Sustainability

Inside LVMH’s annual sustainability report, & how it is and isn’t leveraging technology.

Last month, I had the absolute pleasure of attending LVMH’s annual sustainability conference in New York City, “Future Life”. As one of the biggest presences in an industry notorious for its environmental detriment, all eyes are on LVMH with regards to how fashion should and could become more sustainable. The fashion conglomerate has made promising press with its efforts so far, most recently by naming Stella McCartney, best known for her brand’s devotion to sustainability, CEO Bernard Arnault’s environmental advisor. 

My primary intrigue regarding attending Future Life revolved around better understanding how fashion companies are using technological innovations (or buzzwords like AI/data) to aid their sustainability efforts. I’ve previously written about my interest in image recognition tools as they pertain to better trend prediction, thus resulting in more accurately targeted production cycles and, consequently, less waste within a company. This image recognition fad has been picked up by younger, tech-savvier brands such as H&M, who recently reported that AI has “already helped predict trends & allocate garments to stores.” This is great news, especially considering H&M loses over $4 billion a year from unsold clothing. 

So, how is the world’s biggest fashion & luxury conglomerate integrating technology into its sustainability efforts? Most importantly, in which ways might its use of technology be lacking?

The Report

According to the report, LVMH invests $11.3 million into sustainable efforts– a significant number until you realize it’s a mere 0.0001% of its CEO’s net worth.

LVMH’s sustainability report highlights changes and statistics mainly revolving outside of the production process. For example, LVMH states a commitment to animal well-being, specifically citing its creation of the “first standard in the world for the responsible supply of crocodile leather”, perhaps in response to PETA claims the conglomerate was abusing the animals. Additionally, in its statement regarding reduction of CO2, the report reads that LVMH will “[adopt] low-CO2 transportation with the development of rail transportation for Hennessy.” 

In fact, an entire segment of the report is dedicated to transportation initiatives in CO2 reduction. Transportation in the fashion industry has recently become a topic of intrigue, notably with Gucci stating it would offset the transportation emissions of all its guests, models, and production team for its fashion shows. How will Gucci do this? Through donating millions to programs which eliminate equivalent emissions elsewhere. Claps to Gucci, for essentially changing none of its behavior and solely making tax-deductible donations.

However, fashion’s detrimental carbon footprint does not primarily stem from transportation-related causes. The textile production industry alone produces more greenhouse gas emissions than “all international flights and maritime shipping trips combined.” Thus, while any carbon-reducing incentives are positive, high-fashion houses and conglomerates should focus the majority of their sustainability efforts on the eye of the hurricane: the actual production of goods.

Such seemingly small changes are a significant step for an industry commonly criticized for its treatment of animals and monumental carbon footprint. However, these incentives seem to address fixing what the conglomerate has already been doing wrong, instead of steps which could propel its brands above and beyond competitors.

The Technology

LVMH reported frequent uses of technological tools during its product production process. For example, the report states the use of Edibox, a cloud-based tool for “measuring the environmental efficiency of packaging”. Edibox, in essence, is a tool to digitize receipts with and allow companies to process & analyze sales from receipts with “no paper, no mailing, no CO2, no hassle”.

The conglomerate’s most impactful use of technology, however, pertains to its incentives to have 100% of its waste reused, recycled or converted to energy by 2020 (the report boasts 91% performance in 2018). In order to achieve this goal, LVMH states its use of CEDRE, an in-house recycling platform which (in 2018) treated over 2 tons of various waste products and utilized “selective sorting streams” such as glass, cellophane, cardboard, and plastic. 

Though, again, managing waste after inventory has been produced is only one small portion of the fashion industry’s sustainability problem, and an advanced recycling system is not a technologically savvy nor advanced tool in today’s age of AI & analytics solutions. 

Edibox, unfortunately, was really the only cloud-based software mentioned in the entire report, and the sole time software was even mentioned as a sustainability solution. So, these “technological tools” are simply scraping the surface of what technology truly can contribute to the sustainability efforts of the fashion industry.

Predicting the (Sustainable) Future

My main area of interest regarding fashion sustainability revolves around high-fashion companies making better consumer demand predictions. In essence, if you can look into the future and see exactly what people are going to buy, you can adjust your production cycle accordingly, see less unsold inventory (or waste), and allocate your resources wholly to products which will not end up landfill. In this case, predicting the future can be as simple as running a couple lines of code. 

Training a machine to recognize even the most specific trends in the industry can allow companies to run faster, more accurate analytics on what people are wanting and searching for. For example, a high-fashion house like Dior can then use this machine (or image recognition API) to detect which of their products are standing out most to social media users and publications (i.e. post-fashion week digests on Vogue.com). 

Here are two sample images from Dior’s recent Cruise 2020 collection, labeled by examples of what one could teach a computer to identify:

Neither of these images were used for AI training, solely as an example. Image Source.

The saddle bag, in fact, is a perfect example of the potential benefit of such specific image recognition tools. Dior’s famous saddle bag was the “it bag” right at the turn of the 21st century, which was brought back to life eighteen years later through a massive social media marketing push. How well this product launch was responded to, how frequently users were posting/talking about it, and the success with which it gained traction could all have been easily and quickly measured metrics through an image detection API. 

The day that saddle bag popularity begins to decline by, say, X%, Dior can reduce its production of the purses by a similar X% and result with less future unsold inventory. In short, Dior (and other high-fashion/luxury brands) can use computers to predict the future of trends and their products, thus revolve around quickly obtained, higher accuracy forecasts.

Fortunately, it seems that the fashion industry could be hopping on the AI-for-better-consumer-predictions train. Just a couple of months ago, H&M Chief Executive Karl-Johan Persson stated the brand’s investment in AI was “already helping predict trends and allocate garments to stores”. High-fashion & luxury conglomerates such as LVMH, who own century-old brands notoriously lacking engineering expertise, need to step up to the plate.

Sustainability reports from high-fashion conglomerates must start including more references to software innovations and big-data analytics. Nowadays, sustainability efforts extend far beyond replacing paper with online documents and heavy recycling, especially in an industry for which such solutions are a mere bandaid on a mountain of environmental detriment.

I’m currently a computer science student at Harvard University, focusing on the intersection between technology and fashion. I’m using image recognition models to analyze trends from high-fashion collections, which you can view here.

About The Author

Writer for the Harvard Technology Review.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

You don't have permission to register