Though not her main focus, sustainability was something that came up repeatedly in conversations with senior financial and marketing executives at major brands. McClay realised Google was in a position to help.
“What came up time and time again was data,” McClay said over coffee in one of Google’s London offices. “I thought to myself, ‘wow, we have this team, the Google Cloud team,’ and essentially what they do is they stitch together data.”
Those conversations led to a months-long project that has been quietly gaining momentum within Google. To figure out where to focus its data-processing capabilities, the tech giant brought on the fashion innovation consultancy Current Global and partnered with eco-luxury brand Stella McCartney. They settled on a widely known, but difficult to quantify problem: the environmental impact of producing cotton and viscose--two of fashion’s most commonly-used materials.
It’s a call to action … trying to get the fashion brands and people with data that’s very close to the source to work with us.
Over the coming months, the companies plan to pull together information from diverse sources - including years of data Stella McCartney has gathered on its own supply chain - and then run it through Google’s machine-learning algorithms, which can interpret complex data sets practically in real time. The companies aim to complete the first iteration of their supply chain tool by early next year, with a goal of creating versions that can be used by other brands, and on other links in the supply chain.
If successful, the tools they develop will give brands a clearer view into their supply chain, one of many efforts underway to demystify fashion’s “black box.” Fashion ranks among the world’s most polluting industries, but the data available to measure and understand its impact is patchy at best. Fashion’s supply chain is notoriously complex and murky, and many brands struggle to trace where their clothes are manufactured, let alone the source of raw materials or their environmental impact.
The group will announce the cotton and viscose experiment Wednesday at the Copenhagen Fashion Summit, the industry’s annual sustainability gathering. Their first step: asking brands, manufacturers, NGOs, academics and others to supply relevant data points.
“It’s a call to action … trying to get the fashion brands and people with data that’s very close to the source to work with us,” said Ian Pattison, Google Cloud’s UK head of engineering for retail customers. “We really want to … bring that data together, and put it in one place and make it visible to everybody.”
The goal is to be able to provide granular, local insights on a range of different metrics--including emissions, water use and pollution, and soil impact--that brands can use to help guide design and sourcing decisions. Long-term, the ambition is to make the tool freely accessible to brands, and give other materials beyond cotton and viscose the same treatment.
This topic hadn’t come up in our collective consciousness until a couple of years ago really, so no one was thinking to be able to track emissions down the value chain.
That’s something that isn’t even close to possible at an industry level at the moment. For many brands, digging into raw materials represents a final frontier on their sustainability journey. It’s the area furthest removed from the finished product and often where they have the least visibility or understanding of where things are coming from.
The data that does exist is often out-dated, based on best-in-class examples, or without geographic differentiation. Even as scores of brands layout targets to improve their environmental footprint, they’re often relying on bad or incomplete information to figure out where and how to focus their efforts. It’s an issue that’s plagued even the sector’s most high-profile initiatives to tackle the problem. When dozens of brands pledged to reduce the industry’s emissions by 30 percent by 2030 last year as part of the UN’s Fashion Industry Charter for Climate Action, they didn’t know the baseline they had committed to reduce from.
“This topic hadn’t come up in our collective consciousness until a couple of years ago really, so no one was thinking to be able to track emissions down the value chain,” said Elisa Niemtzow, managing director at non-profit consultancy BSR. “There’s a tremendous amount of work needed in terms of maturity, in terms of how we do this technically.”
While a growing number of companies are delving deeper into their supply chains, the information they collect is often held in proprietary databases. Efforts to compile industry-level tools are resource-intensive and have often proved slow-moving. Beyond the fashion industry, governments, non-profits and academic institutions gather data on water use, pollution, deforestation and more, but the information isn’t consistent globally and hasn’t been broadly applied to fashion’s supply chains.
Part of Google and Stella McCartney’s project is identifying which sets of data don’t yet exist, and how to create them.
Analysing the data is easy; getting the data is difficult
“Analysing the data is easy; getting the data is difficult,” Google’s Pattison said. “But you’ve got to try.”
The decision to focus on cotton and viscose is the result of dozens of conversations with brands, as well as multiple design sprints with teams of engineers at Google to figure out what might be feasible.
“We literally looked through the entire supply chain,” said Current Global’s Co-Founder Rachel Arthur. Ultimately, they decided that focusing on raw materials made sense because it’s the part of the supply chain “where the most impact was happening, but also where there were the biggest gaps for data,” she added.
Cotton was an obvious place to start because it’s such a big market, accounting for around 25 percent of global fibre demand, according to Edinburgh-based consultancy Wood Mackenzie. It’s also a thirsty crop that sucks up vast amounts of water and is often treated with large doses of pesticide. Viscose, which is made from wood pulp, made sense as a test case because it has a relatively consolidated supply chain that is comparatively easy to map.
Don’t get me wrong, it’s a massive goal, but if there’s any business that can have that level of ambition, it is Google.
The companies are already drawing on data they have between them and in publicly-available databases. For instance, they’re looking at Google’s satellite mapping technology to understand land usage, drawing on success the company’s already had applying machine learning to track patterns of deforestation.
Following on from their Copenhagen launch, they’re hoping to tap into information sitting with other companies and institutions willing to help--giving Google’s machine-learning algorithms the best data possible to understand the environmental impact of viscose and cotton production around the world.
“This is one of the biggest technology companies in the world putting a stake in the ground, saying we are going to do something to benefit the fashion industry towards sustainability,” Current Global’s Arthur said. “Don’t get me wrong, it’s a massive goal, but if there’s any business that can have that level of ambition, it is Google.”