In today’s world, it pays to be green. A recent study conducted by Nielsen revealed that 73% of consumers are willing to modify their consumption habits to limit the impact they have on the environment; and in this context, 46% of that same population say they would forgo a brand and switch to an eco-friendlier alternative. Whilst a large part of the movement towards sustainability lies in renewable and ‘clean’ resources, improvements in the usage of data- and AI-empowered business processes can also have an impact in reducing our environmental footprint.
In the past few years, the UN Conference on Trade and Development listed the fashion industry as the second-largest source of pollution in the world; with more than 70% of clothing ending up in the landfill, the environmental impact is estimated to be just behind that of the oil production industry. In response to this kind of observations, fashion retailers are increasingly acknowledging their share of the responsibility and actively promoting eco-friendly initiatives, often launched by freshly hired ‘Chief Sustainability Officers’.
These conversations around sustainability led to industry leaders signing the G7 Fashion Pact in 2019 to formalise their promises and translate talk into action:
H&M has committed to using 100% sustainable materials by 2030 after launching its Conscious collection of eco-friendly fashion pieces in 2012
Zara intends to fulfill the same promise by 2025
With a new generation of customers that are more purpose-driven and increasingly favourable towards brands that champion ethical and eco-friendly values, companies that are quick to clean up their act will be rewarded on the short and longer term.
With a heightened focus on sustainability and the risks/rewards highlighted, consumer brands are increasingly looking for ways to rethink their business models and product portfolios to hit commercial and sustainability targets. The solutions vary by industry and commonly discussed solutions are certainly to adopt eco-friendlier materials and production methods, limit the use of water or toxic chemicals, rethink packaging, and engage in recycling activities. But an area that gets less public attention is around the end-to-end supply chain efficiency level with the aim to limit wastage, unnecessary transportation, and superfluous production. This is where data science can help.
In this article, we explore a holistic approach that brands can use to rethink their end to end supply chain and ultimately limit their environmental footprint from production planning all the way to the customer’s door.
Getting a sense of customers’ needs and wants through data is crucial; it can help businesses steer their planning proactively, manage inventories effectively, and meet demand more accurately. Brands that understand demand well can limit production and distribution inefficiencies which often bear tremendous financial and environmental costs.
Zara, well-known for its disruptive distribution network, commits only 50% of inventory when entering a new season, leaving the remaining half agile to respond to customer trends of the time. Furthermore, the smart chip technologies integrated in the brand’s clothing act as real-time data hubs, allowing the retailer to regulate its in-store replenishment process but more importantly, support its ‘just-in-time’ production. Although this approach is widely praised, Zara is relatively alone in this space. Whilst the latter waits for sales data to inform production, Rue21 takes a more proactive approach by engaging the customer in the decision process through surveys & questionnaires before any production begins. This allows the teen fashion brand to have a better sense of which products customers want in order to allocate production resources more efficiently and bring to market promising products only. The common thread between the two brands the willingness to understand demand and tying it to the production & stocking efficiency that benefits the bottom line and reduces environmental impact.
But despite major leaps in the field of data science, the ways of forecasting product demand in many large corporations have not changed a great deal over the past 20 years and are based on either expert-driven or tool-driven fundamentals. While the expert-driven approach struggles with consistency and operational efficiency, the tool-driven approach suffers from a lack of adaptability, interpretability, and a scope limited to existing products. Fully leveraging the power of AI to take product-level demand sensing to another level is a tough challenge, but the business & environmental value potential is immense.
We believe the answer lies in a flexible approach that encompasses both trend detection for product development, as well as demand forecasting for production that factors in external data sources for early signals, and allows for human input. The success of this process has been proven time and time again with our clients.
For more details on how to successfully build a demand sensing capability, check out these article.
With the explosion of online purchases, it is fair to ask ourselves what type of sales channel is the most environmentally efficient. While customers minimise their number of trips to the store by purchasing online, the e-commerce industry appears to be a carbon-intensive alternative to traditional brick and mortar retail, especially for large items.
First because online deliveries require additional packaging
Second because customers tend to make smaller but more frequent purchases which increase the number of micro-deliveries
Third because items often come from different distribution centres which augments the number of dedicated shipments
And fourth, because most of the time customers are likely to return their purchase, due to the little effort it represents, which leads to extra-large orders and thus bulkier deliveries on the way out and additional delivery on the way back
While technologically advanced gadgets (using AR / VR) can tackle that last point by offering the opportunity to customers to ‘try on’ clothing and be more conscious of their purchase, fashion retailers can look at some ‘quicker’ wins.
Leveraging the power of data & AI can help brands nudge their customers to opt for the ‘right option’ and place the environmental responsibility in their hands. Taking inspiration from Farfetch’s recent tool that allows customers to track the environmental impact of their buying decisions between new versus used item bought, or Reformation’s Refscale, brands can educate their clientele by measuring the environmental footprint of various means of receiving an order (based on the size of the parcel, the delivery location, its proximity to stores/collection points, etc.). Such best practices can go a long way in helping brands reach their cost-saving and environmental impact goals.
Decades of price slashing and discounting has created a culture of disposability in the fast fashion industry. But the race is now back on to put the value at the centre of the purchasing decision. As customers emerge out of the COVID-19 crisis, they are likely to be more conscious of consumerism and their discretionary spending. It is, therefore, a pivotal moment for brands to build back value into their image and products.
Even before the COVID-19 crisis, many brands were still falling into the trap of all-year-round discounts cutting into margins without compensation in volumes. Higher discounts, on the other hand, yield a bigger volume response but are not sustainable. Outside specific promotional events, such as Black Friday or clearance, where large incremental sales are gained and brands cannot afford not to participate, the volume increase from small to medium discounts rarely compensate for the revenue losses. One thing is certain – there is almost always a trade-off to be made between revenue, volume & the bottom line when it comes to promotions.
In order to optimize revenue & profits, fashion brands need to understand the customers’ willingness to pay (elasticities) at a product level, allowing them to differentiate discounts to yield the optimal response against their objective. Working with a fashion retailer, we also observed the halo a key promotional period can have on full-priced items by driving footfall, stressing the importance of setting up the right promotional strategy in terms of product that is designed to maximize revenues without the need to slash prices across the board.
Of course, the fashion industry has always had a need to clear stock before the next season, however, with the advances in demand prediction and a fresh approach to the frequency of seasonal product drops as a result of the shake-up during COVID-19, the sheer volume of the challenge should reduce.
Assuming overall demand has been accurately assessed, the next step is to create the most suitable product assortment per point of sale to avoid replenishment issues and unnecessary product travels; translating into economies of scale on distribution costs and a reduction of harmful emissions. But whilst many companies define assortment according to the point of sale’s qualitative demographic and socio-economics surroundings, or simply by store size & type, only a few companies leverage the power of AI to optimise the product offering through the consumer lens. Our clients often look to us to help them adapt assortment according to purchasing patterns observed at a store/account and geographical level. This will become that much more important in the wake of the pandemic as the lack of tourism will result in fundamental shifts in the make-up of the customer in each location. Moreover, for retailers that are able to leverage their stores for e-commerce delivery, it becomes crucial to get assortment & stock supply right to meet the omni-channel demand locally.
Beyond profiling consumers on purchasing behaviours, understanding product affinities can also help further enhance product portfolio decisions. By applying machine learning to the data behind customer receipts, companies can better identify complementary products which could be promoted & displayed together, as well as substitute products that cannibalise each other and should not form part of the same store assortment. Executing against these findings will increase average order value and enhance sell-through at a store level. Moreover, all this data can be fed back through the value chain to fuel demand sensing algorithms, fostering efficiencies across the different production & distribution phases and proving the impact data science can have on boosting sustainability.
Gradually, a sustainable mindset has become a pre-requisite rather than a differentiator in the fashion industry. Whilst companies work towards a longer-term material revolution, there are quick wins to be had by leveraging the power of data. Market players need to consider a wider scope of business operations in establishing their sustainability agenda – from using market signals for demand sensing, all the way to empowering the consumer to make more conscious choices. The correct exploitation of the data stemming from this extended supply chain can lead to an enhanced and more competitive business model by tying economic and environmental successes together.
Ready to talk or simply curious to learn more? Reach out to our experts to understand how your operational efficiency challenges can be tackled through data science.Review our Business cases in the Retail industry