Bestseller and IBM Garage bring sustainable fashion forward with

Integrates AI to predict best merchandise and drive supply chain efficiencies

By | 3 minute read | September 11, 2020

When it comes to driving change in business priorities and practices, data and AI can be an industry leader’s best ally – and that in turn, can be better for the environment.

Greater transparency and deeper insights introduce much-needed changes to the way organizations produce, sell, transport and consume, with the aim of dramatically reducing harm to the planet.

The fashion industry is well-suited to transform its notoriously wasteful manufacturing practices with AI. The second-most-polluting sector after oil and gas, fashion consumes thousands of liters of water to produce a single item of clothing that takes more than 200 years to biodegrade.

What’s worse, many fast-fashion seasons occur monthly, leaving 80% of a season’s garments out of fashion in weeks. A whopping 24% of all products remain unsold – even after deep discounts on those unsold at full price. The industry churns out USD 300 billion in unsold inventory that ends up in giant landfills. It’s a problem amplified by our socially distanced “one-click” shopping habits.

Lacking the ability to try before we buy, 40% of us are likely to return impulsive purchases to the retailer. This is costly to the merchant and creates significant additional transportation and CO2 emissions. But using AI to help customers purchase items that fit their physical dimensions and their preferred style, merchants can slash their return rate dramatically.

Waste not want not. Are fall wellies in your cart?

Overseeing 20 fashion brands for women, men, teens and children such as Jack & Jones and Vero Moda, Bestseller wanted to recover a 2019 dip in popularity of its ONLY brand while at the same time, face the elephant in the room – sustainability.

At the corporate level, Bestseller launched its “Fashion FWD” initiative to drive sustainable fashion. To support this worldwide effort, Bestseller India saw a high-value opportunity to develop and digitize smarter design and planning tools to minimize waste early in the creative process. By designing and producing apparel that more accurately matches market demand, the clothing company could reduce the high economic and environmental costs of unsold inventory, answering such questions as:

  • Why aren’t these boots selling better?
  • What impact would changing the price have?
  • How do these boots sell across various stores?
  • How have similar products from prior collections performed?

Together with IBM® Garage™, Bestseller launched, the fashion industry’s first AI project aimed to increase sell-through rate and reduce unsold inventory. Fabric AI is an IBM Cognitive Enterprise innovation project delivered by IBM Services®. A managed platform as a service (PaaS) offering that met Bestseller India’s requirements, the solution is deployed to the IBM Cloud® Kubernetes Service.

IBM Garage ran an intensive IBM Garage Enterprise Design Thinking Workshop to help Bestseller India experts create a roadmap for co-creation, beginning with user research. From there, the project focused on creating intelligent workflows for key business processes. IBM Watson® AI tools predict the best products to incorporate into new offerings, determine the right product mix for each store and improve the efficiency of the supply chain. unveils the mystery of fashion trends

Integrating IBM Watson APIs – Natural Language Search, Visual Browse and Visual Search – into existing workflows, is an explainable sales analysis and forecasting asset that examines sales to help predict future sales for new products at different stages of product design and development.

Explainable sales forecasting models not only improve the trustworthiness of model outputs but also provide transparency for all stakeholders involved in product development and launches –improving accountability and fostering a collaborative environment among stakeholders with competing needs. Secondly, designers, buyers, planners and merchandisers benefit from explainable AI-based interventions during preseason, in-season and end-of-season decision making. For example, for markdowns, provides designers with information about the product’s probable success down to the attribute level. This means the client receives advice about products to mark down, when and by how much.

Other features:

  • Compares new products with similar products from the past season to reveal how these products may performed in the future.
  • Forecasts sell-through rate for new products based on product attributes and/or product images.
  • Uses explainable sales analysis for past seasons to reveal why the product did – or did not – sell well.
  • Provides what-if-analysis of product attributes to enable designers and buyers to make informed decisions for product attributes.
  • Delivers customized data to help designers, merchandisers and buyers choose optimal assortments with their domains.

In this presentation, hear from Bestseller how IBM data and AI enhances Bestseller’s brand and brings sustainable fashion forward.

Power digital transformation with new ways of working, leading technologies and multidisciplinary experts. See how IBM Garage works.