How machine learning & AI optimises retail supply chains
Case Study
Background
Retail supply chains have grown increasingly complex in recent years, thanks to the need for multiple sales channels and transport options to encourage growth.
Nowhere is this truer than in apparel, where seasons are short, lasting approximately six weeks. In this complex and dynamic sector, success relies on being able to dynamically and flexibly adapt to changes in demand.
Commonly, forecasters in this sector rely on methods, such as moving targets, to achieve this. However, these calculations struggle to handle complexities and non-linearities of relationships within the supply chain, making this a lengthy process at risk of error.
And a costly one, too. Each year, millions in revenue is lost to discounts and stockouts.
The Challenge
The Solution
From the outset, T-DAB recognised that the application of smart technology through machine learning and AI was key and that only a data driven technical solution could deliver the required magnitude of performance improvement to the supply chain.
Through an initial project discussion, the decision was made to focus the project on improving the supply chain for products once in the UK.
To achieve this, T-DAB proposed that a data driven system would need to identify where stock should be geographically held and then dynamically repositioned within the system to most closely match demand while minimising costs.
Intelligent by design, the solution would be capable of guiding a complex supply chain in how to work flexibly and intelligently, by dynamically re-distributing stock ahead of demand.
The strategy & design highlighted that during the “pre-season”, the machine learning system forecasts UK-wide demand of SKU-level lines, prior to a season of accurate purchasing of stock to meet demand.
Management would then be able to access forecasting of store level demand for SKU-level items, while, in the background, the system would optimise stock distribution to stores, regional storage and distribution centres to meet expected store demand.
During the pre-season, the system would then forecast the total amount of stock needed to satisfy demand, how to distribute it to ensure the greatest chance of having just the right amount at different locations at the lowest cost.
Alongside this, the supply chain management tool could determine how to geographically locate stock within the current infrastructure, as well as illustrating how this can be further improved by using optimal storage and distribution points not yet in existence.
The system would respond dynamically to the beginning of the “season”, by responding to changes in demand. It would adapt by redistributing stock to different locations as conditions and rates of sale change. Importantly, these locations need not be restricted to the physical. They can be substituted or supplemented by channels, such as online.
The Result
Thanks to T-DAB’s understanding of the next-generation of scalable, pervasive AI and machine learning applications, the high street retailer now has a detailed data & AI enabled supply chain strategy & roadmap to increase their yearly profit by 20%.
The retailer can target a reduction in management complexity and reduce risk, mitigating the effects of increasing multiplicity of channels, and consequently maximise profitability.
Furthermore, the responsive and flexible design would deliver an intuitive interface connected to an automated system. This would enable them to stay one step ahead in a highly competitive marketplace.
About T-DAB.AI
T-DAB.AI is data science and data engineering innovation company. We develop innovative, bespoke machine learning-driven solutions to allow anyone to infuse technology with the spark of predictive intelligence.