Project partner: Co-operative England
According to the Met Office, even a four-degree rise in temperatures can lead to different buying habits. When the weather changes, retailers risk losing millions through low stock levels, wasted goods and disappointed customers.
National supermarket chain the Central England Co-operative approached Birmingham City University to help them manage risk and maximise profits by building a system that suggests stock requirements based on the weather.
The challenge – knowing what the customer needs before they need it
Birmingham City University’s Salameh Abu Rmeileh, Assistant Lecturer in the School of Computing, and BSc Computing student Norman Farrow-Butler are developing machine learning intelligence that will directly inform Co-op’s stock-taking systems. The system will compare live and forecasted weather data with shopping basket data from Co-op’s customer purchasing databases in order to predict which items need to be stocked and made accessible depending on current and future weather conditions.
Birmingham City University’s Data Analytics and Artificial Intelligence research centre is conducting pioneering work in the machine learning space and are keen to share their knowledge and help businesses understand more about this technology. Machine Learning is a type of AI that enables software applications to become intelligent and accurate in predicting outcomes without human intervention. In practice, this means the AI built by Birmingham City University can identify correlational trends between the items purchased at any store and the weather in that area, allowing Co-op to maximise profits by only ordering what is needed and improving the customer experience by maintaining the availability of high-demand items. As it utilises machine learning, the system will continue to adapt to changing trends, providing Co-op with a live view of customer needs in relation to varying weather patterns.
The delivery
Salameh and Norman were able to use their skills and knowledge base to work directly with the Information Technology team at Co-operative England. After initial discussions about their needs and systems, the BCU team were able to formulate a plan for a proof of concept and identify the data they needed to feed into the machine learning AI. Focusing on a single store in Birmingham, Salameh and Norman were able to use weather data from that area and combine that with Co-op’s databases to enable the machine learning AI to make initial predictions.
Looking to the future
With this technology, Co-op will be in the best position to maximise profits in an increasingly challenging retail environment. With the rise of online shopping, the role of the physical shopping outlet is changing and a quality experience is the best way to retain customers. Using machine learning AI, Co-op can ensure the availability of high-demand items and predict the best store layout to encourage purchasing. The increased accuracy of inventory will also help Co-op meet their environmental targets by reducing food waste and losses from spoilt stock.
Aside from the immediate benefits, machine learning tools are a key research asset to a business. Starting with this project, Co-op and Birmingham City University could go on to adapt machine learning systems to increase product and market research. Having a good understanding of the shifting needs of their customers could give Co-op an advantageous market position and lead to an increased drive in data-driven solutions.
The collaborative partnership
Birmingham City University has a history of successful collaborations with The Co-operative group and in August 2018 announced a partnership to collaborate on a new programme with the aim of developing skills and improving regional productivity through research. Salameh’s work is one of many projects that utilise new and emerging digital technologies to help Central England Co-operative grow as a business.
