When AI meets store layout design
Store layout, floor area design, and product placement are important in-store marketing tactics that can influence customer decisions and in turn boost sales and profitability. Therefore, retail stores such as supermarkets and warehouses are prioritising their budget for sophisticated marketing strategies to ensure an efficient store layout. This would attract customers to the store, keep them browsing longer, expose them to more merchandise, and make them happy within the store, which consequently increases customer spending.
The current approach to designing store layout is passive and may not reflect customers’ actual behaviour nor the nuances between different residential populations and different stores. Moreover, conventional design process overlooks customers’ navigation through store aisles, time spent in each section, and customers’ emotions toward a product. Such data can be obtained and analysed by integrating artificial intelligence (AI) into the in-store CCTV infrastructure.
The retail sector has increasingly leveraged the power of AI and invested in a range of applications, including pay-with-your-face, check-out free grocery stores, visual and voice search, automated customer service agents, shopping advisers, and product recommendation platforms. Research has also emphasised the role of AI in capturing customer movement in stores and focused on consumer perceptions of AI, yet lacks a framework for how AI-derived insights can inform store layout design.
This research conducts a comprehensive review on existing approaches in store layout design and modern AI techniques that can be utilized in the layout design task. Based on this review, an AI-powered store layout design framework is proposed, applying advanced AI and data analysis techniques into existing CCTV video surveillance infrastructure to understand, predict, and suggest a better store layout.
Conventional approaches to store layout design
The main goals of layout design are to expose customers to more products, increase browsing time, make it easy for customers to find related products, and manage costs and inventory.
There are four major layout types.
1. The grid layout: a rectangular arrangement of parallel displays and long aisles. This is often a popular choice for supermarkets, grocery stores, and chain pharmacies. End caps in the grid layout where essential products are placed help expose consumers to a range of other products and potentially increase sales.
2. The freeform layout: a free-flowing and asymmetric arrangement of displays and aisles, employing different sizes, shapes, and styles of display, to increase customer’s time spent in store and their chance of engaging in exploratory behaviour.
3. Race track layout: offers a major aisle to control customer traffic to the store’s mul- tiple entrances, known as a loop layout. This leads the customer along a specific path to ensure that they are exposed to as many store sections as possible.
4. Circulation spine layout: a traffic loop around the entire store, but the layout also includes a customer path right through the middle of the store
Conventional store layout design approaches rely on three criteria:
1. Product categories: displaying the same categories of products on the same shelf or in the same aisle. For example, supermarkets frequently group products in a bakery area, vegetable area, and fruit area.
2. Cross-elasticity: placing product categories side by side following cognitively logical pairs, considered as cross-elasticities. This captures cross-product interactions in demand via prices as the sales of one product responds to price changes in another product.
3. Consumption universes: clustering products around consumer buying habits. For instance, breakfast products including tea, bread, cheese, and cereal are presented in the same place.
Current AI applications in retailing
Modern visual AI techniques can be used to obtain a variety of insights into customer shopping behaviour. The presence and location of customers, shopping items, and shopping carts can be detected. Pose estimation can detect the action and interaction of a customer with surrounding objects. Heatmaps provide a visual summary of information by using colours to represent data, for example, using warmer colours for locations where all customers spend more time there.
Customer identification tasks involve facial recognition using biometrics data and characterisation of customers through important demographic attributes, including a person’s age, gender, and ethnicity. Such information could be valuable for customer-tailored targeting.
Tracking customers over a whole CCTV camera network is also of great interest to business analytics. Human tracking within a camera focuses on locating human objects in each frame of a given video sequence from a camera, while human tracking across multiple cameras establishes associations between detected humans from non-overlapping cameras.
Customer emotion recognition can detect observable facial expressions. Multimodal emotion recognition infers a combination of facial expression, speech, and body gesture that reflect human emotions. These tools can help marketers and managers understand customer reactions to the products they sell.
Customer actions can also be recognised based on the body gestures of the customers in relation to the shelves, the products, and the trolley/basket they are using. This reveals their interest in the product and elements in the decision-making process that drive the purchase.
Employing AI techniques to improve layout design
From the analyses and review, an AI-powered Store Layout Design framework was proposed to apply visual AI and data analytic techniques to designing store layout to improve customer satisfaction. The design of the framework consists of four phases with a multiple-layer architecture:
1. “Sense”: Raw data as video footage from CCTV is collected in the data layer. The pre-processing layer applies several techniques to improve the quality of captured images and videos such as cleaning, de-noising, de-blurring, and transformation correcting.
2. “Think”: The intelligent video analytic layer uses advanced AI techniques and data analytics to interpret and extract meaningful information from images and videos.
3. “Act”: The execute layer uses the obtained knowledge and insights from the analytic layer to improve and optimise store layout.
4. “Learn”: The process operates as an iterative learning cycle to enable machine learning and continuously improve store layout, which should enhance customer satisfaction.
Following this model can reduce the gap between customer expectations and store design. It also demonstrates the power of AI to boost business productivity. Marketers and managers can also utilise the insights to segment customers based on different behaviour patterns, such as time in store, size and frequency of purchase, price, and sensitivity to store display promotion.
Researcher
More information
The research article is also available on eprints.