Publication Date

2018

Document Type

Dissertation

Committee Members

Pratik Parikh (Advisor), Xinhui Zhang (Committee Member), Frank Ciarallo (Committee Member), Nan Kong (Committee Member), James Munch (Committee Member)

Degree Name

Doctor of Philosophy (PhD)

Abstract

The retail industry in the U.S. contributed 1.14 trillion in value added (or 5.9%) to the GDP in 2017, an increase of 3.7% from the previous year. While store closures have dominated the news in the recent past (e.g., Toys-R-Us, Sears, and Bon-Ton) due to ineffective supply chain practices, inadequate in-store experiences, and competition from e-tailers, other retailers such as Ross, T. J. Maxx, Burlington Coat Factory, and Kroger have been expanding their footprint. Brick-and-mortar stores are unique as they allow shoppers the ability to see, touch, and try products, in addition to exploring new products. Kohl’s CEO has even indicated that 90% of their revenue is still generated in brick-and-mortar stores. Besides reducing supply chain costs, retailers have been paying considerable attention to redesigning their stores by varying layouts and displays to improve shopping experience and remain profitable. However, a lack of scientific methods that correlate layout changes to improved experience has often led to time-consuming and expensive trial-and-error approaches for the retailers. This research focuses on the design of such brick-and-mortar stores by developing a quantitative approach that models the visual interaction between a 3D shopper’s field of view and the rack layout. This visual interaction has been shown to influence shopper purchasing habits and their overall experience. While some metrics for visual experience have been proposed in the literature, they have been limited in many ways. The objective of this research is to develop new models to quantify visual experience and employ them in layout design models. Our first contribution consists of quantifying exposure (which rack locations are seen) and the intensity of exposure (how long they are seen) by accounting for the dynamic interaction between the human 3D field of regard with a 3D rack layout. We consider several rack designs/layouts that we noticed at nearby retail stores, ranging from the typical rectangular racks placed orthogonal to the main aisle to racks with varying orientations, curvatures, and heights. We model this 3D layout problem as a series of 2D problems while accounting for obstructions faced by shoppers during their travel path (both uni- and bi-directional). We also validate our approach through a human subjects study in a Virtual Environment. Our findings suggest that curving racks in a layout with racks oriented at 90° could increase exposure by 3-121% over straight racks. Further, several layout designs could increase exposure by over 500% with only a 20% increase in floor space. In our second contribution, we introduce the Rack Orientation and Curvature Problem (ROCP) for a retail store, which determines the best rack orientation and curvature that maximizes marginal impulse profit (after discounting for floor space cost). We derive impulse profit considering the probability a shopper will see a product category, the probability the shopper will purchase a product from that category if seen, and the product category’s unit profit. We estimate the probability that a shopper will see a location through a novel approach that considers (i) the effective area of that location, (ii) probability distribution of a shopper’s head position based on real shopper head movements, and (iii) exposure estimates from our approach in Contribution 1. To solve the ROCP, we design a particle swarm optimization approach and conduct a comprehensive experimental study using realistic data. Our findings suggest that layouts with either high-acute and straight-to-medium-curved racks or high-obtuse and high-curved racks tend to maximize marginal impulse profit. Profit increases ranging from 70-233% over common rack layouts (orthogonal and straight racks) can be realized depending on the location policy of product categories. The sensitivity of these solutions to shopper volume, cost of floor space, travel direction, and maximum aspect ratio is also evaluated. The implications of our proposed models and findings are wide-ranging to retailers. First, they provide retailers with insights on how design parameters affect both exposure and marginal impulse profit; this can help avoid expensive experiments with layout changes. Second, they reveal hot-warm-cold spots for specific layout designs, allowing for effective product location assignments. Finally, these insights can help enhance shopper interactions with products (i.e., ability to see more products, find products faster), which can improve their shopping experience and drive up sales.

Page Count

160

Department or Program

Ph.D. in Engineering

Year Degree Awarded

2018


Included in

Engineering Commons

Share

COinS