Using gravity model to make store closing decisions: A data driven approach

Area of Science:

• Retail analytics
• Spatial modeling
• Operations research

Background:

• The COVID-19 pandemic has led to financial difficulties for many businesses, necessitating store closures.
• Decisions on store closures are often based on individual store performance, potentially overlooking broader chain impacts.
• Existing research lacks comprehensive models for strategic store closure decisions.

Purpose of the Study:

• To develop and validate a novel model for predicting the impact of store closures on chain revenue.
• To provide a data-driven approach for optimizing store closure decisions in retail chains.
• To analyze the interconnectedness of stores within a retail chain.

Main Methods:

• A modified Huff gravity model with a new attractiveness factor was proposed.
• A forward-backward approach was utilized for model training and prediction.
• Large-scale spatial, mobility, and spending datasets from New York City department stores were analyzed.

Main Results:

• The proposed model's store closure recommendations may differ from those based solely on single-store performance.
• Store closures can lead to significant revenue loss due to complex inter-store customer flow dynamics.
• The performance of a retail chain is an emergent property of store interactions, not merely an aggregation of individual performances.

Conclusions:

• The developed model offers valuable insights for strategic store closure decisions.
• Managers can utilize this approach to mitigate revenue loss associated with downsizing.
• Understanding store interdependencies is crucial for effective retail chain management.