The odds of survival is a risk faced by individual retail businesses. Economic and urban planning agencies have strong affinities to tackle both the causes behind failure of retail businesses and in predicting the likelihood of their success. The failure susceptibility of a retail business may be ascribed to various controllable factors, such as the quality or price of the store’s product, operating hours, and customer satisfaction. So, how can urban planners utilize location intelligence?
The vast amounts of urban mobility and social media activity provides unique insights that survey data, small scale research, field research and focus groups cannot offer. For instance, mobility data reveals the urban dynamics of different locations, such as whether a neighborhood attracts visitors from other neighborhoods, whereas social media data provides spatio-temporally referenced digital traces left by human use of urban environments. With easy access to the intangible aspects of urban life, such as people’s behaviors and preferences, researchers can investigate human activities and uncover key insights by joining and analyzing these disparate datasets.
As a pioneer in the geolocation data industry, Foursquare creates proprietary technology for understanding the places people go, offering a trustworthy, privacy-safe source of movement data. Researchers from New York University, University of Cambridge, and Singapore Management University leveraged the power of Foursquare’s location data – they conducted a longitudinal study using Foursquare’s venue check-ins observed in ten cities across the globe including New York and Singapore, that spans three years and over 75 million check-ins. Additionally, they received venue-related data using the public Foursquare API to understand the credibility of food & beverage establishments closure labels through time-stamped, public shared activities at related venues. These activities consisted of tips, public notes, short user reviews, or photo posts from visitors, allowing researchers to observe the activity at a venue beyond their prediction window and proxy the lifetime of the venues considered.
The Foursquare datasets and its variation of customer visits trends to such locations revealed important insights into how businesses are faring. Based on coefficients for the hourly temporal skew feature, they uncovered an interesting finding in which venues (both F&B and retail) that are popular around the clock, and not subjected to specific hours, have a better chance at survival. This finding suggests that restaurants that only cater to specific customer segments (eg. lunchtime office workers or dinnertime visitors) are more likely to experience failure. In Singapore for instance, the failure rate for restaurants in the Central Business District (dominant lunchtime presence) and Clarke Quay (geared towards tourists, leisure traffic; more active at night), is distinctively different. The researchers picked all restaurants from the two neighborhoods (139 venues in total), and ranked them by hourly popularity entropy (proportion of check-ins the venue received during that hour as compared to total check-ins received over all hours). When comparing the top 30 restaurants (highest entropy) and the bottom 30 restaurants (lowest entropy), they found a clear difference: only 73% of the bottom 30 restaurants survived the next 6 months, while 100% (all 30) of the venues in the top 30 survived. Additionally, the estimated coefficient of place entropy (the distribution of venue types in the vicinity of a venue) suggests that a decrease in entropy improves the likelihood of retail survival. In this case, venues that are in the middle of clustered neighborhoods tend to survive longer.
Over a six month period, analyzing information from Foursquare’s location data of venue check-ins proved to be indicative of a venue’s performance. As anticipated, the results showed a positive coefficient, indicating that venues that experience an upward trend in check-ins have a higher likelihood of survival. Meanwhile, when comparing the two cities of New York and Singapore, researchers found that in Singapore several of the features tied to a venue’s locality or neighborhood are statistically significant, such as the misalignment, which suggests venues that operate outside popular hours of the locality have an advantage over their neighbors. In instances like this, visitation patterns can be used to profile venues by prevalent activities and areas, forecasting movement and preference trends in order to predict how demand for venues changes over time.
This study illustrates how location data can be used to provide insights into the survival of retail businesses. Through extensive examination using the logistic regression model based on Foursquare’s data set, researchers were able to derive a better understanding of how certain factors of urban areas and human mobility significantly affect venues in different cities. Visit trends are a significant feature for all ten cities, and temporal alignment to competitors and distance of travel are also significant features in many cities, while temporal popularity skews. These results suggest that the dynamics of urban environments have a strong influence on venues. Additionally, with the consistent sign of the coefficient across all cities, they deduced that the role these features play in a business’ success is consistent.
This analysis demonstrates how large-scale multi-city generalizations pose challenges as different cities are unique in their own attributes. Although venue closure itself is predictable, the weight of features vary by city in their contribution to business failure. This is where location intelligence can be leveraged by urban planners to provide insights into both contextual and human dynamics, and to create better experiences for the population, retail business and urban development.
Foursquare believes in the power of location and its potential to positively influence the future. Our geospatial solutions present distinctive opportunities for knowledge creation in urban planning and beyond.
Contact us to learn more about leveraging our technology in scientific research, or check out the location data available for free on the AWS Marketplace.
