Geo-Tracking Consumers and its Privacy Trade-offs

48 Pages Posted: 5 Oct 2023

See all articles by Unnati Narang

Unnati Narang

University of Illinois at Urbana-Champaign

Fernando Luco

Texas A&M University

Date Written: August 3, 2023


In recent years, firms have become capable of constantly geo-tracking consumers’ locations and movement patterns through mobile apps. Geo-tracking data can allow firms to better predict consumers’ future actions. However, geo-tracking also raises privacy concerns among consumers and regulators. Using rich geo-tracking data with over 120 million driving instances for 38,980 app users, we quantify the extent to which these data allow restaurants to better predict their consumers’ visits one week ahead relative to using consumer demographics and past behaviors. We also examine how restricting geo-tracking data under counterfactual privacy policies impacts the performance of prediction models that use these data. Using a machine learning framework, we show that geo-tracking data increase the prediction accuracy of our models by 3.1% relative to models that use detailed demographic and behavioral information on past visits and by 14.8% relative to models that use only baseline demographics that do not include any location information. The results from our counterfactual policy analyses further show that privacy restrictions that limit what geo-tracking data are tracked, which users are tracked, and where and how frequently they are tracked reduce the predictive performance of geo-tracking. However, the decrease in performance varies by the type of policy restriction; policies that restrict what data are geo-tracked (i.e., user-level summary of driving behaviors rather than geo-coordinates) and where users are geo-tracked (i.e., within a few miles of a business location), result in the largest decreases in predictive performance relative to complete geo-tracking. Overall, models with restricted geo-tracking still generally outperform models that do not use any geo-tracking information. Our research can assist managers and policymakers interested to assess the risks and benefits associated with the use of geo-tracking data for making predictions.

Keywords: Mobile app, location tracking, privacy, prediction, machine learning, transformers

JEL Classification: C23, C45, M1, M3, Z23, Z28

Suggested Citation

Narang, Unnati and Luco, Fernando, Geo-Tracking Consumers and its Privacy Trade-offs (August 3, 2023). Available at SSRN: or

Unnati Narang (Contact Author)

University of Illinois at Urbana-Champaign ( email )

601 E John St
Champaign, IL Champaign 61820
United States
N/A (Fax)


Fernando Luco

Texas A&M University ( email )

4228 TAMU
College Station, TX 77843
United States

Do you have negative results from your research you’d like to share?

Paper statistics

Abstract Views
PlumX Metrics