Using sensor technology to evaluate the effects of a price promotion on store arrival traffic
利用传感器采集的顾客到店时间数据，通过伽马分布建模预测促销期间的反事实客流基线，从而分离出促销活动带来的真实客流增量，帮助零售商诊断客流与转化问题并优化运营。
• Sensors located at store entrances collect shopper-specific arrival times. • Gamma distributions jointly model the mean and dispersion of inter-arrival times. • Inter-arrival times outside the promotion period provide a counterfactual baseline. • Counterfactual inter-arrival times assume that the promotion has not occurred. • The effect of the promotion compares observed and baseline inter-arrival times. Retailers often evaluate the effect of marketing actions (MAs) on store performance via sales volume, a practice that obscures the role of store arrival traffic as a primary driver of profitability. This paper develops a modeling approach to measure MA-driven traffic surges using sensor-captured, shopper-specific arrival times. Our approach isolates the specific impact of an MA from natural traffic variability by predicting a counterfactual baseline for the arrival traffic during the MA period. Our method provides four key managerial contributions: (i) quantifying true “lift” of the MA beyond natural traffic variability, (ii) diagnosing traffic versus conversion failures, (iii) aligning staffing and inventory with predicted traffic flow, and (iv) using early patterns in traffic as real-time warnings to guide later operational adjustments. We illustrate our method with shopper-specific arrival data from a two-week price promotion that generated 2,880 additional arrivals.