Authors: Joel McCune*, Esri
Topics: Geographic Information Science and Systems, Business Geography
Keywords: business, machine learning
Session Type: Virtual Paper
Start / End Time: 8:00 AM / 9:15 AM
Room: Virtual 49
Presentation File: No File Uploaded
Quantitative consumer forecasting is a methodology for structuring quantitative measures describing the demographic and geographic factors between customers and the surrounding retail landscape. This facilitates application of machine learning methods for forecasting customer engagement. The behavior of the customer is being forecasted and summarized to each brand store location to provide decision making insight.
This methodology models and enables forecasting the engagement between customers and the surrounding retail landscape by taking into account demographics, the geographic relationships between customers and the surrounding brand locations, and customers and competitor locations.
Data is then added quantifying the engagement of customers with a brand store location. Sources of this data can include data from loyalty programs, online sales (for halo forecasting), and human movement data - cell phone location tracking. This metric of engagement is then used as the label for training a machine learning model. This model can subsequently be used to evaluate hypothetical scenarios, adding or removing locations. This can even include systematically evaluating different combinations of adding and removing locations across an entire market to provide valuable insight for determining if there is headroom in a market, and creating a market strategy.