Compared to online shopping experiences, physical retail stores present a static, single experience at all times of the day for all user groups. Store layouts have minor changes every few weeks and occasional more major seasonal changes, but little day-to-day and no user-by-user customization. Online shopping experiences, on the other hand, are extremely customized -- every user’s Amazon page, for example, is highly tailored to his or her preferences and web navigation patterns. Online retailers can accomplish this with data-driven techniques for recording how users mouse over, click through, and interact with webpages.

​

Interestingly, some pilot retail stores (such as “Zippin” and “Amazon Go”) hint at near-future possibilities for physical retails to generate and respond to use-pattern data, just like online retailers. Given that these sensor-laden retail stores will soon have access to a similar level of user data as online shopping, how could this data be leveraged to create more customized, adaptive experiences over the course of a day, better responding to different user groups’ shopping needs and preferences?

​

This project draws from online retail data collection concepts such as “dwell time” and “conversion rate” to imagine new types of data creation for physical retail stores. Without changing the general design of the store, sensors are discretely gridded into the ceiling and used to create heatmaps of dwell time, purchase conversion, and other forms of customer use-pattern data.

​

These mappings could influence real-time how the store is laid out. Using robotic furniture technology (similar to the robotic warehouse shelves already used in Amazon warehouses), stores could reconfigure to target different user groups over the course of a day. This short project tests how a store might reorganize itself in response to use-pattern data, creating a more customized, personalized, and streamlined experience for shoppers visiting and different times of day.