Deep learning for customer insights discovery from videos

Measuring customer experiences has historically been based on analyzing traditional structured data, which mainly consists of surveying questions/answers as well as evaluating purchases and returns. However, using this type of inherently constrained data
alone misses the bigger consumer picture. Deeper insights can be derived from unstructured data such as videos of the customers at the store, where tremendous untapped insights exist. In fact, retailers are not just interested in what do customers buy, they also want to know how do customers shop. The ability to create a bridge between digital and physical behavior to understand the entire journey and connect store transactions to digital marketing investments, is currently a huge black hole in business intelligence and customer analytics.
The way companies bridge this typically is by using very expensive research techniques like focus groups or shopper diaries. TO BE CONT’D

Faculty Supervisor:

Marco Pedersoli

Student:

Rafik Gouiaa

Partner:

Faimdata

Discipline:

Visual arts

Sector:

Information and communications technologies

University:

Program: