Retail has been a busy sector for a few years now. The growing popularity of e-commerce is destabilizing the well-established physical stores, which feel a little helpless and less well equipped to fight back.

E-commerce tracking marketing techniques knows everything about you or almost everything thanks to your online history and the tracking of your clicks. They know your desires based on your research and visualized articles, while in-store, the path that precedes the purchase remains obscure.

In-store analytics solutions

Various solutions are offered for making better business decisions based on tangible data. In reality, these solutions are for the most part solutions only.

retail analytics

Why ?

Take the most commonly used solution: Wi-Fi analysis. The advantage of this solution is that the technical device is relatively financially accessible. It will actually detect smartphone owners whose Wi-Fi is turned on, and therefore the accuracy of counting people is relatively poor. The radius of action is often such that the difference cannot be made between people actually inside the store and those nearby. The connection breaks sometimes making the visualization of the visitor’s itinerary also uncertain. Although inexpensive compared to other in-store data acquisition solutions, the implementation of this type of device is debatable because of its unreliability. The accuracy and nature of the data collected is limited, it is unthinkable to make real business decisions on this basis. Retailers must have customer knowledge that is both fair and comprehensive. Similarly, they must be actionable and generate ROI for sales, merchandising and operations teams.

Live video analytics

A much more reliable and complete solution is to acquire data on visitors by analyzing the video stream at the point of sale. The analysis is done in real-time, without storing images for obvious reasons of GDPR compliance, thanks to cameras installed at the point of sale.

retail analytics

Artificial intelligence techniques, and more specifically computer vision techniques, make it possible to analyze these images with a high degree of accuracy in order to obtain data on the profile of visitors, but especially on their behavior, which cannot be done with Wi-Fi analysis. The cameras placed at the entrance of the stores allow to accurately count the visitors. In addition, the image analysis reveals interesting information on their profile (age group, gender, recurring visits) and the cameras installed at strategic points at the point of sale transmit data related to the behavior of the visitors. Thus, the customer journey, the commitment for a particular category or article, the satisfaction or the waiting period are all data that the video analysis makes possible to acquire. This data is compiled automatically and presented within reports and dashboards.

In conclusion, we can say that the in-store analytics collected via video flow analysis by computer vision rivals web analytics. They go even further, as a more accurate profile of visitors is established, while on-line shared computers are numerous (which distorts profile data) and private sessions prevent the collection of unregistered users profile data.