The ask
An industry-leading soft drink and food manufacturer had millions of IoT connected fridges with the ability to remotely monitor and report on various attributes such as location, temperature, door status, stock levels, compressor health, and much more.
Within CPG industries, it is well known that the shelf placement of products has a huge and direct impact on sales numbers. The client was looking to leverage this, however their existing machines didn’t yet have the ability to monitor which items were already stocked within individual units, or the placement of each item at any given time.
They asked us to create a solution that would help to identify which items were placed in each tray on the fridge. Using this information, the company was capable of taking actions accordingly.
What we did
We worked with the client and their IoT provider to integrate a retrofitting camera that would photograph the contents of the fridge each time the door was opened.
The photographs were automatically uploaded into the Cloud, where we built a system that would analyze each image to identify the type of product and its position on the shelf.
The process
As the solution used photographed images to monitor stock levels, we decided that the best approach would be to use Intelligent Process Automation which incorporated Deep Learning techniques and Convolutional Neural Networks.
This information gathered from these techniques enabled the client to gain a further understanding of how their products are stocked in each fridge, and adjust any incorrect placements.
Summary
By enabling remote management of planogram compliance across all units, this solution gave the client easy and comprehensive control over their stock. This allowed them to better manage their retail partners with new incentives/penalties that would help maximize product sales. They have also saved time and resources by removing the need to visit stores and now have access to 24/7 monitoring of their assets.
Results
ACCURACY OF BRAND AND LOCATION
Additional uses
Solutions like this one would benefit any business with a need to monitor planogram compliance, or a need to visually check their stock automatically. Using Intelligent Process Automation is a powerful way to achieve this, as it has the ability to break down complex images into constituent parts and pull out data from each part of the image.
