How Technology Can Triple the Profit of a Grocery Store? PART 2
Updated: Aug 23
By Jussi Nummela, CTO, Blockstore Group
This blog text is divided into two parts, this being PART 2 and PART 1 was published in November.
PART 1: General tech benefits, RFID, and sensors
PART 2: Character tracking, robotics, image recognition, mobile apps, edge computing, and cloud services
After previous PART 1, I here describe the rest of Blockstore’s technology stack and summarize both parts.
When a customer has entered a Blockstore-powered store (with a mobile app or payment card), character tracking is used to determine, which cabinet she/he is approaching to pick up the desired products.
Within a fraction of seconds, algorithms in our edge computing resource conclude if and when buying process activation should be initiated, and customer can simply open the cabinet door, pick the products they like and continue to the next cabinet. Simple as taking milk from your home fridge or bread from the kitchen cabinet.
An important thing I do wish to emphasize is that our system is tracking a character. In other words, no personal information is handled.
Robotics is something the grocery business is already now very familiar with. More and more processes over the supply chain are taken care of by robots, lately even the last mile delivery. However, robotic solutions are mainly used in warehouses, DC, or MFC.
In Blockstore solution robotics are also used in those same locations. That is how product tagging can be done in the most effective way. RFID tag is added to each product and needed product data is encoded (such as unique ID, product information, and best-before date).
The main difference between our robot and existing tagging/labeling robots is that we provide one single robot that can handle approximately 1000-2000 different products. Multiple batch sizes, multiple SKUs, and all in the same operational setup. To compare, typically traditional tagging or labeling systems attach labels to one or a few different types of products with strict form factors.
The main challenge for the robot with these 1000-2000 SKUs is that packages are obviously multiform and different sorts of materials, i.e., bags, cans, bottles, cartoon boxes, etc., and can also be covered with frost and moisture taking place in cooled environments.
These kinds of robots we did not find ready-made anywhere and therefore we decided to develop tagging robot at Blockstore (if someone already has this kind of robot for sale, please let me know a.s.a.p.). Our robot is not production ready yet, but we expect to launch it in H2/2023. Until then we will provide our other tagging solutions for use: for example, our Tagger 1.0 can serve up to 20 stores from one tagging hub.
In addition to robotics at the warehouse, and stores we also have other thrilling robotic topics in our development roadmap to increase the efficiency of autonomous grocery store. I will share more information later.
Image recognition is already today used in cashier-less environments in grocery retail. At Blockstore we use image recognition with 3D cameras to expedite product tagging within our robotized tagging process.
The robot uses a rendered 3D model of every single product not just to confirm the product itself, but also to detect the position and orientation of the package. This information is then used in our supply chain, in the store, and in overall planning.
We call our own set of software applications “BLOCKSTER Platform”, and it’s been built to fulfill the needs of automatic and autonomous grocery stores. Blockster platform consists of server and separate applications enabling for example tagging, shop-floor operations, buying by app, buying by card, customer service, and lunch service operations.
Even though they are separate applications, they all work together with the Blockster cloud server.
With efficient software architecture management together with utilizing the latest SW development tools, the development cycle becomes short.
It is then often very cost-efficient to simply create dedicated custom-made applications for certain need rather than source and modify from the market. Especially when creating something totally new.
As described earlier, our Walk In – Walk Out shopping process utilizes character recognition. The buying process initiation is very delay sensitive and therefore we do part of the calculation locally on-site.
The character recognition algorithm runs on an on-site middleware server, which further is connected to cloud with the necessary connection.
We utilize Microsoft Azure Cloud services, and we are also a qualified member of Microsoft Startup Program. The main part of our system, Blockster Server, is located in Azure Cloud which brings us easy access to many crucial components of our automated grocery store system, such as databases, robotic process automation, user authorizations, integrations, user interface, reporting, data analysis, etc.
In addition, cost-effective data integrity, backups, and multi-tenancy capabilities are the functionalities where we trust in Redmond.
To summarize, despite all the technological details explained above, the main purpose of all this is to revolutionize the profit of local grocery stores.
We have used quite a lot of time and effort on profitability calculations during the last 12 months and verified even the smallest details with not just one or two, but four different grocery giants in Europe. Separately.
After all these efforts I can proudly declare that when the presented technology stack is working seamlessly together, it really is possible to achieve even triple profit compared to what grocery stores are doing today.
Please contact us for further opportunities!