Oh no, another COVID-19 post. Well hopefully not, but I admit that enforced confinement to my house, and my experience of trying to shop has prompted this post.
Anyone who has tried to buy food in the last month will have experienced an empty shelf or two. In fact one report in retail-week quoted that 72% of us have experienced this (personally I think it is an understatement). Out of Stocks (OOS) as the industry calls it has risen significantly since mid February, a report from Edge by Ascential has shown OOS has stock has soared from and all category average of 1.5% to over 12% across all the majors on 18/3/2020.
In normal times OOS is lost revenue, today it means in some cases people don’t get the food/goods they need. Thankfully supermarkets have implemented demand side restrictions (rationing to you and I) to ensure availability.
But what I want to talk about is the next generation of OOS management, powered by AI and computer vision.
At Neurolabs, we have implemented a highly accurate system for precisely measuring real-time OOS that doesn’t rely on humans visually inspecting shelves or scanning barcodes to manage replenishment. Using inexpensive, IoT sensors embedded in the shelves our system can accurately identify where there are minimum presentations of an item on the shelf and report in real-time alerts not only to the local replenishment team, but to head-office to show an aggregate availability.
Today the process has a significant lag in the system, primarily down to the manual way it is implemented. First a staff member walks the isle, noting stock. Today this is once a day at best (other than fast moving lines such as milk). If it is in the back room it is replenished overnight ( this is why we see the 8am till 8pm restrictions at the moment) if its not in the back room, someone has to enter an order into the replenishment system (usually when the assistant manager has an hour). Depending on what is already on that stores lorry already and stock at the depot it could be several days before that item returns to the shelves.
What if..
The sensors identify and report back out of stocks to the replenishment teams in real time? So there is no need to close and replenish overnight.
What if..
The the data is available to not only the store manager, not only to the head office, but to the logistics supply chain so that the velocity of a particular line is known accurately? So that stocks can be redistributed appropriately at the warehouse.
What if..
The data is converted via the supermarkets forecasting tool into an automatic order to the supermarkets supplier? So that manufacturers can plan ahead with confidence and react accordingly.
None of the “what-ifs” are particularly difficult to do, in fact they are routinely done in large scale manufacturing, like the automotive industry. What is different is how it can revolutionise the grocery industry. The pandemic we are dealing with has drawn into sharp focus some of the holes in the replenishment process. Perhaps now, while many of us have the time, is the moment to plan that future and look to improve the process and with it the profitability of the grocery industry.