Augmented reality (AR) is not just changing the lives of video gamers. It also has the potential to solve many of the problems in the consumer goods and retail industries, including the last mile in retail execution.
Here’s how it can improve process efficiency, boost data accuracy and ultimately increase sales.
For all the recent innovations in retail, actual in-store execution often uses outdated methods. When CPG companies set out to track if their products are available on retail shelves in the right assortment, in the correct quantities, at optimum visibility and with the correct prices and promotions, many still send their sales reps to conduct manual audits. Observations on store conditions are captured on a notepad or, in some cases, using mobile applications. It’s time-consuming. It’s prone to human error. But thankfully, it’s starting to change.
An image worth 1000 insights: Disrupting retail execution with computer vision
Computer vision technology is disrupting the sector by allowing CPG brands to digitize the physical world of retail.
The foremost application of computer vision technology is image recognition (IR). IR uses digital images to provide manufacturers with near real-time and actionable insights into retail shelf conditions. These insights drive greater value by allowing field personnel to immediately fix out of stock issues, optimize product displays and improve planogram compliance. An IR-powered solution thus offers a more objective and consistent way to track in-store conditions compared to manual audits. Brands that use Trax Retail Execution, for instance, typically report 96 percent SKU level recognition accuracy and a 3–5 percent sales uplift in a single category. Expect these gains to increase, as Trax takes IR to the next level with AR.
Icing on the cake: Using AR to strengthen IR’s value
Currently, one of the biggest challenges of IR lies in accurately capturing the images of products on the shelves. The photos can overlap, leading to product duplication, or sections of the shelf can be missed altogether. In the absence of large teams of people who check for discrepancies, this results in lower data quality, and incomplete or inaccurate insights.
This can be overcome with AR.
With built-in AR capabilities that use visual inertial odometry (VIO), the Trax app’s Shelf Capture Assistant automatically detects which sections of the shelf are being captured. It displays this information visually by colouring the captured sections of the shelf on-screen, so sales reps can see at a glance if they’ve missed any areas. So, the next time a user is interrupted by a shopper passing by or a chatty store manager, they know exactly which portion of the shelf to resume taking pictures from, leading to faster audits. What’s more, since there are minimal overlaps to start with, the IT overhead involved in digitally stitching multiple images together is considerably lower.
Trax’ Virtual Ruler can also analyze the scene presented by the camera view and find horizontal planes in the store. This means that it can detect shelves and floors, and use this data to work out the exact location and size of each SKU. Such uses of AR allow the Trax solution to provide more accurate data, which in turn gives more useful and faster insights. But this is just the beginning as far as what AR can enable.
Long-term vision: Unlocking AR’s true potential
Imagine a future where sales reps see instant corrective actions on-screen: For example, a graphic instructing them to move a certain product a few positions to the left to maximize visibility. This would allow them to make meaningful changes immediately, greatly improving the time from insight to action.
With features such as these – and many more yet to be imagined – AR can solve the challenges of retail’s last mile. By providing an augmented digital solution to what has always been an analog problem, Trax is bringing retail execution into the 21st century.
For more information and to see Trax’s AR-powered app in action, visit https://traxretail.jp/technology/augmented-reality/
