Trax can identify over a billion retail products from a smartphone photo.
Using image recognition and machine learning, a sales team from a manufacturer like Coca Cola – one of the company’s clients – can snap a photo in a supermarket chain and within 10 minutes have data back on whether the retailer is compliant with the display terms they agreed.
Even with mustard bottles half-turned away and shade on the wine bottle label, it has a 96 percent accuracy rate.
The Singapore-headquartered company, which an R&D centre of excellence in Tel Aviv, can also use its data to gather intelligence on more esoteric aspects of retail performance. As Martin Smethurst, Trax’s MD for EMEA tells Computer Business Review in an interview at our offices, products behave very differently on the shelf.
“Data shows that some will sell higher volumes if you stock more; they’re elastic. Others will sell the same volumes even if you reduce shelf space; they’re inelastic. Trax technology helps retailers and manufacturers identify precisely which those are, so they can get the most out of their share-of-shelf.”
With the company today announcing its largest investment round to date – raising $125 million in a round led by Boyu Capital, one of the largest private equity investment firms in Greater China – both investor and manufacturer interest is growing. The company, founded in 2010, has raised approximately $235 million in total funding and now operates in over 50 countries with more than 175 client engagements.
The attraction for manufacturers is clear, Martin Smethurst says: “Manufactures can find it really hard to understand whats happening on shelves. They understand what they send to a retail store and how much stock it’s got, through the reports they get, but not whats going on on shelf itself. They’ll send in sales teams to check layout and report back, but they are about 60 percent accurate at best and the process is clunky.”
A Trax camera on a shelf
From Smartphones to Smart Shelves
Trax is also running a “Shelf Intelligence Suite”. This involves continuous shelf tracking with wireless IoT cameras for retailers; the aim is to help ever-watchful manufacturers keep closer track of performance, spot gaps in shelves and optimise sales.
A pilot in the UK is focussing on wines and beers. A small camera on the opposite shelf takes photos of the segment. The cameras were designed in house by Trax and designed to sit on a shelf for years without needing a change of battery.
As Martin Smethurst puts it to Computer Business Review: “With retailers picking for e-commerce buyers, gaps are appearing in shelves much faster. We can start to share where those gaps are immediately and link in to the supply systems; there’s a reasonable chance it’s not out of stock, it’s stuck in a back office. Our aim is to digitalise the shelf; almost become the Google of the shop shelf.”
He adds: “Ninety-five percent of our business right now is manufacturers. We’re moving into the retail space too; retailers could for example gather data, process it on our platform and sell it back to manufacturers.”
With the main Trax product downloadable as an app for end-users, it’s also spawning a nascent crowd-sourced retail intelligence industry.
“Where industry doesn’t have a sales team, they use auditors. Now it’s very easy to spin up a crowd to capture data; they just download our app and have a strict schedule on when to visit and the image quality provided. It has really helped industry broaden its reach.”
The company’s new investors are buying into the vision: Joey Chen, MD of Boyu Capital, said in a release: “The investment in Trax is driven by our conviction in the compelling value proposition of new technology solutions that enable digitalisation of brick and mortar retail.”
He added: “We are impressed by the wide recognition of Trax’s cloud-based one-stop-shop solutions by the world’s largest CPG companies and retailers. Compared to developed markets such as Europe and the United States, the use of digital retail solutions by Chinese consumer goods companies and retailers is still in its early stages.”