Aura Vision, which is part of the current batch of startups at Y Combinator, helps retailers understand who’s visiting their stores and what they’re doing there.
In other words, if you want to see the demographics of who’s visiting the store, or which displays and products are actually prompting customers to linger, or how long customers have to wait in line, Aura Vision can use existing security camera footage to tell you.
“We are focused on specialty retail — everyone on the retail market that isn’t grocery,” CEO Daniel Martinho-Corbishley told me. “We provide them with insights to help them innovate successfully.”
The company was founded by Martinho-Corbishley, CTO Jamie R. Lomelí and CPO Jonathan Blok. Martinho-Corbishley said he and Lomelí both did Ph.D. research at the University of Southampton on machine learning and computer vision, and they “saw the potential for deep learning in the retail industry,” particularly after they “had a look at what else is out there.”
There are companies trying to use security footage to provide in-store analytics to retailers — for example, there’s Prism Skylabs, which launched at Disrupt in 2011 and is backed by CrunchFund. Others are using technology like Wi-Fi and Bluetooth to provide similar data.
However, Blok pointed out that installing new sensors in a store can be “a big upheaval.” With Aura Vision, on the other hand, retailers either use the security cameras they’ve already set up — or if they do need to install new cameras, “you’re going to get a security system” out of the process.
In addition, Martinho-Corbishley pointed to the sophistication of Aura Vision’s technology, which can provide “very precise and accurate insights out from the camera themselves — any camera in the store.” That includes distinguishing between staff and customers in the footage, and determining the demographics of a customer, even if their face isn’t captured.
As for what this kind of analysis does to customer privacy, Martinho-Corbishley noted that the company was “born at the time of GDPR.”
“In that very first year, we made a decision very early on to not identify anyone, so the data that we provide back to our clients is entirely anonymized,” he said. In other words, it will describe the behavior of your customers in aggregate, but “we never link that to the person’s identity.”
Aura Vision charges a subscription fee based on the number of cameras a customer is using each month — something that Martinho-Corbishley said is “a very simple charge” without “crazy hidden fees or crazy retainers.”