By implementing an immersive virtual-reality environment, some AI applications claim it is possible to test products or retail ideas that have not been brought to on the market, putting them on a virtual shelf to study consumer reactions and behavior to real-time merchandising.
By integrating an eye-tracker with a head mount display, these companies claim it may also be possible to monitor the consumer gaze and behavior in regard to certain products.
According to a study by Deloitte, worldwide revenues from the virtual and augmented reality (VR/AR) market will grow from $5.2 billion in 2016 to more than $162 billion in 2020. This represents a combined annual growth rate of 181.3 percent from 2015 to 2020.
To date, eight of the world’s top 10 technology firms have invested in VR, including Apple, Google, Microsoft, Facebook, Samsung, and IBM. In retail specifically, some of the biggest supermarket chains have built their own technology labs to explore how VR and other new technologies can bring benefits to the business. These include Lowe’s Innovation Lab and Walmart’s Store No. 8.
The same study noted the following advantages of virtual reality:
- In-store engagement: Helps customers navigate a store and find products, gain store incentives or rewards as they move
- Product customization: Allows retailers and brands to visualize ideas before investing in execution
- Experiential campaigns: Immersing customers through gaming, storytelling and branded experiences
In this report, we explore some applications in subsections where VR is currently being used or studied, with the aim of helping business leaders determine whether their organization would benefit from this technology:
- Fitting and Makeup Previews
- Home Improvement
VR for Fitting and Makeup Previews
Companies that have developed virtual mirrors or fitting rooms claim they simulate the process of trying on clothes in a retail store. These technologies claim to allow retail stores to present more options to shoppers in less time.
While we have discussed virtual mirrors for clothing, we will highlight a few similar applications which claim to offer shoe and cosmetic previews.
The company claims that the technology uses face tracking or facial motion capture, a computer-vision technology that collects data from images and videos by mapping 64 data points of facial landmarks in real-time to determine head poses, facial expressions, the different skin tones on the lips and face, and facial contours.
The company also claims that the application’s face recognition algorithms are trained to determine age, race, gender, facial hair, and other attributes.
To use the application, users must download the app from the Apple or Google Play stores. From the open interface, the application scans the user’s face and then prompts the user to choose the product and shades to preview on their face. The catalog includes the products and shades of lipstick, eyeshadows, eyeliner, and foundations. The application also offers complete makeup looks that the user can try.
If a customer likes how a shade looks, they can purchase products directly from the application.
The 2-minute demo below shows how Image Metrics’ client, L’Oreal has implemented the technology into its own mobile app:
In the L’Oreal store, a shopper can opt to scan an actual item to display it on the screen and try it on virtually instead of using the product testers, according to the company. When the image of the product appears on screen, a shopper can also test its available shades.
According to the L’Oreal case study, the algorithms are also trained on the customer’s real-time data to deliver relevant recommendations. For example, if the customer has been exploring “smokey eye” makeup, the application will display more products that can be used to create a smokey-eye look.
Image Metrics claims that the app was downloaded 35 million times and resulted in Improved customer experience, engagement and loyalty although no actual numbers were released.
The company also claims that Nissan and Hooya also use its face recognition
Dr. Kevin Walker is the Founder and Chief Technical Officer of Image Metrics. He received his doctorate in computer vision from the University of Manchester. He is also on the advisory board of Faceware Technologies.
Trya Srl created Snapfeet, a mobile application that claims to allow shoppers to virtually fit shoes, based on a 3D biometric scans of their feet. After scanning the shopper’s foot, the application is able to recommend a shoe size and model that best fits the the customer, according to the company.
Using photos of the customer’s foot taken from several angles, the application’s biometric capabilities will create a 3D image of the foot.
This short video demonstrates how a shopper would take the photos from all angles:
Once the images are saved within the app, its algorithms will arrange the pictures and send a notification once the biometry is ready.
Trya claims the app will then reference those biometric measurements each time the user logs in to shop for shoes.
The app also uses deep learning algorithms to place the shopper’s foot image into the 3D image of the shoe, as shown in this 1-minute video:
The company has not made available any case studies, but claims that most of its customers are in the Riviera del Brenta and Marches areas of Italy. The company website reports that one of its shoe retail clients is Spider Shoes.
Piero Donaggio, is the Co-Founder and Principal Scientist at Trya, where he leads initiatives focused on statistical data analysis, machine learning, full stack development and computer vision research. Prior to Trya, he served in machine learning-related roles at EnginSoft, VideoTec, and M31. He earned his Master’s degree in Telecommunication Engineering from the Universita degli Studi di Padova. It is not clear how directly his studies involved AI.
VR for Home Improvement
It is estimated that the number of VR users will increase to 171 million by 2018 from 200,000 in 2014. This interest in VR is compelling brands to explore the potential of VR to create an immersive marketing and shopping experience.
According to the company, the Aria Player’s algorithms can be applied to mobile phones, personal computers, and VR headsets.
Swarovski and Mastercard launched a virtual-reality shopping app for the Atelier Swarovski home decor line designed in collaboration with world-renowned architects and designers. Using Aria, Swarovski says the shoppers are immersed in a home where they can browse and purchase the pieces with Masterpass, Mastercard’s digital payment service.
In the 1-minute video below, a woman uses headset to access the app, walkthrough a virtual room, and purchase jewelry using head gestures and gazes:
After launching the app on a VR headset, Mastercard says shoppers will be asked to begin a secure session by providing their Masterpass account details. Once logged in, they will be able to navigate the home’s different rooms and see the collection by moving their heads left to right or up and down.
The app’s eye-tracking sensors determine which object the gaze is fixed on, which prompts the app to display product information, pricing, and videos about the design inspiration and manufacturing process.
After selecting a piece, the shopper can add an item to their cart or purchase the product by focusing their gaze on the cart or Masterpass buttons at the bottom of the product description. When the headset is removed, the app automatically logs the shopper out of their Masterpass company, according to the company.
Eye tracking technology informs the device where exactly the individual’s gaze is focused, according to the company. This data is then analyzed to determine the person’s presence, attention, focus, drowsiness, consciousness or other mental states or intent.
YouVisit claims to also serve Yale, HP and the US Army as clients.
We were unable to find any C-level executives with AI experience on the company’s team. However, one of its data scientists, Zhen Liu, earned a doctorate in Geophysics from Yale University and attended a data science program at The Data Incubator. Earlier in his career, he was a research assistant and teaching fellow at Yale. It is not clear if that position involved AI.
Specular Theory’s FAB (Families and Babies) is an eCommerce platform that claims to help parents find baby-proofing products from an online catalog using VR, AR, and AI.
FAB was chosen as one of five finalists in the 2017 Innov8: V-Commerce competition presented by Store No.8. Store No.8 is Walmart’s technology lab. Formed in 2017, the lab claims to “incubate” startups that are developing new technologies for retail.
No details were announced regarding the FAB application, but according to the Spectacular Theory website, FAB’s VR technology could enable families to identify child safety hazards around the home and provide product recommendations.
The FAB application claims to consider the child’s age and other contexts when recommending products and can be virtually tested through a Samsung headset or mobile phone before the parents purchase products.
To use the application, the website explains that the user must wear the Samsung Gear VR or download the app on a Samsung mobile phone. Samsung Gear VR is powered by Oculus which is driven by 3D pose reconstruction algorithms.
As part of Facebook’s initiatives, Oculus is also building other AI capabilities into its headsets, such as eye- and hand-tracking capabilities into its headset.
Specular Theory reports that they will expand to more platforms, including IOS.
The FAB site claims that the application’s algorithms will also have the ability to identify the potential child hazards at homes such as electrical outlets, stovetops, and door handles The application’s computer vision will then highlight the hazards within the VR to show parents what items and places need baby proofing, according to FAB.
The company says the application will layer these hazards with recommended childproofing products that parents can use to potentially make the highlighted areas safe. The recommended products will be accessed from Walmart’s online catalog and can be purchased within the application, according to Specular Theory.
We could not find a video demonstration or a high-res product image.
Spectacular Theory also claims Dolby, Sony Music, Accenture, Youtube, and Google as its clients.
While Spectacular Theory is supported by Store No. 8, we could not find any C-level executives on the team with an AI background. We caution readers to be wary of companies that claim to do AI without any C-level AI experts on their team.
VR for Merchandising
InContext developed ShopperMX, a virtual tool that allows for retail and brand teams to simulate new retail concepts before creating anything in the physical world.
On the company website, Shopper MX claims that it can provide retailers with a cost-effective way to test new ideas — such as new product displays, packaging; store layout; or signage on the shelf — to see how they will translate inside the retail store.
Aside from testing new store concepts, the company says another use case for Shopper MX is A/B testing. Retail stores and brands can create different versions of a concept to see which will be better received by shoppers, according to the company. The company claims its software’s machine-learning algorithms are trained to recognize shopper behavior patterns by collecting data about the gaze, store flow or aisle changes.
The 3-minute video below shows how the application can be used to simulate store layout and product displays:
Within the virtual environment, shoppers can interact with items by clicking on them and seeing the product attribute and pricing. The company claims it can also simulate adding items to the cart and the check out process.
According to the demonstration video below, a dashboard will show data about the virtual shopper’s behavior and which version performed best in terms of visibility:
The company claims that the software works on PC, mobile, full-wall screens and VR headsets.
InContext says Kellogg initially used physical test rooms and shopping labs for simulations to study shopper behavior but found that they were expensive and did not produce realistic results. According to the case study, the company turned to Shopper MX.
InContext says Kellogg wanted to test in-store displays in time for a back-to-school campaign. Using Shopper MX, the company reportedly created three versions of virtual displays, designating them as gold, silver, and bronze to represent the cost it would take to set up a physical display. Prior to the test, the company thought the gold display would fare best. Context said Kellogg through it was the biggest and most attention-grabbing of the three.
However, the study showed that the bronze display drew the most attention among the test participants, saving the company 50 percent on our display costs. The company, however, did not provide specific numbers regarding to the size of the study, duration, and actual savings.
InContext’s other clients include Samsung, Wendy’s, Johnson & Johnson and Coca-Cola. The company has raised $42.5 million in venture funding but were unable to find any C-level executives with AI experience on the company’s team.
Takeaways for Business Leaders
From this research, we found a number of similarities among the companies that develop VR applications covered in this story.
The companies above were generally creative design firms that focused on VR, and most were not transparent about how AI was used in their technology. We also found that the leadership teams employed technology executives with image processing or creative technology expertise.
We see that the companies we covered developed the software but relied on third-party headsets such as Facebook’s Oculus and Samsung Gear VR. Some applications also render across personal computers and mobile phones.
One challenge to the advancement of business-to-business VR technologies is the cost of VR headsets, other equipment, and software development. When researching what companies charge for their products, we found that a Samsung Gear VR headset is $129, an Oculus version is $399, and an HTC Vive headset is around $499. At the low end, the Google Cardboard costs $25 to $30, according to Amazon listings.
Header image credit: Pixabay