E-Commerce · Case study 11 of 12

AR/VR Commerce

Walmart's innovation team commissioned an AR/VR-enabled commerce experience to explore how immersive technology could personalise product discovery at scale — moving beyond keyword search toward preference-driven spatial shopping.

AR/VR Commerce — E-Commerce interface design by Yellowfirst
Client
Walmart Labs
Year
2022
Industry
E-Commerce
Role
AR/VR Experience Design · Personalisation · Ideation

The challenge

E-commerce search is fundamentally broken for discovery: it optimises for known items, not for preferences. Customers who don't know exactly what they want get lost. Returns rates for furniture and home décor — categories where AR could have the highest impact — ran at 18-22% due to size/fit mismatch that could have been prevented with spatial preview. The challenge was to design an experience that felt like a natural extension of how people actually shop, not a technology demonstration.

The solution

We designed an AR product discovery experience built on a preference graph — a living model of the user's aesthetic, spatial, and functional preferences that builds from implicit signals (dwell time, rotate gestures, save actions) rather than explicit ratings. Users can place true-to-scale product previews in their actual space via AR, and the preference engine learns from what they keep versus remove. A separate ideation mode lets users explore style boards that connect to shoppable products — bridging inspiration and purchase.

Our process

  1. 01 Preference Modelling Research

    Studied how people make furniture and décor purchase decisions. Identified 7 implicit preference signals available through natural AR interaction.

  2. 02 AR Interaction Design

    Designed touch, gesture, and spatial placement interactions specifically for ARKit/ARCore constraints. Every interaction validated on device, not in Figma.

  3. 03 Preference Graph UX

    Designed the visible surface of the preference engine — how does the system show users that it understands them, without being creepy?

  4. 04 Ideation Mode

    Built a style board experience that bridges editorial inspiration with shoppable inventory, powered by visual similarity search.

  5. 05 Returns Impact Testing

    Measured AR preview impact on return rate in a controlled study of 400 participants. Validated at scale before broader rollout recommendation.

Outcomes

  • 61%Reduction in size/fit return reasons for AR users
  • 4.2×Higher session depth vs. standard browse
  • 7Implicit preference signals captured per session
  • 2.8×Conversion rate for AR-previewed products

Technology

  • ARKit
  • ARCore
  • React Native
  • AWS Personalize
  • GraphQL

Read more