From Frustration to Flow: How We Boosted AR Shelf Capture Completion Rates

Store 360 is feature full mobile app for merchandizers doing champion sales. But when they capture images for wider shelves it consumes their time more than expected. For this reason the business have brought AR Camera feature for seamless capturing.

Challenge

Though the feature supports continues capturing of shelves, user need to follow multiple instructions to capture perfectly. Currently the users are abandon this feature without using it to capture the wide shelf though it consumes less time for capturing the wide shelf.

Role, Duration & Process

Role

Lead UX Designer

Duration

6 Weeks

Process

Non-Linear

Discovery

Usability Testing

Participants: 5–6 users who are both familiar and unfamiliar with the AR Camera.
Primary Users: Merchandisers responsible for capturing shelf images.
Criteria:Must have attempted the AR Camera at least once. Mix of users who abandoned quickly, those who partially completed tasks, and a few who adopted successfully.
Demographics: Range of device types, regions, and experience levels.
Tasks: Capture a wide shelf SKU using AR Camera vs. normal camera.
Measure: Time on task, number of errors, completion rate, and frustration moments (think-aloud protocol).

Understandings from Usability Testing Screen Recordings

Feature Discovery & First Impressions
- Several participants hesitated at the entry point of AR Camera (Do not know how to detect a plane).
- Users often tried tapping the screen as if it were a regular camera, showing confusion about the AR-specific workflow.
- Lack of an intro tutorial or inline guidance led to trial-and-error behavior.
Calibration & Setup Frustrations
- Most participants struggled with the calibration step (e.g., aligning shelves, moving phone left/right).
- Users repeatedly waved their phone around but received low-overlap / poor-detection errors, causing visible frustration.
- Some quit halfway through calibration, saying it felt “time-consuming” compared to just snapping a normal picture.
Task Completion & Abandonment Triggers
- On average, users took 2–3 times longer to capture a wide shelf via AR than via standard camera.
- Frequent error messages (e.g., “low overlap,” “adjust device angle”) made participants feel they were failing the task, leading to abandonment.
- When errors stacked, participants instinctively switched to normal camera as a workaround.

Environmental Constraints
- In dim lighting or crowded shelves, AR detection was less accurate, forcing users to restart capture.
- Narrow aisles limited participants’ ability to step back far enough for the AR capture, resulting in incomplete scans.
- Background noise (customers walking, manager pressure) pushed users to prioritize speed over accuracy, often abandoning AR altogether.
Mental Model Gaps
- Users expected AR to be as quick as snapping a photo, not a multi-step scanning process.
- Several assumed AR would automatically stitch shelves into a panoramic image, and were confused when they had to guide the capture manually.
- Participants didn’t clearly understand the value-add (e.g., why AR was “better” than just using the phone camera).
Positive Observations
- Once calibration was successful, AR capture produced clearer, wider shelf coverage, which users acknowledged as “useful.”
- Users who successfully completed the task said they would use it again if it were faster and simpler.
- A few participants showed interest in combining AR with standard capture (fallback option).
User Suggestions Captured During Testing
- “Just let me take 2–3 photos and auto-stitch them.”
- “Show me a ghost outline of how far I need to move.”
- “Give me a skip button if AR doesn’t work in this store.”
- “Explain why AR is better than a normal photo before I start.”