User Behavior Analysis

User Behavior Analysis

User Behavior Analysis

Identifying UX issues through data, forming hypotheses, and validating with usability testing.

Identifying UX issues through data, forming hypotheses, and validating with usability testing.

Identifying UX issues through data, forming hypotheses, and validating with usability testing.

2023

2023

2023

Background

This project focused on how users leave and switch between live rooms. Existing behavioral data suggested that the current exit flow was creating unnecessary friction.

Process

Pain Points

Inconsistent exit behavior across free and paid rooms

  • Free rooms supported both swipe and Close.

  • Paid rooms disabled swipe and required an additional confirmation step.

  • This made switching behavior dependent on room type.

Optimization Goals

  • Enable one-handed switching between live rooms to help users explore content more easily.

  • Reduce the time and hesitation to leave a live room.

  • Help users explore other live rooms with less friction.

  • Enable one-handed switching between live rooms to help users explore content more easily.

  • Reduce the time and hesitation to leave a live room.

  • Help users explore other live rooms with less friction.

  • Enable one-handed switching between live rooms to help users explore content more easily.

  • Reduce the time and hesitation to leave a live room.

  • Help users explore other live rooms with less friction.

We hypothesized that removing the confirmation step and enabling swipe in paid rooms would make switching easier without preventing users from returning.

Baseline Metrics

Tap Close → Confirm to leave

  • Paid live rooms: 87%

  • Early bird ticket buyers: 93%

  • Paid live rooms: 87%

  • Early bird ticket buyers: 93%

  • Paid live rooms: 87%

  • Early bird ticket buyers: 93%

→ High exit rates suggest the extra confirmation step isn’t effective in reducing exits.

→ High exit rates suggest the extra confirmation step isn’t effective in reducing exits.

→ High exit rates suggest the extra confirmation step isn’t effective in reducing exits.

Leave & Return Rates

  • Paid live rooms: 31% returned

  • Early bird ticket buyers: 78% returned

  • Paid live rooms: 31% returned

  • Early bird ticket buyers: 78% returned

  • Paid live rooms: 31% returned

  • Early bird ticket buyers: 78% returned

→ Users may leave intentionally and still return later.

→ Users may leave intentionally and still return later.

→ Users may leave intentionally and still return later.

*Data is for illustration only, processed and not real.

*Data is for illustration only, processed and not real.

UX & UI Optimization

  • Removed the leave confirmation popup.

  • Enabled swipe-to-switch across all live room types.

  • Added swipe prompts to improve discoverability.

Usability Test Design

Task
After purchasing access, try to leave and explore other live rooms.


- Version A
Original flow with confirmation popup, swipe disabled.

- Version B
Confirmation popup removed, swipe enabled with prompts.

*Usability Testing Tool: Maze.co

Task
After purchasing access, try to leave and explore other live rooms.


- Version A
Original flow with confirmation popup, swipe disabled.

- Version B
Confirmation popup removed, swipe enabled with prompts.

*Usability Testing Tool: Maze.co

Task
After purchasing access, try to leave and explore other live rooms.


- Version A
Original flow with confirmation popup, swipe disabled.

- Version B
Confirmation popup removed, swipe enabled with prompts.

*Usability Testing Tool: Maze.co

Test Results

  • Participants: 39

  • Direct Success Rate: A: 76.9% | B: 97.1%

  • Avg Duration: A: 44.6s | B: 8.7s

  • In Version A, even though swipe was disabled, many participants still attempted it by instinct.

  • 20% of participants in Version A failed the task, unsure how to leave.

  • All participants eventually completed the task in Version B, with a 97.1% direct success rate.

  • Participants: 39

  • Direct Success Rate: A: 76.9% | B: 97.1%

  • Avg Duration: A: 44.6s | B: 8.7s

  • In Version A, even though swipe was disabled, many participants still attempted it by instinct.

  • 20% of participants in Version A failed the task, unsure how to leave.

  • All participants eventually completed the task in Version B, with a 97.1% direct success rate.

  • Participants: 39

  • Direct Success Rate: A: 76.9% | B: 97.1%

  • Avg Duration: A: 44.6s | B: 8.7s

  • In Version A, even though swipe was disabled, many participants still attempted it by instinct.

  • 20% of participants in Version A failed the task, unsure how to leave.

  • All participants eventually completed the task in Version B, with a 97.1% direct success rate.

User Insights

  • Participants frequently attempted to swipe even when the gesture was disabled, suggesting that swipe matched their expected interaction.

  • Inconsistent swipe behavior across room types made the exit interaction harder to discover.

pinxichen.design@gmail.com

© 2026 Pinxi Chen

pinxichen.design@gmail.com

© 2026 Pinxi Chen

pinxichen.design@gmail.com

© 2026 Pinxi Chen