UX Research
Published:
2026/06
Tools
Claude in Chrome
Workflow
1. Define the research objective
Start with a clear question to avoid collecting irrelevant insights. For this experiment, I wanted to understand which project entry point attracts the most traffic when visitors decide to explore additional work.
2. Map user journeys and research scenarios
Identify all available project entry points and document how users may navigate through the portfolio.
Entry points
Homepage
Project List page
"More Projects, Broader Impact" section at the bottom of each project page
Research purpose
The purpose of this research is to understand how users discover projects and identify opportunities to encourage deeper exploration.
Initial assumptions
Featured projects are not visually emphasized on the homepage.
The recommended project section lacks enough visual distinction.
Project categories are lack of the project list page.
3. Define the expected deliverables
Provide a structured research report based on simulated user behavior.
The report should include:
Entry-point performance comparison
Desktop vs. mobile behavior
Project discovery patterns
Key friction points
Actionable design recommendations
4. Conduct AI-simulated user testing
Use Claude Code to simulate 100 participants with different browsing behaviors and goals. Analyze how they navigate the portfolio and which paths lead to project exploration.
5. Synthesize findings and opportunities
Summary Table
Entry Point | Traffic Volume | Engagement Depth (projects/session) | Primary Friction | Priority Fix |
|---|---|---|---|---|
Homepage | High (52%) | Low (1.2) | No hierarchy, no "View All" CTA | Featured hero tile + View All CTA |
Project List Page | Medium (31%) | High (2.9) | No category filter | Clickable tag filters |
"More Projects" Section | Low (17%) | High (3.0) | No exit to full list, mobile UX | Add "See all projects" link |
Claude identified six friction points, but three assumed the portfolio would continue growing. Since my strategy is to keep the project list intentionally curated and focused on featured work, those suggestions were not aligned with the current direction.
Key Friction Points
Friction 1 — No Visual Hierarchy on the Homepage
Users described difficulty understanding which projects were most important or representative of Tang's best work. The only path to the full list is the nav bar, which some users overlooked.Friction 2 — No "Back to Projects" or Breadcrumb Awareness at Project Bottom
Users who finished a carousel item and wanted to return to the project list had no immediately obvious path downward. They needed to scroll back up to the nav.Friction 3 — Mobile Project Placement
First-time mobile visitors who do not immediately notice the nav bar's "Projects" link may have a session in which they never encounter any project content.
Actionable Design Recommendations
Action 1 — Elevate One Featured Project on the Homepage
Redesign the homepage bento grid to give one project a larger, visually distinct "hero" tile (approximately 2× the current tile width). Include a brief and a clear "Read Case Study →" CTA.Action 2 — Add a "View All 6 Projects →" CTA to the Homepage
Place a clearly labeled call-to-action button below the project tiles on the homepage, like "See All 6 Projects".Action 3 — Add "View All 6 Projects" Link Inside the "More Projects" Section
Beneath the "More Projects, Broader Impact" carousel, add a secondary link: "See All 6 Projects →" on the project page.Action 4 — Prioritize Mobile-First Project Visibility
On mobile, place at least one project tile (or a "See my work →" CTA) above the skills animation in the stacking order. Currently, mobile visitors scroll through the hero intro and skills ticker before seeing any project.
Reflections
Claude identified several valuable friction points, while some recommendations conflicted with the product strategy. For example, it suggested solutions for managing a larger project library, whereas my portfolio is intentionally curated to keep attention on a smaller set of featured projects. The experience reinforced that AI can generate insights, but designers still need to evaluate them within the context of product goals and strategy.