Selected work 01
giathuoctot.com
Understanding the why before redesign
I researched what was getting in customers’ way, chose a fix we could ship and measure sooner, and designed a one-tap voucher button.
- Role
- Lead Digital Customer Experience
- Dates
- 10/2025 – 06/2026
- Type
- B2B e-commerce · Pharmaceuticals
- Team
- Vu, UI Designer. Collaborators: Engineering, Customer Service, Marketing, Finance
The context
giathuoctot.com is Duoc Vuong’s e-commerce platform. Pharmacies and drug counters across Vietnam use it to restock medicine, usually about once a week. I joined in October 2025 as the first design and customer experience hire, with a full v2.0 redesign already planned.
- My role
- Research, prioritization, UX and UI, requirements, stakeholder alignment and measurement.
- Teammate
- Vu, UI Designer.
- Collaborators
- Engineering, Customer Service, Marketing and Finance.
Step 1: Listening
A survey of 300 customers
I wrote and distributed a survey, collected 300 responses from 3–10 December 2025, grouped the open answers by theme and compared ratings across app, mobile web and desktop.
- Were satisfied
- 84%
- Found the site easy to use
- 87%
- Understood promotions, the lowest rating
- 69%
- Named promotions as a difficulty
- 28%
- Named promotions as a reason to choose the platform
- 14%
The customer service team
- Customers missed the small voucher button and placed orders without a voucher.
- They called Customer Service to add one, or cancelled and ordered again.
- Many did not know that several vouchers could apply to the same order.
A screen-by-screen audit
I logged 105 issues across 15 areas, prioritized each one and assigned it to myself or Vu.
- Missing states
- 75
- UX problems
- 25
- Missing flows
- 5
Step 2: Choosing what to fix first
I weighed how often each problem came up against the work required to fix it. Search was the most common complaint, named by 33% of survey respondents.
- Vouchers affected the moment before ordering. Work was concentrated in the cart and voucher list, and we could compare order data within weeks.
- Search needed deeper backend changes. Engineering rebuilt it in parallel while I wrote the customer-facing requirements.
We shipped Cart and Vouchers before the rest of v2.0.
- The cost
- Old and new screens sat beside each other until the remaining redesign shipped.
- The gain
- Earlier measurement and evidence for the remaining redesign.
Step 3: The problem up close
Step 4: The solution, one tap
The decisions behind the button
I designed Auto-apply available vouchers: one tap applies every voucher an order qualifies for.
Apply all eligible vouchers
- Why
- Vouchers on giathuoctot.com can stack. Applying every eligible voucher, or the higher eligible tier within a group, gives customers the best discount.
A button, not silent auto-apply
- First idea
- Apply vouchers automatically when products changed.
- Why not
- The server could not recalculate eligibility on every change.
- Trade-off
- Calculation stayed on demand, leaving customers one tap to make.
Keep the list, but fix it
- Why
- Some pharmacists wanted to check vouchers themselves.
- Trade-off
- Two paths to maintain.
- What changed
- With Vu and Engineering: clear amounts, descriptions and expiry; conditions and applied products on tap; near-expiry warnings; grouped vouchers.
Fix the data, not just the screen
- Why
- Finance had put discount amounts and conditions into voucher titles, so the product could not show them properly.
- What we did
- With Finance and Engineering, I gave those details their own fields.
- Trade-off
- Work beyond the interface, but the product now shows the details and Finance manages the data more easily.
Final design
The numbered markers identify the changes in the original design.
The impact
Comparing the month before launch with the month after:
- Vouchers used per order
- +40%
- Items per cart
- +30%
Cart also introduced an upsell banner above product cards. Neither increase can be attributed to the voucher button alone.
What I learned
- Complaint counts are one input to priority. A smaller fix can give us evidence sooner.
- Customer Service explained the behavior behind survey scores.
- Internal workarounds can expose customer problems: Finance’s titles revealed a data issue that needed cross-department work.
- Early results helped me make the case for the remaining v2.0 redesign.
Keep exploring
Other projects

Web app · Property management
Trudi Portal
One place for a fragmented working day
Explore the project
Website · Property management
Trudi Website
Showing the product through its website
Explore the project
Mobile app · Property management
Trudi Tenant & Landlord App
From separate requests to one conversation
Explore the project
Mobile app · Healthcare
Hough Rapid
Clarity through an at-home test
Explore the project




