Your Orders
Lead Product Designer • iOS, Android, responsive web
An app and web experience that helps patients check their prescription orders and know what to do next.
Business context
1.8B+ prescriptions filled in 2025
60M+ digital customers
No. 1 U.S. pharmacy by prescription sales.
The opportunity
💊 For patients
Understand prescription status.
Know when action is needed.
Manage orders without calling the pharmacy.
👩🏻⚕️ For CVS
Improve patient satisfaction.
Keep customers coming back.
Reduce pharmacy calls and costs.
The problem
Usability testing with 10 participants revealed three issues in the designs I inherited:
01 — 🧠 Status clarity
Important prescription details were overlooked.
02 — 🔧 Self-service
Similar button labels made some actions unclear.
03 — 🤝 Trust and confidence
Conflicting status messages and timelines created confusion.
My contributions
Product design
Interaction design
Usability research and testing
Design leadership and mentoring
Cross-functional collaboration
AI concept development and prototyping
User research
My approach
Defined two scenarios: One complex and one simple to evaluate different prescription situations.
Led prototype development: Guided a designer in creating prototypes for both scenarios.
Partnered with UX Research: Helped define recruitment criteria and interview questions.
Helped setup the study: It included ten moderated remote sessions through UserTesting. We varied which scenarios participants saw first to reduce the influence of seeing one before the other.
Research objectives
🧠 01 — Status clarity
Can patients understand their prescription status and important details?
🔧 02 — Self-service
Can patients identify and complete actions independently?
🤝 03 — Trust and confidence
Do status messages, timelines, and system feedback help patients understand what happens next?
Complex
Simple
What we learned
I rated usability issues on Nielsen Norman Group's 0–4 severity scale considering how often they occurred, how difficult they were to overcome, and whether they kept causing problems.
01 — 🧠 Status clarity
Participants focused on high-level messages, overlooking important prescription details.
Severity: 3 (major usability problem)
Recommendation: Make the prescription issue and required action clear in the primary status message.
"It just says 'not filled, action required'… Oh, OK, now I see it says it's expired."
02 — 🔧 Self-service
Two of 10 participants were confused by the difference between "Options" and "Review options."
Severity: 2 (minor usability problem)
Recommendation: Use clear, distinct, action-oriented labels.
03 — 🤝 Trust and confidence
1 of 10 participants overlooked a required action on an insulin prescription that also displayed a “Ready by” date.
3 of 10 participants were confused by the date.
Potential risk: Conflicting messaging could lead someone to expect their insulin without realizing they needed to act.
Severity: 4 (usability catastrophe)
Recommendation: Prioritize required actions and hide potentially misleading pickup dates until the issue is resolved.
"It's not filled yet… It will be ready on December 14."
"So I put in a request to receive it early, I am not sure what exactly that will give me."
Redesign and ✦ AI exploration
After transitioning to another project, I independently revisited the research findings to explore design improvements and see how AI could help patients better understand their prescription orders and next steps.
Design decisions
01 — 🧠 Status clarity
Because participants often overlooked important details, I audited and rewrote some of the status messages to make specific issues more prominent.
Examples:
Before: "Delayed” tells you there’s a problem.
After: "Doctor approval required" tells you there’s a problem and why.
02 — 🔧 Self-service
Since participants were confused by "Options" and "Review options," I replaced these generic labels with clearer, action-oriented language.
Examples: "Manage order" and "Message pharmacy"
03 — 🤝 Trust and confidence
To prevent conflicting information from obscuring required actions, I redesigned the order cards to:
Remove misleading pickup dates when patient action is required.
Prioritize one clear CTA explaining how to resolve the issue.
Add a warning icon to highlight orders needing attention.
Simplify messaging so patients can quickly understand the issue.
One image could not load. Check both image URLs in the Code Block.
One image could not load. Check both image URLs in the Code Block.
04 — 💡 Ideas to test next
Beyond the research findings, I identified three additional opportunities to explore:
Color-coded parent status groups: Help patients scan and distinguish order statuses more easily.
Medication images: Help patients identify prescriptions more quickly.
Simplified status messaging: Remove redundant information, such as “Available now” when “Available until Wed, Dec 21” already communicates availability.
Exploring AI in the experience and design process
Product concept: Designed an AI assistant within prescription order cards to answer questions, explain statuses, and direct patients to pharmacy staff when human support is needed.
Design process: Used AI to explore concepts, challenge design decisions, and identify edge cases. Connected AI to my Figma designs through MCP and used Codex to build and refine the prototype.
Tested design
My redesign
Testing the redesign
I’m testing my redesign against the original to see if it fixes the three problems we found.
Version A: The designs we tested at CVS
Version B: My redesign
Who: 20 people who manage prescriptions online, 10 per version.
How: Unmoderated test with clickable prototypes in Lyssna
Setup: Each person sees only one version, so what they learn in one can’t help in the other. In each group, half start with the simple order and half with the complex one.
Tasks: The same tasks as the original study, reworded to fit both versions.
Research objectives
I’m using the same three goals as the original study and comparing both versions on each.
🧠 01 — Status clarity
Can patients understand their prescription status and important details?
🔧 02 — Self-service
Can patients identify and complete actions independently?
🤝 03 — Trust and confidence
Do status messages, timelines, and system feedback help patients understand what happens next?
I’ll also check whether people who feel sure are actually right, especially about their insulin.
Results
Results coming early November 2026. I’ll rate each problem on Nielsen Norman Group’s 0–4 severity scale, the same scale I used in the original study.
Measuring success after launch
If the redesign launched, I'd partner with Product and Analytics to measure:
Self-service success: More patients resolving order issues independently.
Support demand: Fewer avoidable calls about prescription status.
Required actions: More patients completing necessary steps on time.
I'd also monitor patient satisfaction to understand the broader experience.
Team feedback
“Time and time again I’m impressed with the care, thoroughness and great quality of Nancy’s work—both how she works with people and the designs that she produces. She’s able to make good progress while collaborating with fellow designers and SCRUM team members. I note particularly how she gracefully handles many unexpected challenges that come up as details need to be worked out with development teams. When Nancy’s working on something, I’m happy to know I don’t need to worry about it, and I can look forward to excellent work and positive collaboration.” — UX Design Senior Manager