Baskin-Robbins Taiwan's first membership app
Built the loyalty surface Taiwan didn't have—points, rewards, and a store-ready redemption path—by scoping a credible MVP against a Japan reference model and real CRM constraints.
- Role
- Lead UX/UI Designer
- Timeline
- Dec 2024 – Mar 2025 (design phase)
- Team
- PM · 4 engineers · client + JP HQ
- Scope
- End-to-end app UX/UI · UI kit · handoff
Japan had the program; Taiwan had the gap
In Japan, BR31 already runs a mature digital membership: rewards, exclusives, and repeat engagement. In Taiwan, the brand was still missing a first-party loyalty surface that could compete with chains already owning mobile rewards and personalized offers.
The product mandate wasn't “make pretty screens”—it was to convert occasional buyers into repeat members through a coherent earn-and-redeem loop, while staying aligned with headquarters and shippable on a fixed calendar.
Copying Japan wouldn't work—and neither would scope creep
- Market rulesJapan uses points to drive membership tiers; Taiwan's commercial and CRM setup emphasized spend-based progression. Treating “parity” as identical UI would have shipped the wrong mental model.
- Delivery ceilingThe client wanted Japan-like breadth; engineering needed minimal backend churn and predictable surfaces(including webviews and external URLs where native wasn't justified for MVP).
- Omnichannel truthRewards only matter if redemption works in store. The design had to anticipate member identity at POS—QR or equivalent handoff—so digital and physical stayed one account, not two stories.
Own the path from ambiguity to build-ready files
I led UX and UI with the PM and four engineers. That meant requirements synthesis, feasibility negotiation with backend reality, and continuous alignment with Taiwan stakeholders and Japan headquarters—not a handoff after a big reveal.
- Functional map merging client goals with CRM capabilities and API boundaries—so scope was explicit before pixels.
- Flows + page logic after wireframe approval: branching, edge cases, and what ships as webview vs native.
- Hi-fi UI + UI kit (components, color, type, icons) for consistent iOS/Android implementation.
- Developer handoff: annotations, slicing notes, scroll vs fixed regions, and Figma links so engineering didn't interpret layout by guesswork.
What we optimized for under pressure
MVP scope as a contract—not a wish list
The functional map forced decisions: what lands in phase one, what waits for phase two, and which experiences stay thin wrappers (webview/URL) to protect the launch date. That reduced rework when product, client, and engineering meant different things by “launch.”
Localize the membership mechanic, not just the copy
We reassessed points vs spend when Taiwan had no initial one-to-one redemption plan like Japan. I reframed UX around the mechanism the business could operationalize—so screens matched backend rules users would actually hit.
Visual direction as a business choice
I presented three distinct directions (energetic pink-led, minimal/fresh, gradient/premium). The selected route balanced brand energy with scanabilityfor dense menus and promo surfaces—because loyalty apps fail when users can't complete tasks in line or at the counter.
A member hub, not a brochure
The app centers on earn → track → redeem: points, vouchers, and transaction visibility, with quick actions surfaced on home so habitual use doesn't depend on hunting through catalogs.
Flavor and merchandising content stay visually rich but structured: grids, promos, and navigation patterns that repeat so learnability compounds across sessions.
Cross-platform execution was specified, not implied: PingFang TC on iOS, Roboto on Android; BR pink and blue anchored to components; modular buttons, cards, nav, and inputs so future features don't fork the system.
What “done” meant in this engagement
This project measured success in delivery quality under constraint: a client-approved MVP packaged for engineering, aligned with HQ, and honest about what required backend work vs front-end-only wins.
- ~4 monthsFrom kickoff to dev-ready mockups and UI kit—while continuously trimming and reprioritizing for the MVP launch window.
- Fewer build-time surprisesFlow diagrams with system logic + annotations reduced ambiguous tickets; webview vs native was decided early to match capacity.
- Phase-one MVPScoped for launch on an agreed timeline (including post-design build), with larger parity moves queued for phase two instead of collapsing the schedule.
- Skill carryoverClearer judgment on when UX needs backend support vs when front-end patterns suffice—critical for enterprise-style delivery.
Quantitative engagement metrics post-launch can be added once the client releases them; the portfolio story here is judgment, scope control, and ship-ready craft.