STHLM · 59.33°N · 18.06°E
How a single in-app feedback prompt rescued the McDonald's app rating in Saudi Arabia
Mission
McDonald's needed to repair its rating overall in the App Store — a 2.9★ average that was actively suppressing acquisition and threatening the brand commercial position.
Challenge
Build, ship and validate a feedback prompt that filters happy users into the App Store and unhappy users into a private support channel — without redesigning the app, without an A/B test, and without disrupting the loyalty flow. Five weeks from kickoff to live in production.
Client
McDonald's
KSA pilot · 67+ market platform
My role
Product Design Lead
Services
Scope note: This case study focuses specifically on the accessibility (a11y) initiatives delivered for this client. In my role as Product Design Lead, my responsibilities extended well beyond accessibility and included product strategy, UX and UI design, and the evolution of the design system. Those parallel streams of work are intentionally not detailed here and will be shared in upcoming case studies
Digital Product Design
Product Strategy
Project Mgmt
Digital Accessibility
Impact
An actionable card in Home Screen shipped in five weeks, that can turn the App Store rating from a brand liability into a marketing asset across 67+ markets.
App Store rating in KSA tripled in 23 days. The same pattern we expect to roll across 67+ McDonald's markets in a new release window.
STORE RATING
1.55 4.31
Of users who opened the prompt, 79% selected 'I'm lovin' it' — a real, measurable signal of in-product satisfaction the team had no way to capture before.
USER SENTIMENT
79% Lovin' it
Discovery to live in production in five weeks, owning the spec, the design, the Jira ticket, and the dev co-working sessions through the build.
SHIP CADENCE
5 weeks to ship
The action card being an existing documented component reusable for future surveys, NPS, and rating prompts across markets.
SHIP CADENCE
1 Reusable Pattern
"In my role, I constantly look for scalable solutions that can move the needle across markets—not just locally, but system-wide. What stood out in this initiative wasn’t only the lift in ratings in a single market, but how intelligently the ratings flow and UX were redesigned to influence customer behavior without disrupting the core journey.
The solution proposed by the Senior Product Designer helped turn a fragmented experience into a clear, data-driven ratings flow that customers actually engage with. The results in KSA show that when the experience is structured well, improvements in ratings and feedback quality follow quickly.
What’s most exciting is the repeatability. This isn’t a one-off win—it’s a blueprint. The same approach to ratings, feedback loops, and action-driven UX can be adapted across other markets and even extended into different verticals within McDonald’s. It’s a strong example of how thoughtful product design can unlock measurable business value at scale."
%2012_10_15.png)
LET'S START
The problem behind the rating.
KSA was sitting at a 1.55★ App Store average. The product wasn't broken — analytics showed strong daily engagement, healthy retention, and a steady stream of orders. The reviews were outliers from a vocal minority, and there was no in-app mechanism to capture the silent majority who were satisfied.
The market needed a way to redirect frustration into a private support channel and surface satisfaction into the public rating — without redesigning the home screen, without launching a survey tool, and without breaking the loyalty flow that drove most of their session value.
DISCOVERY
Reading the dashboard, not the reviews.
Pulled GMA Lite analytics across all McDonald's markets to compare KSA against control markets at similar maturity. The pattern was clear: high engagement, healthy retention, low review volume — and a tiny pool of vocal 1★ reviewers dragging the average.
In parallel, I assessed landscape-specific constraints such as hand positioning, reach zones, and changes in visual hierarchy, clarifying how orientation impacts usability and accessibility at scale.
Cross-referenced the action-card surface logic in the existing app to find a non-disruptive insertion point: the home screen action stack. Already battle-tested for promotions and loyalty nudges; perfect for a feedback prompt.
Reviewed Apple's App Store Review API documentation to confirm we could trigger the native rating prompt only after a positive in-app signal — protecting users from being asked when frustrated.
STRATEGY & DESIGN DIRECTION
Two journeys, one card.
Designed a single in-app action card with a fork. One question — “How are you liking the app?” — branches into two paths: a public review path for advocates, and a private feedback path for critics. The negative path never reaches the App Store.
If user taps 'Lovin' it' → 'Rate us': native review flow opens and the prompt never shows again.
If user taps 'Needs work' → 'Leave feedback': routed to support, never shown a public rating ask.
KEY FOCUS AREAS
Where the work actually went.
The visible deliverable is small — one card, two buttons, two follow-up screens. The hidden deliverable is the dismiss-window logic that decides when the prompt comes back so it never feels nagging.
Mapped a five-state matrix covering every outcome: closed without interacting (return in 3 months), tapped 'Needs work' but no feedback (return in 3 months), tapped 'Needs work' and submitted (return in 4 months), tapped 'Lovin' it' and chose 'Maybe later' (return in 3 months), and tapped 'Lovin' it' and rated (never show again).
Aligned the matrix with engineering before any UI exploration. The logic was the contract; the screens were the surface.
Default: 3 months. Negative + feedback: 4 months. Positive + rated: never again.
VALIDATION & ITERATION
Owned the spec. Co-authored the build.
Wrote the Jira ticket myself — GMAL-25904 — to control the brief, not just the design. The ticket carried the full context: session-start trigger logic, two-flow branching, dismiss-window matrix, native review prompt fallback, and the constraint that markets that have loyalty enabled use prior session history while others use a counter.
Paired with engineering through the build in three co-working sessions, resolving backend tag limitations in real time. Tags couldn't track when the user interacted with the card, so we moved the timing logic into the app itself, kept GA4 firing for analytics, and centralised the source of truth.
The card shipped on 21 August 2025. By 13 September the App Store rating in KSA had climbed from 1.55★ to 4.31★. No A/B test, no soft launch, no rollback.
VALUE DELIVERED
Value Delivered
A 23-day swing from brand liability to brand asset, validated in the live App Store. The pattern is now ready to scale across the McDonald's market portfolio — Saudi proved it works at production scale, with no special-casing required.
