FLICKR · PLATFORM REBUILD
Flickr App Reimagined
Flickr rebuilt its iOS and Android apps in React to modernize the experience and let any engineer build for mobile. I led the research and designed core features, from onboarding to uploading, on a new design system.
ROLE
Lead product designer
TEAM
2 designers (with an agency), 2 engineers, 1 PM
PLATFORMS
iOS and Android, rebuilt in React
OUTCOME
Limited beta, then rolled back. Lessons below.

01 — WHY REBUILD
Increase retention on a legacy app
Years of tech debt left Flickr’s apps slow and dated, and missing features members relied on from desktop. Rebuilding in React would modernize the app and open mobile work to every engineer at the company, not just a few native specialists.
65%
of new users abandoned the app within 3 days of signing up
40%
of surveyed users felt disengaged when landing on their home feed
82%
were frustrated that desktop features were missing from the app
02 — RESEARCH
Grounding the rebuild in what members needed
I ran the research that shaped what belonged in the new app: an iOS survey, an outcome-driven (ODI) workshop, and concept tests for new ideas like photo spots, private group sharing and photo feedback.
TOP PAIN POINTS FROM THE iOS SURVEY
Slow loading and errors
Dated, high-friction UI
Organizing photos is hard
Hard to find inspiration
Members unaware of features
WHAT IT CHANGED
From findings to features
Research moved groups into the main feed for better discovery and made uploading a priority. Ideas like changing a photo’s license and sorting by date taken or uploaded went into a ranked backlog for scope reasons.
MY IDEA
A benchmark before launch
I ran an NPS survey on the existing app to set a baseline. The score was low, and the plan was to re-run it on the new app to measure whether the rebuild worked.
AI IN MY PROCESS
Faster synthesis, global copy
I used AI tools to synthesize open-ended NPS responses into themes, and prototyped in Lovable to pressure-test flows and make sure copy worked for an international audience.
03 — FROM VISION TO PLAN
Turning a brainstorm into something engineering could build
After a three-day leadership brainstorm, features landed in Jira as empty tickets. Engineering leads struggled to scope work in new code, in a separate app, alongside native engineers who knew the old app best.
01
Pushing for what the team needed
Our team pushed back on the timeline and asked for more time, resourcing and data. That brought in an agency to build a dedicated design system.
02
A system connected to code
I contributed to and audited the new system. Because its Figma components mapped directly to React components, I could design iOS and Android screens quickly, and engineers could build them with fewer handoff gaps.
04 — WHAT I DESIGNED
Home, search and uploading: the core of a photo app
I owned three of the app’s most-used features end to end. I also designed onboarding through empty states, and brought settings to parity with desktop.
01 · HOME
A home feed you can steer

Following feed

Top 500, swipeable

Pro banner in the feed
02 · SEARCH
Search that gets you there fast

Recent searches

Priority search

Related searches

Group results
03 · UPLOAD
Uploading, end to end

Pick photos

Edit in app

Add details

Upload in progress
05 — OUTCOME
What happened at launch
THE HONEST VERSION
Under pressure to ship, the new app launched to a small percentage of users but unfortunately was incomplete. The response was harsh, and leadership rolled back to the existing app.
The design system, research and feature work still left Flickr with a clearer picture of what members need on mobile, and a React foundation connected to design.
06 — WHAT I LEARNED
Quality and parity are the launch bar
Parity before novelty
Long-time members judge a new app by what it’s missing. Matching the essentials matters more than adding something new.
Measure before you ship
My NPS baseline was the right instinct. I’d push even harder for instrumentation and quality gates before any beta.
Ship less, finish more
A smaller, polished first release would have earned the trust to keep building, along with more structured feature announcements and feedback loops.
NEXT PROJECT