Case studies
Chaos → clarity.
Five products you already know. The thing they share: the breakthrough wasn't the feature set — it was the decision they dared to make.
Linear
Before
- The category was owned by Jira — bloated, opinion-less, configurable to death
- Every team drowning in fields, views, and custom workflows
- Speed was treated as a performance problem, not a design problem
After
- Decided: speed IS the product. Everything else bends to it
- Removed configurability that didn't serve the 80% case
- Keyboard-first, opinionated defaults, ruthless visual restraint
“Linear didn't win by adding features. It won by deciding which ones to refuse.”
Figma
Before
- Sketch on Mac, Photoshop for pixels, Invision for hand-off — designers stitched together five tools
- Everyone agreed design→dev was broken. No one had a unified stance
- The obvious play was a better Sketch. Faster. Prettier.
After
- Decided: move it all to the browser. Multiplayer by default, not as a feature
- Chose WebGL and the pain of inventing its own rendering layer
- Made the file the source of truth — for designers, devs, PMs, everyone
“The best product decision in design software this decade wasn't a feature. It was a posture.”
Notion
Before
- Every team owned Docs, Drive, Confluence, Trello, Evernote, and a wiki graveyard
- Founders pitched "one tool to replace them all" — most ended up being none
- Flexibility usually means shallow. Opinion usually means rigid. Can't have both.
After
- Decided blocks were the atom — and everything else (pages, databases, dashboards) was arrangement
- Turned the onboarding page itself into a marketing surface
- Chose community templates over in-house tutorials — users taught users
“Notion sold the same feature set as its competitors. What it actually sold was the decision to stop using them all.”
Duolingo
Before
- "Serious edtech" flopped again and again: Rosetta Stone, Pimsleur, textbooks on iPad
- Retention was catastrophic — adults dropped off in weeks
- The category equated rigour with discomfort
After
- Decided gamification wasn't a layer — it was the product
- Owl-led push notifications, streak anxiety, leagues, hearts — choices that sounded unserious, worked brutally well
- Made the product slightly embarrassing to talk about. Billions of daily sessions later, nobody's embarrassed.
“Duolingo's engineers built the learning engine. Duolingo's designers decided it should feel like a game, not a class.”
Arc Browser (The Browser Company)
Before
- Every "Chrome alternative" in a decade had died — Opera, Brave, Vivaldi, etc.
- Users didn't want a browser. They wanted tabs to stop ruining their day
- Incremental UI tweaks would not unseat Chrome
After
- Decided to rebuild the browser around Spaces and Members — not tabs and bookmarks
- Moved the address bar to the sidebar. Broke a 20-year convention, on purpose
- Every aesthetic choice was positioned: "the browser for creative people who refuse Chrome's chrome"
“Arc wasn't a better browser. It was a better decision about what a browser is for.”