Can you evolve a product's design without breaking user trust or existing mental models?
At Meta, I led the redesign of a core messaging feature used by 500M+ users. I started by auditing our design system for patterns that were outdated but heavily relied upon. Using Figma’s component libraries, I created a token-based system that made updates atomic—changing a color or spacing globally without side effects. For innovation, I prototyped a novel gesture-based reply flow in UserTesting, recruiting 200 existing users. The key was defining a 'consistency boundary': core navigation and reading patterns stayed identical; new interactions were additive and optional. I ran A/B tests with feature flags over four weeks, measuring task success (up 12%) and support tickets (down 8%). The innovation won because it solved a real friction—slow reply retrieval—while maintaining familiar iconography and placement. We then documented the pattern in our design system as an optional variation. The lesson: protect the user’s cognitive load like a sacred contract. Innovation must feel like a natural extension, not a foreign language.
At Google, emphasize quantitative rigor—use A/B testing in a live environment to prove the innovation doesn’t regress core metrics like search success rate or navigation latency.
At Apple, focus on the human interface guidelines as a foundation—ensure the innovation feels ‘inevitable’ by deeply integrating with existing behaviors like swipe-back or 3D Touch.
📚 Recommended Resource
The 0-1 PM Interview Playbook (2026 Edition)
Product design thinking and UX interview frameworks used at Google, Apple, and Meta.
Get it on Amazon →