Earth Hero.
A redesign of Earth Hero's homepage and actions experience - focused on why users were dropping off and what it would take to make them stay.
I was the design lead.
I hired and managed the design team, set the design direction, and kept everyone aligned - across research, strategy, writing, and the stakeholder. We worked in sprints, tested continuously, and made sure every decision had research behind it.
Half of users were gone within six weeks.
That was the number we kept coming back to. ~50% drop-off in the first six weeks. And when we dug into why, it came down to three things: users didn't know what to do, couldn't find their way around, and didn't feel like their actions actually mattered.
The app wasn't broken - it was just unclear. Users weren't disengaged because they didn't care about climate action. They left because the product made it too hard to care.
What we were designing toward.
- Know what actions to take without having to figure it out
- See their impact in a way that actually means something
- Move through the app without getting lost
- Cut the early drop-off rate
- Get more users engaging with recommended actions
- Give the product a clear direction to build from
We started with a card sort.
Before anyone touched a frame, the research team ran an open card sort to understand how users actually grouped and thought about the content - not how the product assumed they would.
My role in research
I worked with the researchers to decide what content to test, built the card sort testing pages, sat in on sessions to take notes, and then translated what we found into IA changes.
The work that shaped the direction.
Google Analytics - We pulled drop-off data to understand where users were leaving the app. The steepest fall-off happened within the first two weeks, which pointed us toward onboarding and the homepage as the primary problem areas.
Card Sort - Round 1 - Users grouped content in ways that didn't match the app's existing structure at all. This told us the information architecture needed a full rethink, not just surface-level fixes.
Categorization Diagram - After the first card sort, we mapped out how users were naturally organizing content. This became the foundation for the new IA and navigation structure.
Card Sort - Round 2 - We ran a second sort to validate the new categories. Results were much more consistent this time, which gave us confidence we were moving in the right direction before touching any designs.
Testing told us where we got it wrong.
Usability testing pushed us to rethink several things we thought were working. Here's what changed and why:
Keeping the stakeholder in the loop.
I ran weekly check-ins with the stakeholder to share where we were and talk through tradeoffs. When their instincts didn't line up with what research was telling us, I pushed back - with evidence. That's part of the job.
High fidelity, ready to build.
Once the direction was validated, I led the final hi-fi designs and put together handoff documentation so the work was actually usable by a dev team - not just pretty files sitting in Figma.
What I'd do differently.
Leading a cross-functional team taught me a lot about how to keep people aligned without slowing things down. I got better at advocating for research when it conflicted with gut instinct - mine included.
If I did it again, I'd define success metrics earlier and pull in engineering before handoff. Getting developers in the room sooner would've made the final designs sharper and more realistic to build.