EcoTrack

Overview

People who genuinely care about climate change often feel paralysed by it. Not indifferent. I spent four weeks researching why motivated users abandon eco apps, then designed the experience I wish already existed.

End-to-end solo: everything from screener survey to two rounds of usability testing.

Role

UX Designer (solo)

Timeline

4 weeks

Platform

iOS

Type

Personal project

Research

5 interviews · 2 test rounds

Why motivated users still quit

Motivated users abandon tracking not because they don’t care. Existing apps optimise for data completeness over daily usability, making the act of tracking feel worse than not tracking at all.

• iOS concept only. No backend. Carbon estimates are averaged. Approximate over complex.
• Scope: daily habits. Offsets, social, leaderboards cut. The core loop had to work first.
• 4-week sprint. A self-imposed deadline forced real prioritisation.
• No longitudinal data. Insights from 45-minute sessions, not observed behaviour over time.

“I opened Joro once, saw my score, felt terrible, and deleted it. I wasn’t looking for a report card.”

Five phases

Double-diamond. The Define phase upended my assumptions entirely.

• Empathize. 5 semi-structured interviews and a 12-person screener survey. I recruited people who identified as eco-conscious but inconsistent, not existing eco-app users.
• Define. Synthesised 30+ raw findings into one framing: users don’t abandon eco apps because they don’t care. They quit because tracking makes them feel worse, not better.
• Ideate. Listed 23 possible features, then cut them with one filter: can a first-time user do this on day one without help?
• Prototype. Mid-fi first to validate IA and flow, then hi-fi to test visual decisions, with usability tests between both rounds, not just at the end.
• Test. Two rounds with 5 participants each. Round 2 tested the specific changes identified in Round 1.

What I expected to find, and what I found

I expected to fix a UX problem. Research surfaced an emotional one: 68% said existing apps made them feel judged, not helped. The brief wasn’t “make tracking easier” - it was “make it feel safe to be imperfect.”

• 3 of 5 had quit an eco app within the first 10 days. So first-time logging became the primary design problem. Not the dashboard.
• No one wanted to manually log food in detail. So quick-select cards replace free-form input. Speed over completeness.
• Every extra navigation step caused 1-2 testers to abandon the task. So logging happens without leaving the home screen. The floating add button exists because of this single observation.

Three segments, two personas

Priya, 22, student

“I care about the planet, but I don’t know where to begin.”

Arjun, 28, PM

“I’ve built good habits. I want to know if they really matter.”

• Eco-curious beginner. Cares but overwhelmed. Needs wins, not warnings.
• Active eco-enthusiast. Built habits already. Wants proof they matter.
• Lapsed eco-app user. Burned by guilt-driven data. Trust is the entry point.

What each competitor taught me

I used each app for 2 to 3 days and paid attention to the moment I stopped wanting to open it.

• Joro. Automatic tracking feels magical but creates trust problems when users can’t see why a purchase triggered a number. I kept logging manual but made it fast.
• Capture. Every screen has 6+ competing data points. It hardened my one-primary-number-per-screen rule: completeness is the enemy of daily usability.
• Aerial. Proved a beautiful narrow-scope product beats a complete ugly one. Scope constraint is a feature, not a limitation.
• UNEP Carbon. Education and action live in separate silos. Learning and logging need to share the same screen - a standalone learning tab is a tab users won’t open.

No single app was bad. Each excelled at one thing. EcoTrack’s brief: make daily logging so frictionless that users don’t need anything else to build a habit.

Every decision had an alternative I chose not to use

I tested three green values. Lime-green was too optimistic, celebrating before the user did anything. Forest green felt preachy. #2D7D43 sits between: calm, credible, grown-up.

For type, DM Sans was too friendly - the wrong register for a data tool - and Nunito’s rounded terminals looked playful at small sizes. Inter won: trustworthy, data-appropriate, excellent at every weight.

IA first, screens second

One filter: can a first-time user complete this on day one? Social sharing, streak mechanics, carbon offsets and lifecycle analysis all failed it. What remained: home, activity log, impact breakdown, learning, progress.

Onboarding, one commitment at a time

V1 asked for too much before showing any value - two testers dropped off before setup was complete. V2 earns trust first, then asks.

Core screens

Each screen answers one question without requiring the user to navigate away.

Iteration 1 · The add button

Hypothesis: a centred tab bar add button mirrors Instagram and Spotify. Reality: 4 of 5 testers couldn’t find it within 90 seconds. One said: “I kept tapping the logo.”

The update moved it onto the screen as a floating action button. In Round 2, all 5 testers found it immediately.

Iteration 2 · From technical to human

Hypothesis: exact CO₂ values with breakdowns would give users the understanding they wanted. They could read the numbers. They just felt worse after - three described them as “meaningless” or “like seeing calories on a menu, just guilt.” The problem wasn’t the design. It was the frame.

V2 uses comparative framing with plain-language equivalents. Same data, different register. After: users said “motivating,” not “depressing.”

Outcome

After Round 1, the word most testers used was “complicated.” After Round 2 it was “simple.” Same data - prioritised, not reduced.

• Log a morning commute: 3/5 → 5/5. The floating button replaced the buried tab button.
• Find an actionable food tip: 4/5 → 5/5. Comparative framing replaced raw CO₂ numbers.
• Add a new weekly goal: 2/5 → 4/5. Goals separated from the Profile tab.

Honest caveat: 5 participants per round validates direction, not production readiness. The real test - 30-day retention - requires a live product.

What I learned

• The problem I thought I was solving wasn’t the real problem. I started out thinking the gap was bad UX. Research showed the real gap was emotional: users felt judged by their own data.
• What I cut mattered as much as what I built. I started with 23 features. The quick-log flow only works because I didn’t also try to handle carbon offsets on the same screen.
• Research findings need a decision attached. The findings that changed my design were written as decision sentences: “Because users said X, I will do Y.” Next time I write those before synthesis, not after.
• What I’d do differently. Design for the unhappy path. Every screen shows a successful, populated state - not day one with no data, a bad week, or a user who wants to stop tracking.