Reducing uncertainty in smart home onboarding

Reframed onboarding from UI simplification to confidence-building, increasing activation by 42%.

Role: Lead Product Designer Β· Duration: 5 months Β· Team: PM, Engineering, Research

Executive Summary

Redesigned a failing onboarding experience into a predictable, confidence-driven system that helped users successfully set up their devices.

Instead of simplifying the UI, we focused on reducing uncertainty during setup β€” the primary cause of drop-off.

Impact:

  • +42% onboarding completion

  • Pairing success increased from 21% to 74%

  • 7-day retention doubled

  • Support tickets reduced by 37%

Why this mattered

  • Smart home adoption depended on successful activation.

  • Most users abandoned setup before connecting their first device.

  • Hardware sales were strong, but activation rates remained critically low.

  • Support costs increased as onboarding confusion grew.

Business risk

  • low activation

  • poor retention

  • increased support dependency

  • reduced ecosystem adoption

Product challenge

The onboarding flow needed to reduce uncertainty for first-time users without increasing complexity.

Context & Problem

Homely was bundled with smart home hardware, but the mobile app struggled to convert buyers into active users.

Despite strong hardware sales, users failed during setup and abandoned the product early.

Key issues:

  • high onboarding drop-off

  • unreliable device pairing

  • increasing support dependency

  • weak early retention

This was not a UI problem β€” it was a behavioral bottleneck blocking activation.

Discover & Insights

We combined analytics, interviews, and usability testing to understand where and why users failed.

Initial hypothesis

We assumed the onboarding was too complex and needed simplification.

What we discovered

  • users didn’t understand device hierarchy

  • setup language created fear of mistakes

  • long pairing moments lacked feedback

  • unclear system states caused hesitation

Key insight

The problem wasn’t too many steps β€” it was a lack of clarity and predictability.

Failed assumptions

Assumption 1

Reducing steps would improve onboarding completion.

What happened

Users became less confident and more likely to abandon setup.

Why it failed

Users didn’t need fewer steps. They needed reassurance and predictability.

Assumption 2

Tooltips could solve confusion without changing the flow.

What happened

Users ignored guidance under stress.

Why it failed

The onboarding architecture itself created uncertainty.

Key realization

The issue was behavioral, not instructional.

Decision

We explored multiple directions before defining the final approach.

Option 1: Simplify the flow ❌

Reduce steps and UI elements to speed up onboarding.

Result: Testing showed that removing steps increased confusion and drop-off.

Option 2: Add guidance layers ⚠️

Introduce tooltips and explanations on top of existing flow.

Result: Users ignored guidance and still felt uncertain.

Final direction: Build confidence through structure βœ…

Instead of reducing steps, we:

  • broke the process into predictable micro-steps

  • made system states visible at all times

  • introduced recovery paths for failures

This shifted onboarding from task completion β†’ confidence building

Solution: rethinking onboarding

We redesigned onboarding as a structured, guided system focused on clarity and control.

Key principles

  • predictable step-by-step progression

  • continuous system feedback

  • clear success and failure states

  • reversible actions and recovery paths

Key changes

  • introduced guided micro-steps instead of long flows

  • added explicit system states (searching, connecting, success, failure)

  • reduced ambiguity in device setup language

  • ensured every action had visible feedback

The goal was not speed β€” it was confidence.

Solution Architecture

The system was designed to reduce cognitive load and guide users through setup with minimal ambiguity.

Interface decisions

  • strong primary actions guiding progression

  • limited use of color to highlight key actions

  • consistent icon system for device recognition

  • predictable layout and spacing

  • clear feedback states for every interaction

The interface reinforced a single idea:

πŸ‘‰ β€œYou always know what’s happening and what to do next.”

Iteration & Testing

We validated the solution through rapid RITE testing cycles. Each iteration focused on reducing hesitation and improving clarity.

Key improvements

  • unclear transitions β†’ replaced with explicit system states

  • long waiting times β†’ enriched with feedback and time expectations

  • failure states β†’ redesigned to be recoverable instead of blocking

Testing guided every major decision and helped prioritize changes with the highest impact.

Outcome & Business Impact

Measured over 8 weeks post-launch:

  • onboarding completion increased by 42%

  • pairing success rose from 21% to 74%

  • 7-day retention doubled

  • support tickets dropped by 37%

  • activation time reduced by 53%

The onboarding became faster, more predictable, and significantly more reliable.

Reflection

This project showed that clarity drives adoption, while uncertainty blocks behavior.

Simplifying UI is not always the answer β€” in complex systems, users need confidence, not minimalism.

Design impact is measured in behavior change, not visual polish.

Previous
Previous

Eliminating approval chaos in onboarding system

Next
Next

Building a scalable UX foundation for a growing SaaS platform