Matterport Cloud Sync

I identified an infrastructure gap that leadership had deprioritized for years and spent six months making the case to get it on the roadmap. Then I designed and shipped it with one engineer. 87% NPS increase.

Company
Matterport (3D capture platform, acquired by CoStar for $1.6B)

Role
Product Designer (only designer on project)

Team
1 engineer, minimal PM support

Impact
87% NPS increase
368,948 restores in first 90 days
Shipped across iPhone, iPad, and Android

What I did

Matterport's mobile app was stuck in 1995. Data only lived on the device it was created on, which meant customers were literally mailing iPads across the country to collaborate. If you scanned a building, added 1,000 tags, and needed to rescan, you'd re-enter everything manually. Mobile created and web edited, and the two never talked to each other.

I pulled the support data, built a financial model, and spent six months selling this to leadership before designing and shipping it with one engineer.

Building the
Business Case

This project didn't start with design. It started with data.

I pulled support tickets, categorized sync-related issues, and ran user interviews. Then I built a financial model tying sync problems to churn and expansion revenue, and took it to leadership framed around NPS, retention, and support costs rather than as a design pitch. It took six months of selling across the VP of Product, engineering leadership, and the executive team. The hardest part was reframing what everyone saw as technical debt into a strategic priority.

Designing for Distrust

Research revealed that users had improvised their own backup workflows over years of the system failing them, and they didn't trust new features not to break something.

So I designed around that. Confirmation friction before destructive actions, the ability to cancel mid-operation, and progress indicators showing exactly what was happening so users felt in control. I separated "Capture Data" (local) from "Spaces" (cloud) in the nav, made upload progress visible but non-blocking, and cleaned up the terminology to use less technical language. The onboarding was value-based ("your work is safe") rather than feature-based ("we added sync").

Shipping with Constraints

One engineer, a ten-year-old codebase that nobody fully understood, and three platforms to ship across. I scoped aggressively with a phased rollout: restore and basic browsing first, iPad polish later, and advanced features explicitly post-v1.

The constraints forced clarity. Every feature got filtered through one question: does this matter enough to spend our limited cycles on? If not, we cut it. Some animations got simplified, some edge cases got deferred.

Let's chat

Resumé (PDF, 107kb) ↓

Email me →

LinkedIn →