Virtual Walkthrough
Visual Design
UX/UI
3D Modelling
Product Management
60% of users scan spaces for documentation, however while 3D meshes capture geometry, fine details like labels, pipes, and material finishes disappear. This feature was the bet that we could recover that detail without asking users to recapture anything.
The insight that made this feature possible was that every spatial capture already contained raw camera images and camera intrinsics. I worked with engineering to reconstruct the original scan path and anchor raw captures precisely to the 3D mesh.
Backwards compatible, this feature enables users to "walk" through their capture in sequence, jumping between the model and original photography for both spatial context and fine details.
Initial activation was lower than expected, not because of usability but because the entry point was buried. I ran rapid experiments lift daily activations, the most successful being a tooltip surfaced at first open that lifted daily activations 70%.
Simultaneously, I ran an A/B test on incumbent generic vs contextual paywalls. Tool specific messaging converted 12% more users - both tooltip and paywall patterns are now used across all new feature launches.
