Polycam Scenes
Polycam was a capture tool. You'd scan something, export it, and move on, and users who wanted to combine captures or build environments would leave for Blender or competitors. We were averaging 5,000 export activations per day. Scenes was our first creative tool, and the proof that Polycam could be a platform instead of just a utility. Daily exports dropped 30% after launch because users finally had somewhere to go inside the product.
Company
Polycam (3D Scanning platform)
Role
Product Designer + Product Manager
Team
1 3D engineer, 1 Front-end engineer
Impact
1,300% increase in unique users
Highest-converting paywalled feature
30% drop in exports (users staying in-app)
Enterprise renewals and seat expansions
What I did
I owned design and product through the entire lifecycle. Two months to deadline with partial designs and a lot of unmade decisions, then six months of continuous shipping that turned a V1 into something real. When the data showed we could expand access beyond enterprise, I put together the case and rolled it out, which drove a 1,300% increase in unique users.
Shipping under deadline.
Partial designs and a hard deadline, so there wasn't room for extended exploration. I had to make calls and keep things moving.
I helped build the 3D gizmo for placing and manipulating captures (translate, rotate, scale with direct-manipulation handles) and designed the layers panel for managing multiple captures: visibility toggles, reordering, and the spatial hierarchy that let users build complex scenes without losing track of what was where.
Launching enterprise-only made sense for V1 because it meant lower server risk, higher-intent users, and a cleaner feedback signal, but I planned for expansion from day one. I ran weekly syncs across two engineering teams, kept a shared dependency tracker, and surfaced blockers early. We hit the deadline.
Continuous Improvement
V1 shipped and feedback was mixed. There were strong customer examples and genuine enthusiasm, but also clear gaps that were blocking professional adoption.
I designed and ran a survey designed to surface actionable answers rather than just satisfaction scores, then created a prioritization framework to weigh customer requests against product vision. I documented it so the team could reference it without me in the room, and then shipped continuously for six months:
Measurement tool for cross-capture comparison. Customers could now measure across different scans in a single scene, which drove a 19% increase in measures created.
Floating toolbar I designed to keep users in spatial context while accessing tools. This was a new pattern for Polycam that's since been adopted elsewhere in the app.
Precision controls: XYZ coordinates, scale locking, and toggleable snapping. Table-stakes for AEC workflows that were blocking enterprise adoption.
AI Generator integration for virtual staging. Architects and interior designers could stage spaces without leaving Scenes, which led to a 52% increase in AI-generated models.
Layer groups, swap, and undo/redo, all based on user feedback. Professional-level functionality that was simply missing.
Design Decisions
Keeping users in the scene,
not in a menu
3D editing tools typically dock controls in a sidebar or top bar, which pulls attention out of the spatial context and into a flat UI layer. That's fine for Blender users but wrong for Polycam's audience. I designed a floating toolbar that stays close to the user's focus point in the viewport so tools are accessible without leaving the scene. This was a new pattern for Polycam, and it's since been adopted in other parts of the app, which is usually the best sign that a design decision was right.
Paywalls at the point of intent, not the point of entry
The default approach to paywalling a feature is to gate it upfront: "upgrade to access Scenes." I pushed for the opposite. Let users in, let them build, and surface the paywall when they hit a real limit like layer count or export resolution. By the time they see the upgrade prompt they've already invested in a scene and understand the value firsthand. This changed the conversion model. Scenes became Polycam's highest-performing paywalled feature, converting nearly 280% better than the next highest.
Cross-capture measurement
Measuring within a single scan is straightforward, but measuring across captures (two separate scans composited into one scene) is a different problem. Different coordinate systems, different scales, different origins. I designed the interaction so it feels identical to single-capture measurement, and the complexity is invisible to the user. This was the most requested feature from AEC customers, and the 19% increase in measurements created after launch confirmed it was blocking real workflows.
Opening Access
The enterprise-only launch was deliberate, but the goal was always broader adoption.
I pulled data on how enterprise users were actually using Scenes: layer counts, feature usage, and complexity patterns. Most weren't pushing the ceiling, which meant we could open access without overwhelming infrastructure. I proposed a tiering model, backed it with the data, and designed contextual paywalls that appeared when users hit actual limits like layer count and export resolution rather than showing generic upgrade prompts.
Scenes became our highest-performing paywalled feature, nearly 280% better than the next highest. It's also our most-viewed paywall, which shows the demand even before users convert.
The Outcomes
1,300% increase in unique users after opening to Free and Pro (2,000 to 28,000). Users who created 3+ scenes converted to paid at a 33% higher rate than other users, and daily exports dropped 30%, which was the clearest signal that we'd actually changed behavior.
Enterprise customers renewed and expanded seats because of Scenes. AEC teams doing facility management went from 3 seats to 10+. What started as a creative tool became infrastructure for professional use cases we hadn't originally designed for.
I'm currently defining the Scenes 2.0 roadmap, rethinking the interaction model and expanding the target market based on what we learned from six months of shipping.