Built the demo for Wix's first DevCon, then keynoted it
77% of the conference built an avatar with it
Wix's first developer conference needed a demo that showed off every new platform capability at once. It had to exist before some of those capabilities did. I built an avatar generator that ran the whole way through the platform, then presented it live as the keynote.
Also in this role
- Asked for an onboarding deck for new hires, argued for a site instead and built it. Time on the homepage doubled after feedback-driven edits
- Ran a developer-experience study whose findings shipped as onboarding UI changes, refreshed documentation and new product features
Building on features that didn’t exist yet
The demo had to cover every new release going out at the conference (Blocks, the ecommerce SPIs, the Bookings API) on a two-month timeline, and several of those weren’t finished when I started. Some I was building against before they were stable enough to build against.
Add a large company’s approval process on top and the constraint wasn’t design, it was sequencing. So I ran it as a scoped agile process with the boundaries agreed up front and regular reviews, because the thing most likely to kill it was finding out in week seven that something had changed in week three.
The demo itself
You create an avatar in a custom component, and it flows through the rest of the platform from there, into the ecommerce catalog as a product, through the new integrations, and out the other side. One continuous thing rather than a tour of disconnected features.
That was the design decision that mattered. A demo that visits each capability in turn shows you a list. A demo where your own avatar travels through all of them shows you a platform.
Building it found the bugs
Being first to implement the new releases meant I hit their edges before any customer did. The demo doubled as the platform’s first real integration test, and the bugs it turned up got fixed before launch rather than after.
The keynote
I presented it live to 300 developers, a community that had started the year at ten. 77% of attendees went on to create an avatar, which is the number I care about: not that people watched the demo, but that they opened the tool afterwards and used it.
What I’d do differently
I built the demo and the talk as one artifact, which made both fragile. A change to an unfinished API was also a change to the script, two days out. I’d separate what I’m demonstrating from what I’m saying about it next time, so a late platform change costs me a slide instead of a rehearsal.