Lead Product Designer/2021/Web3 / Real Estate
Scaling a Web3 real-estate platform from 0 to 70K+ users — acquired by ARK7
I led the design of Robinland — a Web3 real-estate investment platform that let first-time investors buy tokenized fractional shares of real properties. The job was unusual: designing for users who’d never invested before, working with mechanics most of them had never heard of, inside a market most of them didn’t trust. We had one year to scale from zero to something with enough product-market signal to keep the company alive. That work eventually became part of the foundation that led to ARK7’s acquisition of Robinland.
First-time investors and unfamiliar Web3 mechanics, on a one-year clock
Robinland was selling something that didn’t exist in the mainstream consumer vocabulary yet: fractional real-estate ownership represented as blockchain tokens, with yields paid out on a schedule and properties browsable on an interactive map. Every piece of that proposition required user education. And every piece of it was happening in the moment a user was being asked to commit real dollars to it.
The constraint that defined the work was the timeline. We had a year to scale the platform from zero to something with enough signal to justify the company’s next round. Research couldn’t sit in a separate phase, and neither could polish. Both had to live inside the design and engineering rhythm we were already running.
Confidence broke down where the language did
I ran lightweight studies in Maze and UserTesting in parallel with shipping — not as a research phase but as a continuous signal pulled into each iteration. Behavioral data from early users came in through analytics and stakeholder review sessions. The pattern that kept surfacing wasn’t “users don’t understand Web3,” which everyone expected. It was “users stop where the interface stops explaining itself.” Confidence broke down at the seams: when a user moved from browsing a property to clicking Invest, when an interactive map gave them no callout for what was selectable, when filter labels assumed knowledge the user had no way of having.
That insight changed the tradeoff I was holding. Hiding Web3 mechanics — wallets, tokens, yields — would have made the product feel less intimidating but also less honest, and the long-term users we wanted would eventually need to understand those concepts to take the platform seriously. The alternative was to design contextual cues around the language: labels, callouts, inline guidance that let users learn while they invested rather than before. We went with the second path.
Each unfamiliar concept earned the right to be on the screen
The redesign reorganized the discovery flow around the question a first-time investor was actually asking: where am I looking, and what can I do here? The map got interactive callouts that clarified which areas were selectable and how zoom controls behaved, instead of leaving exploration to trial and error. Filters got guided labels that explained their function on first glance, so a new user could narrow results without learning a new vocabulary. The Invest action moved into the primary line of the journey — visible, persistent, and weighted against the secondary actions that had previously competed with it.
Each property listing earned contextual descriptions that gave the user enough to evaluate the opportunity without leaving the page — what they were buying, what the projected returns implied, what the underlying property was. Search and navigation got discovery cues so first-time users found features they otherwise wouldn’t have stumbled into. None of these are exotic patterns. They’re the basic moves experienced product teams take for granted on consumer-facing flows. They worked here because we applied them rigorously to a product where users had no priors to fall back on.
Trust scaled because the language did
The platform onboarded tens of thousands of users in its first phase and crossed 70,000 over the year that followed. Robinland was acquired by ARK7 not long after — the kind of outcome where the design contribution is one of several inputs, not the headline. The work I’d point to as the proof isn’t the user count. It’s that the patterns we shipped in the first quarter still held up at scale, because they weren’t shortcuts to mask the unfamiliarity of Web3. They were a vocabulary built for users who didn’t yet have one.
What I notice looking back is how little of the eventual product was new. The structural moves from the first quarter — the contextual labels, the inline guidance, the way Invest sat in the line of the journey — were still doing most of the work at the time of the acquisition. Whatever I’m designing now, I’m trying to ship work that holds up that way.