A marketplace that actually understands creative work.
Most freelancer platforms are built around transactional logic: post a job, receive bids, pick the cheapest one. That works fine for commodity work. It fails completely for creative, strategic, or specialist work — where the relationship, the working style, and the cultural fit matter as much as the portfolio.
The client — a creative director setting up the platform — came with a clear instinct: this should feel different from Upwork or Fiverr. Less like a job board, more like a professional community. The strategic challenge was translating that instinct into a product architecture that could actually sustain a marketplace at scale.
Understanding both sides of the market.
I started by running discovery interviews on both sides of the marketplace — with freelancers who had experience on other platforms, and with clients who hired creative talent regularly. The goal was to understand not just what they did on current platforms, but what frustrated them, what they wished existed, and what would make them choose something new.
The structural choices that define the marketplace.
Marketplace strategy is mostly about choosing what you won't do. The platform needed a clear position on several fundamental decisions before any design work could be meaningful.
Profile Design
Rather than a standard portfolio grid, profiles are structured around working style, availability, and collaboration preferences — not just output. The bet is that clients who know how they like to work will make better hiring decisions when they can see whether a freelancer matches their style.
Discovery Model
Passive discovery (browse and apply) is supplemented by active matching — the platform surfaces relevant projects to freelancers based on their profile, not just their skills. And relevant talent to clients, not just their brief keywords. This requires a richer data model than most platforms attempt.
From strategy into structure.
With the strategic decisions made, I moved into design exploration — working through the information architecture, the core user flows, and the profile and project page templates. The goal at this stage was to get to something testable with real users as quickly as possible, not to produce a polished design.
The first round of wireframes revealed a tension in the profile design: we wanted profiles to feel rich and human, but too much required information would create friction at signup that would kill supply-side growth before it started. The solution was a progressive profile model — a minimal profile gets you in, but the platform incentivises completion over time.
Where we are now.
The platform is in the late strategy and early design phase. The product roadmap is defined, the core user flows are validated at wireframe level, and the first round of user testing is being planned. The next milestone is a mid-fidelity prototype of the core freelancer and client flows ready for testing with real users.