Worka

Three products under one brand, taken from a proposition to an MVP: the Worka workspace marketplace, Worka Team (the enterprise admin product) and the Partner Portal for workspace partners.

Role

Sole designer on the Worka marketplace for roughly the first month, defining scope with the CTO, the head of engineering and the head of programme.

Team

Inception group of four: the CTO, the head of engineering, the head of programme and me, joined soon after by a front-end engineering partner. Design scaled from about a month in, with a lead designer for each product in later phases.

What I owned

The marketplace concept and core experience, the delivery plan, and the atomic structure and first components of the core design system, before scaling the team.

What shipped

Shipped to clients in closed beta.

Three Worka marketplace phone screens: the home screen with search and nearby desks, meeting-room results including a Company Space listing, and a desk booking on a floor plan at a price of £0 for company space.

Search, results and booking for the person searching. A company’s own space ranks on relevance like everything else; budget and permission show as price, not as position.

See all screens →

Context

The Instant Group is a corporate real estate business that helps organisations and individuals find workspace, in a global market valued at $3.1tn in 2022 (Allied Market Research, 2023). Worka was its attempt to open that market up.

The idea existed when I arrived. The product did not. There was a basic brand from an external agency and no definition of what it would be, for whom, or how it would work.

Demand runs from a sole trader booking a desk for the day to a business signing a five- to ten-year lease. Serving both lets a customer’s workspace needs grow inside the product.

Director, User Experience

The Instant Group

May 2023 – July 2025

Constraints

Eighteen months to an MVP, on a legacy technology stack that would slow engineering if design depended on it.

Two transaction models in one interface. Short stays settle in minutes through Stripe. Nobody commits to a five- to ten-year lease off photographs, so those listings exist to book a viewing and open a sales conversation.

Visibility is scoped per person. An enterprise can list its own space into the marketplace, walled so only its employees see it, and those employees are further bound by permission levels and budget. Two people searching the same city on the same day see different inventory.

Decision

Rank results on what fits the person searching, not on what the enterprise would prefer they book.

An enterprise has an obvious interest in filling its own space first and spending less outside it. That interest sits at the boundaries: budget limits and permissions define what is bookable, and within them relevance decides what surfaces. Discovery with enterprise customers established this was enough for MVP.

Commercial weighting was deferred. Before a marketplace has critical mass, tilting results toward the payer risks suppressing the thing the product most needs, which is people using it.

Trade

Deferring the ranking lever gave up a commercial mechanism, and it put more weight on the budget and permission controls. They became the only place an enterprise’s financial interest was expressed. So they had to be precise enough to satisfy a finance owner, and quiet enough that an employee inside their allowance never felt managed.

It also stores up a harder problem. The ranking question returns once the marketplace has scale, and by then people will expect results to be ordered for them. Adding commercial weighting to a product users already trust costs more than building it in from the start. I took that cost to get adoption.

Process

Ran global discovery research across the English-speaking markets Worka would launch in first, to establish demand patterns and workspace behaviours. It narrowed phase 1 of the launch to the US, Canada and the UK. Ran co-creation workshops with the C-suite to define the marketplace proposition.

Prototyped and validated concepts in v0, Lovable, Figma Make and Claude Code, shortening the path from an idea to something testable.

Designed the three products together: the marketplace for discovery and booking, Worka Team, where organisations set budgets and permissions and manage their own space, and the Partner Portal, where partners set pricing and availability, upload photography and connect their existing systems to keep inventory in sync.

Architected three design systems to match, decoupled from the legacy codebase: a core design system for the marketplace, with Worka Team and the Partner Portal branching from it. That was a trade agreed directly with engineering leads: design took on the complexity of maintaining a branched architecture, and engineering got its velocity back.