Two hundred thousand participants, and nowhere to sell
Uni.xyz had competitions, projects, journals, publications and collections. What it didn’t have was a way for its community to trade with each other, so they left to do it somewhere else.
Five segments, and a hole where the sixth should be.
Uni.xyz connects architects, designers and experts worldwide, and by 2021 it was the largest architecture competition platform there is, with over 200,000 users. Between its five segments, the community could compete, publish, critique and curate. None of them let anyone sell anything.
Design challenges from institutions around the world
Showcasing work, with feedback and visibility
Writing about process, ideas and the field
Research, reports and long-form content
Personal galleries worth coming back to
The community was leaking at the point of sale.
A designer with a template set, a guide or something physical had a platform full of exactly the right audience and no way to offer it to them. So they took it elsewhere: a store on another site, a link in a bio, a DM. Every one of those transactions was a moment a user left Uni, and the platform lost what came next.
The problem, as we framed it“How might we integrate a marketplace within Uni’s platform that lets the community discover, list and purchase products?”
The drop-off wasn’t happening because people lost interest. It was happening at the exact point where they were most engaged.
A sixth segment that reads as part of Uni.
Not a store with Uni’s logo on it. A marketplace built from Uni’s existing library, entered from places people already knew, with two flows for two very different people: one who wants to sell, and one who wants to buy.
Profile→Uploader→Preview→Publish→Browse→Checkout
Centralised listings
Products live in one findable place instead of scattered across profiles and other sites.
Discoverability
Tags and categories so a good listing reaches the person who wants it.
In-platform purchase
Payment completes where the relationship already is, with digital and physical endings.
Two people, four moments.
Sellers want control. Buyers want fewer steps. Each flow is designed for one of them.
List a product from your own profile
The uploader lives inside a seller’s profile, so listing feels like an extension of having one rather than registering as a merchant. A stepper shows how much is left.
See it the way buyers will
Before anything publishes, a preview of the listing exactly as a buyer meets it. The feature sellers needed most and asked for least.
Shop from one new icon
A single cart icon added to Uni’s side navigation opens a marketplace homepage with categories and featured products.
Pay without leaving Uni
Browse, filter, add to cart and complete payment inside the platform. Digital products download; physical ones get tracked.
Supply arrived faster than we projected.
We instrumented the marketplace before launch, so the numbers afterwards meant something. Within three months, sellers were listing without being chased.
Above pre-launch projections, within three months
On the platform the marketplace launched into
Active users across designers, firms and students
From research to launch, 2021
The 25% is measured against what we projected before launch, not against a previous marketplace; there wasn’t one. It says supply showed up faster than expected. It doesn’t by itself say the marketplace was profitable.
- Before
- Sales happened on other sites, and users left with them
- After
- Listing and buying happen inside Uni
- Before
- No way to find what the community was making
- After
- One marketplace with categories, tags and featured products
So, how did we get here?
From what shipped, to why it was designed this way.
We asked five hundred of them.
Rather than assume what the community wanted to trade, we ran surveys and interviews with stakeholders and 500 active users, sampled across designers, architecture firms and students at roughly 40 / 30 / 30.
70%+ of designers would list
Digital resources like templates and guides, and physical items too.
65% wanted to buy inside Uni
Rather than be redirected somewhere they had no relationship with.
Both sides already present
A marketplace with supply and no demand is a graveyard. Here both sat in one community.
Two people, pulling in opposite directions
Selena
Interior designer. Wants to find something specific, judge it quickly and buy without leaving a platform she trusts. Her enemy is friction.
Marco
Freelance architect. Wants listing to take minutes and to see exactly what buyers will see before going public. His enemy is exposure.
Four questions that shaped the marketplace.
Each started with more than one reasonable answer.
One marketplace flow, or two?
Treating the marketplace as one flow is how you end up with a listing form buried three levels inside a shop. I mapped the seller and buyer journeys separately. The buyer flow is one path up to payment and two after it: a digital product downloads, a physical one gets tracked.
A new visual language, or Uni’s own library?
A new segment is a tempting excuse to invent a new look, and the result would have been exactly the bolted-on shop the brief was trying to avoid. I worked from the existing library and extended components only where the marketplace genuinely needed something new.
A seller dashboard, or listing inside the profile?
- Alternative
- A separate merchant dashboard sellers register for.
- Decision
- The uploader lives in the profile, so listing feels like an extension of having one.
- 01
Tags that do real work
Keywords like home decor or modern are what make a listing findable later.
- 02
Sellers choose their call to action
Buy Now or Get It Now: a template and a print aren’t sold with the same verb.
- 03
A stepper across the whole form
The cheapest way to stop someone abandoning it.
- 04
Preview before publish
Nobody says “I’m nervous about publishing”. They just don’t publish.
Where should buyers enter the marketplace?
One cart icon added to a side navigation people already knew. One new element, and a whole segment behind it, so the buyer never feels they’ve crossed into somebody else’s product.
Tested before handoff, measured from day one.
Usability testing before handoff checked whether the flows behaved the way the diagrams claimed. The uploader is where it paid off: a listing form seems obvious to the person who designed it and is full of small stalls for everyone else. Then I stayed close through development with eight engineers.
Traffic and engagement
Google Analytics across the marketplace, set up before launch.
Supply
Listings growth, plus daily and weekly listing counts to see if the uploader was used or merely visited.
Demand
Retention, time in the marketplace, conversion rate and click-through.
The uploader and the purchase center.
Four steps for sellers, with a stepper so they always know how many are left. One icon in the side nav for buyers, with a whole segment behind it.
What Uni taught me
It would have been faster to design the marketplace as its own product and connect it with links. It would also have felt like an advertisement inside a community.
- 01
Native work is invisible, which is why it’s worth doing.
Reusing the library, entering from the profile, one icon in the existing nav: nobody notices any of it when it works.
- 02
A two-sided product is two products.
The buyer wants fewer decisions, the seller wants more control. Designing each side for what it needed made both better.
- 03
Make the reasoning easy to inherit.
With three designers and eight engineers, quality depended less on the design than on how clearly I explained it.
The community was already buying and selling. All the marketplace had to do was stop being somewhere else.
Designed with a team of three at Uni.xyz, two designers and a senior designer, alongside eight engineers, over ten months in 2021.