Project Summary
I spent 18 months designing a marketplace that worked exactly as intended, and that almost no one used.
The product wasn't broken. The flows were clean. The dashboard was scalable. Users onboarded, signed in, and then quietly went back to closing deals over WhatsApp.
That signal forced a reframe I didn't expect to be making. The bottleneck wasn't a missing buyer-seller interface. It was that the people actually closing deals, our internal traders, were running multi-million-dollar transactions on spreadsheets, with no shared view of what was happening across the team.
I led the pivot from a self-service marketplace to internal tools that supported human-led trading.
This is the story of how I got there.
Responsibility
Defining user flows and system structure
Designing the dashboard and key interactions
Creating prototypes and validating ideas
Iterating based on feedback from internal teams and users
Collaborating closely with product and engineering
Contributing to product direction as insights emerged
Tools
Figma (Figma Make)
Miro
Lovable
Claude
The World We Walked Into
A scrap metal trader doesn't start his morning by opening a dashboard. He starts on WhatsApp. He has nine threads open, three of them with buyers he's known for over a decade. He sends a voice note. He gets one back. By 10am, he's brokered a deal for a container of copper worth more than most people's annual salary. This is not a digital-native environment.
Trust in this industry is built through:
Years of repeated transactions
Personal relationships between buyer and trader
Direct communication on familiar channels
Reputation that travels by word of mouth
The users we set out to serve weren't anti-technology. They were optimizing for something a marketplace can't deliver on its own: certainty about who's on the other side of a six-figure transaction.
I didn't fully understand that when I started.
What We Believed
The hypothesis felt obvious at the time. Give people better tools, and they'll prefer autonomy and transparency over phone calls and inbox sprawl.
I designed a dual-sided platform on that premise.
Sellers could create listings;
Buyers could post requests;
Both sides could track orders through a single dashboard;
Status updates would flow automatically instead of through follow-up emails
For 18 months, I refined that platform. Onboarding got smoother. Information architecture got cleaner. We added a marketplace, then bidding. By every internal review metric, the product was complete.
It was also, increasingly, beside the point.
Buyer dashboard:
The First Cracks
The cracks didn't appear all at once. The first sign was that engagement curves flattened almost immediately after onboarding. People created accounts. They looked around. They created a listing or two. Then they stopped.
When we asked them what they were doing instead, the answers were consistent. "I called my contact." "I sent it to the trader." "I prefer to speak to someone." We weren't being rejected. We were being routed around.
We were designing for efficiency. Users were designing their workflow around trust.
I tried one more iteration. We launched a public marketplace with bidding, the closest thing to a self-service feature loop we could ship. Traffic improved. Engagement metrics looked healthier. For a few weeks, it felt like the thesis was being saved.
Then we looked at what bidders did after they placed a bid.
They asked to speak to a trader.
The product wasn't a failed marketplace. It was a fancy contact form for the human relationship users actually wanted.
That was the moment the strategy needed to shift, not the design.
Home-page (Including the bidding feature):
The Thing Hiding in Plain Sight
While I was focused on user-facing flows, a different problem had been quietly developing inside the company.
Our traders, the people actually closing deals, were working out of spreadsheets. Each trader maintained their own. Deal information was scattered across columns that meant different things to different people. There was no shared view of what was in flight, what was stuck, or what had just closed.
Leadership had tried to fix this once already, by rolling out Pipedrive. It didn't take. The tool didn't match how traders thought about deals, so they quietly went back to their spreadsheets.
I'd been so focused on buyers and sellers that I'd missed the most important user in the system.
It was the trader.
And the trader was the one we hadn't designed for.
Making The Case
This was the hardest part of the project. I had spent a year and a half building the marketplace. So had engineering. The business had told the market we were a buyer-seller platform. Pivoting toward internal tools meant acknowledging that the public-facing product wasn't going to drive the business on its own.
I put together a case grounded in what behavior was telling us:
Users were not adopting self-service. They were repeatedly choosing humans.
Traders, the ones humans were choosing, didn't have the tools to scale.
The constraint on growth wasn't the marketplace. It was the trader's workflow.
The reframe leadership accepted: don't replace the trader. Equip the trader.
That decision changed what we built next.
What I Shipped After The Pivot
Internal trading platform A system for traders and ops to track deals from negotiation through completion. I prototyped real workflows in Lovable with traders directly, not through written specs. We tested end-to-end scenarios live. By the time engineering picked it up, the tool already reflected how traders actually worked, not how I'd assumed they did.
Marketing site redesign The previous site sold us as a marketplace. The new product story was different: human expertise, supported by software. I led a redesign that repositioned the company around trust and capability. Interactive prototypes aligned messaging and structure with leadership early, so the strategic conversation happened upstream of any code.
A new design-to-engineering workflow Both projects shipped through a "vibe coding" handoff. High-fidelity prototypes, built to feel real, traveled from design into implementation without losing intent. The gap between what was designed and what was built got smaller. Reviews got faster. Stakeholders signed off on working scenarios instead of on screens.
Outcomes
Product strategy shifted from self-service marketplace to a hybrid model where digital tools support, rather than replace, human-led trading.
Traders moved from fragmented spreadsheets to a shared, single source of truth for
in-flight deals
Stakeholder decisions moved upstream of engineering investment, reducing rework
External positioning rebuilt around what was actually true about the business
What I Now Bring To Every Brief
1. Adoption is the only validation that counts
A coherent design no one uses is a failed design. I treat adoption as a primary success metric, not a downstream effect of "good design."
2. Behavior beats feedback
Users told us the platform was useful, then didn't use it. Watching what people actually do, especially what they do when they think no one is looking, has consistently outperformed what they say in interviews.
3. Trust isn't a UX problem
In relationship-driven industries, the role of design isn't to replace trust. It's to understand how trust already flows through the system, and design tools that support those moments instead of competing with them.
4. The most important user is often not the end user
The leverage in any system tends to sit with someone less obvious than the customer. Asking "who has the most leverage on the outcome we care about, and are we designing for them?" early has changed the direction of more than one project since.
5. Prototyping is a strategy tool, not a delivery tool
The right prototype changes what gets decided, not just what gets built. Lovable shifted my work from showing screens to running scenarios. Stakeholders engaged differently. Decisions got better.
Reflection
I started this project focused on interfaces. I left it focused on systems. I used to believe good design could pull users toward better behavior. I now think behavior in mature, relationship-driven industries is mostly fixed, and the role of design is to work with it rather than fight it.
If I were given this brief again today, I wouldn't start with a dashboard. I'd start by mapping how trust flows through the business. Who people rely on, when they need reassurance, and where digital tools can support those moments instead of replace them.
The most valuable outcome wasn't the product we built. It was learning where not to build.
That has shaped every brief I've taken on since.
© 2026 Nasim Raeesi