Rastro search design
How insights from live auctions improved product discovery for antique hunters.
Role
UX/UI Design
01
Project Context
Rastro is a web app that aggregates products from Europe’s top auction platforms. I led the UX/UI design, collaborating with a Y-Combinator-backed team of ex-Shopify and Uber developers to design Rastro’s search interface—on a tight one-week timeline.
02
Project Challenge
Most online auction platforms, like Interencheres, follow the design patterns of eBay and Amazon—text-heavy, search-first, and designed around product titles rather than visuals. Here are some examples:
Through conversations with antique dealers, I uncovered two key insights about their shopping preferences:
They preferred live auctions because they could visually scan items faster than online listings.
Discovery was often accidental. Their best finds were from recommendations not search queries.
Even among these professionals, the aesthetics of the experience matters: accidental visual finds were not encouraged by the existing information-dense interfaces. Less than 50% of screen space was dedicated to images, and recommendations were buried. These sites were built for keyword search—not for visual browsing.
03
Design Approach
With just a week to design the interface, I focused on one goal: maximizing visual discovery. Instead of relying on text and filters for searches, users needed a way to explore items through images.
I looked to platforms that successfully balance search with discovery, and apps like Vinted, Pinterest, and Gem stood out. Their Masonry grids put the product images first, with visual recommendations systems that guide the user to their next product page.
Image-first using a masonry grid.
Recommendations driven.
Query refinement for power users.
04
Design Solution
The resulting interface dedicated 90% of the screen to images using a masonry grid. The design is recommendation driven, allowing merchants to stumble upon a find through visual algorithms. Power users are also able to double down on a product's details or use search filters to find a niche sale.
05
Results
We achieved an average of 13 page views per session, which doubled eCommerce benchmarks seen by the founders at Shopify. 70% of these page views came through recommendation clicks, proving to be more effective at getting users to product pages than through search queries.
While the market size wasn’t large enough for VC-scale returns, the product was a hit with antique collectors. The team pivoted, but the success of this launch cemented our collaboration, and we still continue working together.
Rastro search design
How insights from live auctions improved product discovery for antique hunters.
Role
UX/UI Design
01
Project Context
Rastro is a web app that aggregates products from Europe’s top auction platforms. I led the UX/UI design, collaborating with a Y-Combinator-backed team of ex-Shopify and Uber developers to design Rastro’s search interface—on a tight one-week timeline.
02
Project Challenge
Most online auction platforms, like Interencheres, follow the design patterns of eBay and Amazon—text-heavy, search-first, and designed around product titles rather than visuals. Here are some examples:
03
Design Approach
With just a week to design the interface, I focused on one goal: maximizing visual discovery. Instead of relying on text and filters for searches, users needed a way to explore items through images.
I looked to platforms that successfully balance search with discovery, and Pinterest stood out. The Masonry grid puts the product images first, while recommendations take center stage on the product details page.
04
Design Solution
The resulting interface dedicated 90% of the screen to images using a masonry grid. The design is recommendation driven, allowing merchants to stumble upon a find through visual algorithms. Power users are also able to double down on a product's details or use search filters to find a niche sale.
Image-first using a masonry grid.
Recommendations driven.
Query refinement for power users.
05
Results
We achieved an average of 13 page views per session, which doubled eCommerce benchmarks seen by the founders at Shopify. 70% of these page views came through recommendation clicks, proving to be more effective at getting users to product pages than through search queries.
While the market size wasn’t large enough for VC-scale returns, the product was a hit with antique collectors. The team pivoted, but the success of this launch cemented our collaboration, and we still continue working together.
Mohana feature redesign
Through conversations with antique dealers, I uncovered two key insights about their shopping preferences:
They preferred live auctions because they could visually scan items faster than online listings.
Discovery was often accidental. Their best finds were from recommendations not search queries.
Even among these professionals, the aesthetics of the experience matters: accidental visual finds were not encouraged by the existing information-dense interfaces. Less than 50% of screen space was dedicated to images, and recommendations were buried. These sites were built for keyword search—not for visual browsing.
Rastro search design
How insights from live auctions improved product discovery for antique hunters.
Role
UX/UI Design
01
Project Context
Rastro is a web app that aggregates products from Europe’s top auction platforms. I led the UX/UI design, collaborating with a Y-Combinator-backed team of ex-Shopify and Uber developers to design Rastro’s search interface—on a tight one-week timeline.
02
Project Challenge
Most online auction platforms, like Interencheres, follow the design patterns of eBay and Amazon—text-heavy, search-first, and designed around product titles rather than visuals. Here are some examples:
03
Design Approach
We wanted to clearly present a user's actions, while linking them to their specific symptoms and blood test results. My role involved redesigning the action cards and description screens to be concise, actionable, and directly tied to these metrics. I focused on improving the information hierarchy to help users quickly implement changes, while keeping educational content accessible for those seeking more details. We aimed to validate the redesign by surveying our beta cohort, with a goal of at least 50% finding the new design more impactful.
04
Design Solution
The resulting interface dedicated 90% of the screen to images using a masonry grid. The design is recommendation driven, allowing merchants to stumble upon a find through visual algorithms. Power users are also able to double down on a product's details or use search filters to find a niche sale.
05
Results
We achieved an average of 13 page views per session, which doubled eCommerce benchmarks seen by the founders at Shopify. 70% of these page views came through recommendation clicks, proving to be more effective at getting users to product pages than through search queries.
While the market size wasn’t large enough for VC-scale returns, the product was a hit with antique collectors. The team pivoted, but the success of this launch cemented our collaboration, and we still continue working together.
Through conversations with antique dealers, I uncovered two key insights about their shopping preferences:
They preferred live auctions because they could visually scan items faster than online listings.
Discovery was often accidental. Their best finds were from recommendations not search queries.
Even among these professionals, the aesthetics of the experience matters: accidental visual finds were not encouraged by the existing information-dense interfaces. Less than 50% of screen space was dedicated to images, and recommendations were buried. These sites were built for keyword search—not for visual browsing.
Image-first using a masonry grid.
Recommendations driven.