This project was to design Lancy, a freelance job aggregator. The platform collects gigs from the major freelance job boards into one central feed and then filters out low-quality work for the user. Then, with advanced filtering capabilities, users can easily find exactly what they’re looking for.
ROLE
UX Designer
TEAM
1 developer,
1 product manager
TIMELINE
6 weeks
TOOLS
Figma, Userbrain
WIREFRAMES & TESTING
Filter Layout & Structure
The first core task I focused on was how filtering the job feed would function—an essential part of the Lancy experience, given its core value proposition of helping users quickly surface relevant gigs.
After analyzing filtering patterns in other job board platforms and wireframing multiple layouts, I narrowed the test down to four variants: (1) a single scrollable row of filters, (2) two stacked rows of filters, (3) a filtering sidebar sorted into groups, and (4) a sidebar with ungrouped filters.
Test 1: Filter Structure
Variant 4– the ungrouped filtering sidebar –performed best across the board, with a 28% faster task completion and a qualitative ease rating of 9.5/10, +.75 higher than average.
Feedback on the other designs was more mixed. The horizontal scroll layout earned an average ease rating of 9.0, but users completed tasks 25% slower than their personal average. The filters in two stacked rows was 6.5% faster than average but received the lowest qualitative ratings overall.
Test:
Unmoderated, 4 participants
Opt 1:
Single, Scrollable Row of Filters
Opt 2:
Two Rows of Filters
Opt 3:
Grouped Filtering Sidebar
Opt 4:
Ungrouped Filtering Sidebar
Follow-Up Test: Were there consequences to showing all the available filters at once?
Even though the ungrouped filtering sidebar performed the best, I ran another test before making the official design call. I wondered: was showing the filters all at once too much information for the user to take in? Even if it was a faster filtering experience, would it have a negative effect on other platform tasks?
I did a quick follow-up test to check for signs of cognitive overload.
Test 2: Checking for Cognitive Overload
In this test, participants were asked to locate specific jobs in the feed, using interfaces with the previously tested filter designs. Task completion times remained consistent across all variants—assuaging my concern that showing all filters at once may detract from the user's ability to quickly scan the job feed.
Test:
Unmoderated, 4 participants
I moved forward with the ungrouped filter sidebar officially.
Opt 4:
Ungrouped Filtering Sidebar
Opt 1: Placement at top of sidebar
Opt 2: Placement above job feed
The Quality Jobs Only Filter
Because the “High-Quality Jobs” filter is tied to Lancy’s pro subscription, its placement and discoverability were critical. Because of its importance, I tested this feature separately from the broader filter structure.
Two variants were tested:
A toggle within the filter sidebar
A toggle at the top of the job feed
Test 3: Toggle Placement
Results were decisive: 4 out of 4 participants immediately moved their cursor to the filter sidebar when asked to show only high-quality jobs in the feed. The toggle placement at the top of the filtering sidebar thus unsurprisingly led to 50% faster task completion.
Test:
Unmoderated, 4 participants
I moved forward with the placement at the top of the filtering sidebar.
Final Designs
Search & Filter
Filters in Use
Search Functionality
Quality Jobs Filter
Quality Jobs Filter On
Filter Off
Saved Posts
Saved Job Posts
Reflection
Lessons Learned
Talking directly to our target audience was invaluable. Our early user interviews revealed unexpected pain points and helped us uncover critical nuances we would’ve otherwise missed. Initially, we assumed users wanted more job posts, but through conversation, we learned it wasn’t about quantity: it was about relevance. Freelancers weren’t overwhelmed by too few options; they were overwhelmed by too many irrelevant ones. This insight fundamentally reshaped our approach, shifting the product’s focus from volume to intelligent filtering. If we hadn’t engaged users early on, we would’ve invested in solving the wrong problem.