Connect
Universal Parts Catalog Redesign
The Universal Parts Catalog lets users browse generic items like tools, fluids, and accessories that don’t require exact vehicle fitment, using a category-based navigation.
Analytics showed that most users dropped off before adding anything to cart, with completion rates as low as 9% in some markets.
Role
End-to-End UX Researcher and UX/UI Designer
Duration
3 months
Tools
Figma, FigJam, Beamer, Userberry, Teams
Problem to solve
Users struggled to find the right universal part due to inconsistent naming, too many navigation steps, and unclear categories.
Solution
A reorganized category structure with fewer choices shown at once, visible subcategories inside each category card, clearer hierarchy, and a new "Top picks" feature letting users pin up to 5 frequently-used categories for quick access.
Final design
Background
The Universal Parts Catalogue was underused, with GA4 showing sub-50% add-to-cart completion across all markets over the previous 30 days.
Our goals were to reduce time-to-order, increase add-to-cart by 15%+, reduce abandonment by 20% per country, and improve the overall UX.
From there, we defined a set of hypotheses to test:
H1
Reducing the number of choices shown at once will help users find the right category faster and reduce catalog drop-off by 20%.
H2
Presenting a clear hierarchy on the catalog main page will help users find the right category more quickly and reduce drop-off by 20%, by boosting search confidence.
H3
Providing representative visuals for categories/subcategories will reduce average session engagement time by 50% and increase add-to-cart by 20%.
H4
Showing subcategories under the related category (rather than hiding them) will reduce drop-off by 20% and cut engagement time by 50%.
H5
Adding shortcuts to favorite categories/subcategories will reduce engagement time and drop-off, since the needed category is found faster.
Research
Analytics
GA4 funnels across 5 markets, plus 83 Microsoft Clarity sessions with heatmaps over a 90-day window.
Survey
117 responses, 54-day duration, avg. completion time 6:23.
1:1 interviews
customers, internal staff, and automotive professionals across markets.
Usability testing
265 responses, 13-day unmoderated test via Maze, 50% workshop owner/employee and 44% sales roles.
Heuristic evaluation
3 evaluators, Nielsen's 10 heuristics, cross-checked independently to catch violations a single reviewer might miss.
Competitor analysis
7 competitors reviewed, simulating the user journey to see how each structures universal/generic parts.
What we found
53%
Rated the catalogue hard to use. Of 117 survey respondents, 53% scored ease of use 6 or below, while only 16% rated it 9 or 10.
55%
55% of sessions showed users clicking unresponsive elements. Across 83 Microsoft Clarity recordings, more than half contained at least one dead click.
36%
36% of users interacted with category tiles, making them the most-used element on the page. But instead of opening the category, they expanded the sidebar, creating confusion.
Pain points
1
Navigation & Click Depth
Too many clicks to reach results; browser back button often fails; breadcrumbs don't update correctly when drilling down.
2
Category Structure & Placement
Items placed in unexpected categories (e.g. garage equipment inside universal parts, batteries misfiled); "clamps" category mixes two different types.
3
Visual Hierarchy
Too many categories shown at once (50+) overwhelms users; large empty white space on cards; inconsistent image quality; text sometimes too small to scan quickly.
4
Search Reliability
Free text search misses synonyms and inconsistent product naming across brands; VIN search unreliable; unavailable items often shown first.
5
Tile/Click Behavior
Users expect tiles to act like links (open a new page); instead they sometimes expand the sidebar, creating rage clicks and confusion.
6
Personalization & Efficiency
No way to save filters or favorite subcategories before this project; no shortcuts to frequently used categories.
Define
Prioritized around reducing cognitive load and click depth rather than trying to fix every issue raised (search reliability, data/naming inconsistency, and VIN accuracy were flagged as separate, non-UX workstreams).
Core UX-addressable problems:
Choice overload: too many categories/subcategories shown at once, no clear hierarchy.
Unclear click behavior: tiles and breadcrumbs not behaving as users expect.
No quick access to frequent categories: every session starts from zero, regardless of past behavior.
Affinity mapping of interview insights.
Clusters ranked by severity and frequency.
How might we…
simplify the category navigation so users can reach the right subcategory in fewer, clearer steps, and return to frequently used categories faster.
Ideation
Two directions were explored for how subcategories should appear inside a category card:
Version 1 showed subcategories directly, with a "show more" button to reveal additional ones beyond what fit on screen.
Version 2 kept subcategories hidden until the user clicked into the category, trading a cleaner initial view for an extra step.
Both went into prototype testing rather than being narrowed down first, since the trade-off (visible now vs. clean now) wasn't obvious without user data.
Design
The category card
Each category card shows a first-level image and name, with second-level subcategories inside the same card rather than behind a click, addressing the "no extra click" request directly. A scrollable list handles cases with more subcategories than fit on screen (this replaced the "show more" button, covered in Testing).
Click behavior
Clicking a tile always opens that category's page, never expands the sidebar. This was the single biggest source of confusion in the original design (36% of all clicks, dead-click and rage-click behavior) and became a hard rule applied consistently across every navigation level.
Top picks
Users can pin up to 5 categories or subcategories at any level for quick return access, directly addressing the "let me favorite my categories" feedback. Pinned items live in a collapsible section on the catalog homepage; hitting the 5-item limit prompts a replace-oldest confirmation rather than silently failing or blocking the action.
Responsive behavior
3 cards per row above 1440px, 2 cards per row between 834 and 1439px (with horizontal scroll for Top picks), 1 card per row below 834px (Top picks condenses to 4 cards to fit the screen without scrolling).
Link types
Distinct icons for internal categories, external links, PDFs, and video, so users know what a click will do before clicking it, addressing the "unclear what this does" feedback from Clarity and heuristic evaluation.
Testing & Iterations
01 / TWO DIRECTIONS, ONE SHARED FLAW
Tested V1 against V2 head to head
V1 showed subcategories directly but hid extras behind "show more." V2 exposed everything but felt cluttered. Both were rated overwhelming past 50+ categories, with wasted white space on cards.
02 / USERS WANTED BOTH, NOT EITHER
Feedback converged on a hybrid direction
V1's clearer imagery, V2's no-extra-click access, plus the ability to pin frequent categories, which became the seed for Top picks.
03 / VALIDATING THE HYBRID
Tested the final design with 265 users
Finding a specific item hit 96.8% success (33.4s avg); returning to the overview hit 93.8% (22.4s avg). Both strong, but misclick rates stayed high (49.8% and 67.8%), showing users got there, just not always cleanly.
04 / ONE MORE FIX AFTER TESTING
Removed the click entirely
The "show more" button was replaced with a scrollable list, echoing a complaint raised separately in sales feedback.
V1 Comparison banner.
V2 Comparison popup.
Results
Since launch, GA4 tracked real usage of the redesigned catalog in Switzerland and Czech Republic. The picture is mixed: strong list-to-purchase conversion where the catalog was used, but the underlying problem it was meant to solve isn't fully resolved.
Switzerland (Jan through Aug 2026)
1.27M items were viewed in the catalog list, 28.8% were added to cart, and 21.3% converted through to purchase. Once someone reaches the catalog, it performs.
Top picks adoption
177 pins recorded against 17 unpins over the same period, in isolated spikes rather than steady use. The feature is being used, but hasn't become a habit yet.
Czech Republic (Mar through May 2026), catalog vs. free-text search
Free-text search drove Kč327M in revenue from 8,845 active users. Category browsing drove Kč32M from 5,400 active users, roughly a tenth of the revenue from a similar user base. Even with the redesigned catalog live, users who can search directly still bypass browsing almost entirely.
What we still don't know
We don't have a clean before/after funnel isolating the redesign's effect from other factors over the same window (seasonal demand, other catalog changes). What the data does show clearly: browsing remains secondary to search wherever search is available. The original goal, making category browsing viable for users without an exact part name or number, was only partly met. It works better than before for the users who rely on it, but most users still avoid it when they can search instead.
Final Thoughts
01
Testing two extremes can reveal the answer is neither.
V1 and V2 were opposing bets, one image-heavy, one click-light. Users didn't pick a winner, they wanted both.
02
Success rate isn't the whole story.
Both validated tasks hit 90%+ success, but misclick rates stayed high (49.8% and 67.8%). Users reached the goal, just not always cleanly on the first try.
03
Iteration doesn't stop at the test.
The "show more" fix came after the Maze test had already closed. It took a separate feedback source to surface it.
04
A better catalog doesn't replace the need for better search.
Even with the redesign live, free-text search still outsold catalog browsing 10x in Czech Republic. Fixing browsing helps, but it doesn't remove the case for fixing search too.