Ashish Sharma
Used car listing page
A browsing experience to search, filter, and evaluate used cars
About
Project is about redesign of the listing page which includes cards, widgets, filter discovery and overall layout of the listing page along with implementation of new DLS.
Company:
Full time with Cars24
Year:
2025
My role
I owned this project end to end.
> Research - Analyse the issue in the current listing page
> Ideation and iterations - New direction for the listing page
> Working with engineering to spec and ship it
Two designer.
Old listing page experience
We did UX audit, Interviews and also look at data to identify the issues in listing page
We conducted 16 interview for the cohort - new and repeat user before and after doing test drive
Overall feedback from audit and interviews
> Page feels like a home page wearing a PLP costume
> Decision fatigue kicks in before even user starts browsing
> Filter and serach are not discoverable
> Lack of explanation of seller types and zero user intent filtering on the page
> Trust is the biggest decision driver in used-car marketplaces
The numbers confirmed it:
> PLP to PDP conversion is 64%
> Filter usage was around 18% and user who use filter convert 2.3x more
> Search usage is 10% and user who use serach convert 1.6x more
> Around 70% new users don't use any tool to refine their listing
👉 In a high-inventory marketplace like cars24 where there can be 2000+ cars listing in a city, narrowing choices is a critical part of the buying journey.
Phase 1
Simplifying the listing experience to improve discovery and decision-making
> Simplified the page structure to reduce cognitive load and improve focus
> Improved how users enter, navigate, and refine their browsing journey
> Brought more clarity to information that supports decision-making
Phase 1
Car card
The car card fails to support confident decision-making
Phase 1 - Car card
Understanding sellers and assurances
Sellers structure
There are 3 main seller types currently at Cars24
Cars are procured as per the business model and mapped to these seller types
👉 New models keep on coming as business find new ways to sell cars like one upcoming model is "Demo
cars" - Cars which are used for test drive at showroom
List of promises
There are 11 promises in total which are mapped to the cars as per the seller and business model
Sometimes, these promises car applied at the car level to give assurance to user for better cars
👉 So there is no clear structure of mapping the promise - is it at car level or seller ??
Phase 1 - Car card
Benchmarking
Seller types & card patterns
We analyzed leading platforms to understand how inventory is structured and communicated.
> Inventory is segmented by seller type or trust (Assured, Certified, Dealer, Owner)
> Trust is communicated via badges and labels
> Seller type is often paired with a value proposition (inspected, certified)
> Car is presented either on the real condition or some platform
> Low clarity in differentiating seller types
👉 While platforms segment inventory using seller types and trust markers, the lack of clarity and consistency makes it harder for users to quickly understand and compare options.
Phase 1 - Car card
Intial ideation
Exploring multiple directions to balance trust, hierarchy, and decision-making clarity
> Important elements to consider - Variant dedicated line, CTAs, Carousel, trust signals
> Explored different information hierarchies (price, trust, image) to guide user focus
> Tested price vs EMI placement and overall information grouping for clarity
> Iterated on card density and CTA patterns to improve scanability and comparison
Phase 1 - Car card
Rethinking Trust
We identified that seller type and trust promises were tightly coupled, limiting clarity and scalability.
> Seller ≠ Trust
> Same seller can have different quality levels
> New business models were hard to accommodate
For this enable we need a structured and scalable trust model and the need to differentiate seller types from promises that we offer. This led to the idea of creating an Assurance system — a tiered framework that represents the level of checks, benefits, and reliability associated with each car.
We did same activity for the Seller name
Finally, we arrive at the seller name as
> Cars24 Owned Stock - Cars24 stock listed on the platform
> Verified Direct Seller - Customer listing cars on platform to sell
> Trusted Partner Dealer - Dealer lisying cars on platform for selling
Assurance tier name as
> Zero Worry max - All 11 promises included
> Zero Worry Plus- Only 8 promises included
> Zero Worry - Only 5 promises included
Phase 1 - Car card
Final design
Important decision for car cards
> Subtle branding for the seller (as centre pill) and assurance
> Car sitting in the platform which changes colour as per seller type
> Price and name against each other for better clarity
> Clear assurance pills as per seller and car
> Actionables
Phase 1 - Listing page
Final PLP
It includes
> Filter and search discoverability
> Business tabs with explanation
> Intent filter via Dynamic inline multi facet filter widget and other single facet filter widgets
> Seller, assurances explanation widgets
Sorting logic of the listing
We change the sorting logic of the edge cases where users have searched or applied filters
> If result not found in Cars24 stock (GS) in the city then expand it to other sellers (private seller, dealer)
> If result not found in the city including all sellers then move to nearest city which has same car in any seller (Preferences remain to Cars24 stock)
> If result not available in same city and nearest city in any seller then show similar cars in users's city
Phase 1 - Car card
User testing & Impact
Testing
We conducted two rounds of usability testing
We tested the redesigned PLP with both new and repeat users, iterated on the identified issues, and validated the updated experience through a second round of testing
> “CARS24 Owned Stock” became the most trusted and easily understandable seller type
> Users better understood assurance benefits and trust signals
> Sticky filters and improved discovery supported browsing flow
> Scrollable imagery and richer interactions increased engagement
> Users appreciated clearer comparison between assurance tiers
> Seller segmentation created initial curiosity and improved discovery
> Colored filter chips got strong visual attention
> Cards have multiple colours
👉 Users responded positively when trust signals, seller communication, and information hierarchy made browsing easier and decision-making more confident.
Number showed this as well
Magic Scale conversion:
After
Before
Conversion
PLP to PDP:
74%❤️
64%
Test drive bookings from PLP:
7.5% ❤️
0
Filter usage:
38%❤️
18%
Seacrh usage:
12.5%❤️
10%
Phase 2
It involves improving the inside filter page of the filters
We redesigned the filter experience from discovery, understanding, interacting and finally applying
Phase 2 - Filters
Current filters
We did UX audit for the filters and understood what was not working
>Inconsistent interaction patterns → lists, grids, sliders, pills all behave differently
>Weak selection visibility → users can’t clearly see what’s selected or applied
>Poor feedback during interaction → no real-time response (especially sliders & counts)
>Find-ability issues → brand search, logos, and selected states are hard to track
>Readability problems → low contrast, small text, truncated labels
What data is telling us
> 22% of the all users interact with filters and go inside
> Nearly 18% drop-off happens:
PLP → Filter page → Apply filters
> Top filter usage is concentrated around 5 key filters: Budget, Make & Model, Transmission, Model Year, and Kms Driven
👉 The most-used filters are scattered across the list, increasing effort and discovery time for users
Phase 2 - Filters
Benchmark
We looked at a competitors and indirect businesses as well to get full understanding of the patterns
We benchmarked around 20 companies.
> Left-panel navigation improves scanability and reduces interaction effort on mobile
> Visual filters (icons, chips, cards) improve recognition and reduce cognitive load compared to text-heavy lists
> Platforms optimize filters for faster decision-making, not just better filtering
Phase 2 - Filters
Final design
Designed to make filtering faster, clearer, and easier to navigate.
> Improved filter hierarchy and grouping for faster navigation
> Reduced interaction effort with quicker selections and better mobile ergonomics
> Added visual cues, contextual information, and richer explanations for better decision-making
> Increased discoverability of high-usage filters through search, highlights, and smart grouping
> Introduced more intuitive and scalable filter patterns inspired by marketplace benchmarking
Phase 2 - Filters
Impact
Number showed this as well
Deeper engagement in filtering Users applying 2+ filters increased
Magic Scale conversion:
After
Before
2 filters:
30%❤️
25%
3 filters:
18% ❤️
13%
4+ filters:
12%❤️
10%
1 filter:
52%
44%
Indicates users are exploring and refining more
Stronger use of decision-driving filters
Magic Scale conversion:
Impact
Budget:
+2.64%❤️
Fuel:
+1.93% ❤️
Year:
+2.82❤️
User who are using combination of 3-4 filters have BI at 8.5% and normal BI is 5%
👉 Users who applied 3–4 filters were able to narrow down thousands of listings into a more personalized set of ~20–30 cars, making decision-making significantly easier
Phase 3
The goal was to help users narrow thousands of listings into a sweet spot of ~20–30 highly relevant cars, making discovery more personalized and decision-making easier.
> Recommend a car flow — Users answer a few simple questions to quickly reach a curated set of relevant cars within the sweet spot.
> Saved filters — Helps repeat users quickly return to previously refined preferences and continue discovery seamlessly.
> Personalized pre-PLP experience — A tailored entry layer that helps both new and repeat users narrow down choices before entering the full listing experience.
For new users:
Use-case based quick discovery (Family, Budget-friendly, City driving, etc.), Guided recommendation flow, Search-led exploration
For repeat users:
Previously applied + saved filters, Continue journey widgets, Recommended cars based on past interactions
Futher scope of improvement
Phase 3 focuses on helping users narrow down large inventories into a smaller set of highly relevant cars, making comparison and decision-making easier.
Thanks for reading 😊
