See full Figma file

Redesigning the page dealers bid on

A ground-up redesign of Spinny's live car-auction listing page - built for speed, scannability, and split-second bidding

1 designer, 2 PMs ~3 months 4 versions shipped iOS & Android
My role — IC designer (individual contributor)
Before and after - Spinny auction PLP redesign
+13%
More retention
+9%
Revenue growth
−40%
Fewer cars missed
NPS jump

* Before shipment vs after shipment · numbers rounded off

What is the auction app

Source stock without chasing the market

Spinny Partners is the app 80,000+ used-car dealers use to source their cars - thousands of inspected cars in live auctions with verified condition reports, so buying stock is fast and low-risk instead of guesswork across the market

1

Browse

Thousands of inspected cars, each with a verified condition report

2

Bid

Every car goes live for a set window - dealers place their bids

3

Win

The highest bid moves ahead on the car

4

Negotiate

Settle the final price with the seller, right in the app

5

Pay & receive

Pay after inspection - the car is delivered to the door

Dealers were losing cars

The busiest screen in the app had quietly become the slowest to read

Every hesitation cost a car -
and the revenue with it
clutter + a countdown = missed bids
What was going wrong
  • Timers ran out before dealers finished judging a car
  • Images so small, dealers saw one car at a time
  • Banners, IDs, tags & CTAs ate the real estate
  • Language dealers misread - losing trust
The cost
73%

of a dealer's stock was going to Cars24 - and it won on two fronts

An easier page
→ the half this redesign owned
More inventory
→ a supply bet, out of scope for this project

The process

1

Audit

2

Benchmark

3

Meet dealers

4

Design

5

Internal discussion

6

Iterate

7

Show dealers

8

Iterate

9

UI sanity

10

Watch & improve

Assume nothing - watch dealers bid

Full research

Two research methods, run in dealers' own shops

20+ dealers
Method 01 · Behavioural

Usability sessions

Sit beside dealers & watch them bid - how fast? what first? where do they struggle?

Usability session 1 Usability session 2 Usability session 3
12 dealers
Method 02 · Attitudinal

Card sorting

Every piece of card info, sorted by importance - thinking aloud

Card sorting session 1 Card sorting session 2 Card sorting session 3

How many of 12 dealers valued each piece of info

Engine rating
12 / 12
Exterior rating
10 / 12
Fitness expired
10 / 12
Seller agreed price
7 / 12
Time remaining
6 / 12
Fair value
4 / 12

Every dealer checked the engine rating - though half didn't fully trust it

Worth adding

  • “Accidental” damage flag10 / 12
  • No-damage & not-flooded tagsasked
  • Colour as a filter10 / 12
  • Engine-sound cueworkflow

Worth removing

  • “Finance available” - misread12 / 12
  • “Recommended” - ignored/not understandable11 / 12
  • Auction ID - they screenshot very rarely9 / 12
  • Fair value on the card9 / 12

26 problems hiding in plain sight

Sorted two ways — what was broken, and what was missing.

What's broken13
Unclear signals
Star ratings are unclear
Location de-emphasised
Tags not understood
Banners shown all the time
“Prime” tag on cars is not understandable
Wasting space
CTAs wasting space
Auction ID wasting space, rarely used
Car make wasting space
Random white space
Shaky, inconsistent UI
Small, un-scrollable images
Inconsistent card sizes based on the content
Missing & delayed notifications (bug)
Chaotic layout
Old Spinny PLP - problems
What's missing13
No dedicated homes
No room for upcoming cars
No room for closed cars
A dedicated section for pending actionables
No user control
Wishlist hidden
“Hide the car” feature missing
No feedback on actions
No easy way to see cars near selected RTOs
No context or discovery
“For you” underused, current bids compromised
Similar cars buried
No link to a car's past
Scrap/fitness-expired cars info not shown
No alert when previously bid cars return
No banners between cards to push features
Competitor · Cars24
  • Bigger, clearer images
  • Plainer copy - “they write clearly, we write superficially”
  • More inventory - the real reason for a 2nd app
Terminology gap
  • “Structural damage” → read as exterior damage
  • “Finance available” → read as a loan on the car
  • We weren't speaking dealer

Where dealers actually go

Daily path to a bid

Daily active dealers
100%
Open the listing page
92%
Move on to bid on a car
37%

Nearly the entire dealer base passes through the product listing page, daily

In dealers' own words

Hamare counter pe kya bikega, usse matlab hai - specs, not badges
UPDealer · Uttar Pradesh
With C24 I'm 99.9% sure I'll get the car. With us, even after OCBs, it often won't close
SIDealer · Sirsa
I'd rather book a ₹4k flight and drive it back than pay ₹20k transit
HRDealer · Haryana
Bigger images help me decide faster - I read colour & variant from the photo
DLDealer · Delhi
I thought “structural damage” meant exterior. “Accidental” is clearer
PBDealer · Punjab
Under 40 seconds left? I won't even open the car
DLDealer · Delhi
“Prime” ka matlab aaj tak clear nahi - bas ek gold tag dikhta hai
RJDealer · Rajasthan
Jis car pe pehle bid ki thi, wo laut aaye to pata hi nahi chalta
GJDealer · Gujarat
Scrap ya fitness-expired car pehle se pata ho, to mera time bache
MPDealer · Madhya Pradesh

Issues surfaced, sorted

Product
  • Photo zoom friction
  • Ratings shown too small
  • Copy felt evasive
Business
  • Low conversion
  • Slow token refunds
  • Thin inventory vs C24
Operations
  • Unresponsive KAMs
  • Inspection mismatches
  • After-sales gaps

Moodboarding

Where we looked

Moodboarding - references we looked at

4 updates, 3 months

Ship, test, improve, repeat

PLP version 1
how it started
PLP version 2
PLP version 3
PLP version 4
how it's going

Back to the shops

10+ dealers rated every change

Remove banners
10
Bigger images
10
Prime/Select inline
10
Remove CTAs
9
New “for you”
9
Engine-sound icon
8
Drop Auction ID
7
Ratings feel accurate
6
“More like this”
4
what stuck
  • Less clutter, bigger images, plainer words
  • Dealers asked us to remove even more
what we heard
  • Ratings must feel accurate to be trusted
  • A few wanted ID/fair value as a safety net
  • “More like this” split the room

The redesign

All screens

Same car, half the noise

Redesigned full screen
Redesign detail 1
Redesign detail 2
Redesign detail 3
Redesign detail 4
Redesign detail 4b
Redesign detail 5
Redesign detail 6

Decisions that mattered

Killed the clutter

Banners, ID, make, unread tags & CTAs - gone 10/10 dealers approved; some wanted more cut

Big, auto-scroll images

Dealers read colour & variant from photos most-loved change

Engine-sound cue

A speaker icon for models dealers listen to (Polo, Vento) loved once understood

Spoke dealer language

“Accidental” not “structural”; Prime & Select defined inline clarity = trust

Car card iterations

Other iterations that didn't make it

Card iteration 1
Card iteration 2
Card iteration 3
Card iteration 4
Card iteration 5
Card iteration 6

Bonus features & upgrades

All screens
Bonus feature 1
Bonus feature 2
Bonus feature 3
Bonus feature 4
Bonus feature 5
Bonus feature 6

Built for foldables & tablets too

The same system, scaled to bigger canvases

17% of our high-transacting dealers use foldables - enough reason to design for them too

Foldable / tablet screen 1
Foldable / tablet screen 2
Foldable / tablet screen 3
Foldable / tablet screen 4
Foldable / tablet screen 5
Foldable / tablet screen 6
Foldable / tablet screen 7
Foldable / tablet screen 8
Foldable / tablet screen 9
Foldable / tablet screen 10
Foldable / tablet screen 11
Foldable / tablet screen 12

Faster decisions, fewer misses

+13%
More retention
+9%
Revenue growth
−40%
Fewer cars missed
NPS jump

* Before shipment vs after shipment · numbers rounded off

The 40% drop in missed cars is the one I'm proudest of - proof dealers could finally read a car and act before the clock beat them. This was the "easier page" half Cars24 was beating us on; inventory stays a separate, ongoing bet

Future scope

We will make this even better

Fixing app notifications

Different car card views

“Not interested” in a car feature

Relation with car type based on history

Timer-based urgency features

AI calls for dormant dealers

What I took away

01

Removing is a feature

On a timer, every element cut is speed returned

02

Watch, don't ask

Shops surfaced what surveys never would

03

Speak their language

“Accidental” vs “structural” = trust or confusion

04

New icons need teaching

Loved once understood - not before

05

Ship in versions

Five releases beat one perfect launch

06

Let data settle it

20+ A/B tests over opinions