Back to work
Reverse pickup,
rebuilt for the field
A guided, camera-first pickup flow that helps Delhivery field executives verify returns at the doorstep and stop wrong pickups before they leave the customer’s hands.
PLATFORM
Delhivery FE App
USERS
70,000+ field executives
TIMELINE
3 weeks
THE OUTCOME
16%
Increase in pickup rate
33% → 14%
False-positive pickups
13%
Drop in reverse pickup losses (45.5 lakh)
01
CONTEXT
Every return travels backwards.
The rider is its only checkpoint.
In last-mile logistics, the pickup and return flow carries goods back from the customer’s doorstep to the dispatch center. Field executives are the key touchpoint on that journey: they collect the item, verify it is the right product, and carry it back. When verification fails at the door, the loss shows up much later, and it is expensive.
CHALLENGE
Complex, manual steps slowing down every pickup
SCALE
70,000+ field executives using the flow every day
GOAL
A pickup flow that is fast, guided and hard to get wrong
02
THE PROBLEM
One tap to approve.
Nothing to check it against.
The old flow asked field executives to confirm a product match with a single Yes or No, often while the customer waited. There was no guidance on what to inspect, how to photograph it, or how to recognise the original product. Wrong items slipped through, and the cost surfaced downstream.
20%
Pickup failure rate
3.4 Cr / month
Lost to wrong pickups
“I don’t want to mistakenly pick up a wrong shipment, or else I will be penalised for it.”
KISHORE, FIELD EXECUTIVE
03
THE FLOW WE INHERITED
Eight screens. One photo. Zero guidance.
1
One question per screen
QC was split into a linear, three-step questionnaire that riders had to tap through one at a time.
2
A ceiling, not a minimum
“You can take maximum 4 pictures” set an upper limit, so most riders stopped at one.
3
Guidelines in fine print
Photo rules were a block of text above a button, easy to skip when a customer is waiting.
4
A single Yes is enough
Nothing stopped a rider from photographing a sealed package and confirming the pickup.
04
RESEARCH
We rode along before
we redesigned anything.
We made several field visits to our dispatch centers, accompanied delivery personnel on their routes, gathered direct feedback, observed how they interact with customers, and documented every pain point along the way.
KEY INSIGHT
Field executives are often in a rush, and typically pick up the product after only a quick inspection.
1,500 wrong pickups, studied one by one.
We went through roughly 1,500 samples of wrong pickups to find out where verification was breaking down. Three patterns stood out.
WRONG PICKUPS BY CATEGORY
65%
Apparel & Accessories
25%
Other categories
10%
Shoes & Footwear
Different categories fail in different ways. Apparel alone accounts for nearly two-thirds of wrong pickups, so QC has to be tailored by category to stay fast and accurate.
05
WHO WE DESIGNED FOR
Meet Kishore. He gets paid
per pickup, not per minute.
Kishore Pal
AGE
24
EDUCATION
12th pass
STATUS
Single
OCCUPATION
Field Executive
LOCATION
Gurgaon
TECH LITERACY
Mid
“I am used to my assigned locality. The more shipments I deliver or pick up, the more I earn.”
BIO
I am currently working in Gurgaon. I want to earn money and buy a motorcycle. I soon plan to get married.
PLATFORM
Delhivery FE App
PERSONALITY
Extrovert · Ambitious · Moderately tech-savvy
06
FROM FINDINGS TO FIXES
Four problems. Four design decisions.
Each pattern from the field and the sample study turned into a specific change in the flow.
Broken flow
The flow was fragmented and unintuitive, making it hard for field executives to move through it efficiently.
Guided pickup flow
A simplified reverse pickup flow with clear, step-by-step instructions at every stage.
Time inefficiency
An average pickup took about 6 minutes, significantly longer than a typical delivery.
Fewer steps
Unnecessary steps removed wherever applicable, so every tap moves the pickup forward.
Poor image quality
Photos were often of unopened packages, especially in the apparel category.
Instructed image capture
Clear guidance on how to take each required image, right inside the camera.
Too few images
Field executives were capturing just a single image.
Minimum three images
Front, back and tag images captured for every pickup, depending on category.
07
THE SOLUTION
The new pickup flow
Every solution from our field and desk research made it into the design.
STEP 01 · PRODUCT SUMMARY
See the original before you pick up the return
Reference images, upfront
The listing’s other photos sit as thumbnails beside the main image, so riders can compare angles at a glance.
Return reason on top
Pickup checks change with the return reason, so it leads the summary and sets up a more thoughtful pickup.
STEP 02 · INSTRUCTION
Unbox it first. Then shoot.
Before the camera opens, a visual prompt asks the rider to remove any packaging or flyer. The camera should see the product, not its polybag.
Show, don’t tell
A photo of the action does the explaining, so the instruction lands even for riders who skim text.
STEP 03 · GUIDED CAPTURE
The camera tells them what to shoot
On-screen prompts walk the rider through the Front, Back and Tag images in sequence, with a reminder to remove packaging on every shot.
STEP 04 · MANDATORY IMAGES
Three angles. No shortcuts.
Capturing three images is mandatory, with views like Front, Back and Tag depending on the product category.
Progress lives on the button
From “Capture Front Image (1/3)” to “Proceed to QC Questions”, the rider always knows what’s left.
STEP 05 · QUALITY CHECK
Every question, on one screen
We moved from a linear, one-question-per-screen QC to a single page that shows every question to be answered. Riders see the whole checklist before they commit.
STEP 05 · QUICK VIEW
The original is always one tap away
The product summary stays accessible as a collapsible section throughout QC, so riders can recheck details without losing their place.
STEP 06 · RESULT
A verdict no one can misread
A full-width success or failure state with one clear next action. When QC fails, the reason is spelled out, so the rider knows to walk away.
08
RESULTS
More pickups. Fewer wrong ones.
Lower losses.
After launch, the redesigned flow drove a substantial increase in the pickup rate and brought pickup losses down.
FALSE-POSITIVE PICKUPS
Before
33%
After
14%
↑
16%
Increase in pickup rate
↓
13%
Drop in reverse pickup losses, worth 45.5 lakh
Thanks for reading
There’s more to the story. These are only the highlights.
