Back to work
Rebuilding the Rider Referral System for 70k+ Last Mile Delhivery Riders
Slab-based referral campaigns and a redesigned Refer & Earn experience that grew weekly leads by 49% and brought CAC down to ₹300.
SURFACES
Partner App · FE App
Rider apps for referrers, Dashboard for the Capacity building team
TEAM
1 PM · 4 Eng · Design & Marketing
SCOPE
Express & Local
Campaigns at State, City and Center level
Summary
!
Problem
Referral was the cheapest, best converting channel, yet with 70k+ MAU only nearly 3% of riders referred, and it brought in under 2% of all leads.
→
Approach
Changed the referral construct, the app UI and the campaign tooling together: slab-based campaigns in the PMT Dashboard, front and center visibility, easy sharing and transparent tracking.
↗
Outcome
Weekly leads grew 49%, referral’s share of new rider onboarding went from 9% to 15%, and CAC dropped to ₹300.
What was the problem?
~3%
of riders refer
70k+ MAU, avg ~2k referral leads a month
2 to 5%
conversion
referral lead to load taken, campaigns run manually
<2%
of all leads
10,835 of 7,35,419 leads in 6 months
Referral had higher conversion rates and lower acquisition costs than other digital channels, yet it was consistently hovering around 2,000 leads a month. Referral campaigns were not live for most riders, the app showed a flat ₹500 that was out of sync with the actual campaign, and the Capacity Building team had no way to configure slab-wise rewards.
Leads created by source, Aug to Jan
Referral is the smallest channel after Agency
Field Leads
270414
NULL
191808
Digital
127430
Organic
102644
Bulk
28967
Referral
10835
Agency
3321
Cost per load taking rider
Sep
Oct
Nov
Digital & Organic
₹17,811
₹2,234
₹2,768
Referral (System + Manual)
₹2,786
₹1,055
₹1,650
Research
We went through six months of lead data, the LMS tables and the existing referral flow, and spoke with 4 to 5 riders. The refer construct is generally well understood in the market. What broke was everything around it.
1
Referral was massively underreported
In 5 days of data only 30 leads had source as Referral, while 425 leads had a Referred by or Filled by value.
2
Stages weren’t actionable
In progress, Ecode initiated, Ecode generated, In review: riders couldn’t act on them, and milestones or deadlines weren’t shown.
3
Campaigns weren’t live
Only 8 referral campaigns were configured since November. January had no referral payouts. In December, 3 riders won an incentive.
We found some of the reasons why referral is low
01
UX with insufficient clarity on benefits for each partner
02
Flat referral payouts limit the ability to design attractive, business-viable incentive structures
03
Not scalable beyond riders (e.g., non-riders), missing a high-potential acquisition channel
04
Low transparency on referral status and payouts, leading to reduced trust
05
No dashboard control for CB team to configure slab-wise referral incentives
Who it had to work for
No comprehensive dashboard to manage their AOR’s referral program.
CB Manager
Stages are not understandable or actionable by the referred rider.
Referrers (Last Mile Agents & Field Executives)
Payout details are not mandatory, so some referrers miss out on rewards.
External Referrer
Team leads and ATLs were excluded from campaigns because of potential fraud.
Internal Referrers
Strategy
OUR HYPOTHESIS
Make referral visible, easy to share and transparent to track.
Visibility of referred leads and their progress on the refer page will lead to improved conversion rates. Milestones and their rewards will push referrers to call and ask their leads to pick loads or reach milestones quickly.
Front & Center Visibility
Easy Sharing
Transparent Tracking
Guided Communication
The solutions we proposed
→
One-tap referrals via WhatsApp and SMS
→
Clear visibility of total referrals and total earnings
→
Simple, guided communication across the referral journey
→
Faster reward withdrawal with clear payout visibility
→
Centralized referral dashboard to manage campaigns without tech dependency
→
Stronger outreach via WhatsApp, print, and social channels
Decisions
01
Front & Center Visibility
PROBLEM WE SOLVED
Referral sat out of sight, and the app showed a flat ₹500 that didn’t match the live campaign.
WHAT WE DESIGNED
•
A customizable banner on the home screen showcasing earning potential, synced to the live campaign
•
An empty state that invites riders to start earning, with the referral code and Invite always in reach
•
Earnings and total referrals at a glance, with a nudge when friends haven’t completed targets
02
Easy Sharing
PROBLEM WE SOLVED
Referral was not scalable beyond riders (e.g., non-riders), missing a high-potential acquisition channel.
WHAT WE DESIGNED
•
One-tap invite on WhatsApp with a prefilled message
•
A static QR code for flyers or posters
•
“Fill for Others” to sign up new riders directly
03
Transparent Tracking
PROBLEM WE SOLVED
Low transparency on referral status and payouts was leading to reduced trust.
WHAT WE DESIGNED
•
Referrals split into Joined and Not Joined, with Not Joined categorized by status (In Progress, ID Created, Load Taken, etc.)
•
One-tap WhatsApp reminder to nudge referrals to complete pending steps
•
See History to track referral and transaction details
Status Tracking: every state tells the referrer what to do next
04
Guided Communication
PROBLEM WE SOLVED
UX had insufficient clarity on benefits for each partner.
WHAT WE DESIGNED
•
A How It Works page that explains how users can refer others in three simple steps
•
Clear communication of how much users can earn through referrals
•
A dedicated FAQ section and T&C for common questions, plus an explainer video
05
PMT, Template and Campaign Creation
PROBLEM WE SOLVED
The Capacity Building team had no dashboard control to configure slab-wise referral incentives.
WHAT WE DESIGNED
•
Campaigns at State, City or DC level, flowing from Campaign to Slabs of Referral Criteria and Reward
•
Templates Type A to D with 1 to 5 slabs each, for Express and Local
•
Bulk Upload for multiple states, cities and DCs, with a default campaign always live
Overall Impact
20%
↗
of weekly fresh onboardings
Referral’s share of new rider onboarding was 9% before launch.
60%
↗
conversion, lead to Load Taken
The highest channel on the basis of conversion.
₹300
↘
new CAC
85% less than performance marketing and 40% less than FOS acquisition.
+49%
weekly leads
1,393 → 2,087 average
2,729
peak week
95% above pre-launch
45%
next target
4,500 new riders a week by end of November
Referral has higher conversion rates and lower acquisition costs than other digital channels.
Challenges & Learnings
01
Fixing attribution before design
Referral as a source wasn’t recorded for all referred leads. LMS attribution fixes for direct and indirect referrals had to land first, or none of the new screens could be trusted.
02
Opening referral without inviting fraud
TLs and ATLs were excluded because of fraud risk. They can now refer, with a check on the last load taken in the last 2 months for existing riders.
03
Paying people outside the system
External referrers shared UPI IDs, but payouts ran on bank accounts through Oracle. That needed its own payout process flow.
