The 90-Day
RTO Reduction Playbook
Go from diagnosis to a 25% cut in RTO in 90 days. Inside, you will find three sequenced phases, and how to measure progress.
.png)
Get the full report
India's D2C report on 51.1 million shipments and the categories that rewrote the cart.
By submitting, you agree to receive communications from ClickPost. No spam, ever.
D2C brands have already run this play.
Every intervention in this playbook is drawn from live ClickPost deployments.
Wellbeing Nutrition deployed AI voice agents to recover failed deliveries with automated engagement, feedback capture and action.
Pilgrim brought its absolute RTO rate down three full points through carrier performance accuracy with AI carrier allocation.
Miraggio reduced address verification escalations to their support team by implementing order editing on thank you page.
.svg)
The benchmark, the true cost, and why most efforts fail.
.webp?width=48&height=48&name=New-Age%20Fashion%20(1).webp){kind=small}
.webp?width=48&height=48&name=Skincare%20%20(1).webp){kind=small}
.webp?width=48&height=48&name=Wellness%20(1).webp){kind=small}
ClickPost FY2026 network data, COD orders, brands with 1,000+ COD shipments
RTO range: Rs 285- 490 depending on the order. The two costs teams forget are CAC you paid to acquire a customer who never converted, and stock blocked in transit.
Every 1 percentage point of RTO reduction = 8,333 fewer RTO events annually = ₹29.2 lakh in recovered margin.
Seven distinct upstream failures show up as one number. Without attribution, teams chase symptoms.
Checkout owns conversion, logistics owns cost, CX owns response time. RTO sits in the seam, owned by no one.
Without a weekly scorecard by carrier, pincode tier, payment mode and category, gains are anecdotal and reversible. Phase 1 fixes this first.
Diagnose
Where, specifically, are our RTOs coming from, by root cause, by carrier, by pincode tier, by payment mode, and by product category?
Two weeks: one dataset, seven buckets, four cuts, one baseline scorecard. By Week 2 you know, to the rupee, what your top three RTO drivers cost.
Pull 90 days of shipment data
Pull these fields for every shipment from the last 90 days. On ClickPost it's one export.
Apply the 7-bucket RTO attribution framework
Every RTO maps to exactly one root cause, the one that first failed in the chain. A wrong address that cascaded into a COD refusal is a Bucket 2 (address) event, not Bucket 1.
Map each carrier's NDR codes to standard buckets. Delhivery, Blue Dart, XpressBees, Ekart and Shadowfax all label things differently, so normalize before you classify.
Check 100 RTOs against IVR recordings and customer comms. Expect to reclassify 15 to 25% on the first pass. That calibrates the whole dataset.
Segment the data four ways
Each cut surfaces a different lever.
Split COD from prepaid. COD usually runs about 7× prepaid. Under 5×, look for fake prepaid orders. Over 10×, your COD verification is broken.
Score every carrier across metro, T1, T2 and T3. The gap between your best and worst carrier in T2/T3 is usually 20 to 25%, the biggest Phase 3 lever.
RTO is lowest at mid-cart values and higher at both ends. Cheap orders customers don't bother refusing; pricey COD orders carry more risk.
First-time buyers RTO far more than loyal ones, dropping from ~29% on order one to ~22% by orders 6 to 10. The clearest reason to push first orders to prepaid.
Identify the top three loss drivers
Rank the seven buckets by absolute ₹ loss over 90 days (₹350/event × count). For a typical ₹100 Cr GMV fashion brand:
Your numbers will differ. The exercise is the point.
Build the baseline scorecard
A single page, refreshed weekly for 13 weeks. It contains:
.png)
Not on ClickPost? Phase 1 takes 3 to 4 weeks instead of 2, mostly carrier reason-code normalization across Delhivery, Blue Dart, XpressBees, Ekart, and Shadowfax.
Teams trust carrier-reported NDR codes. Don't. The 100-order audit catches 15–25% misclassification that distorts every downstream decision.
A 28% overall RTO tells you nothing actionable. It can mean uniform failure or two carriers failing catastrophically in T2. The diagnosis is in the cuts, not the average.
Two weeks is the budget. If data isn't perfect, classify with what you have and refine in Phase 2. Velocity matters more than precision at this stage.
Quick Wins
Which levers can be deployed in under four weeks with no carrier renegotiation, no warehouse changes, and no integrations beyond our existing stack?
Four workstreams, four weeks, all configuration-level. Combined: 8–10% points, about 40% of your 25% target.
Each gate kills a specific bucket before the parcel can fail delivery.
COD verification at checkout
A checkout prompt offering ₹20 to 50 off to switch. Each COD order carries ~₹122 of expected RTO cost, so ₹30 to 50 to convert nets ₹100+.
Before dispatch, Parth calls to confirm the order and offers a prepaid switch with a payment link on the same call.
Above a set threshold, collect 20 to 30% online and the balance at delivery. Skin in the game slashes refusal-at-door.
Set the threshold at the AOV band where COD RTO crosses 40%. Expect a 4–7% checkout-completion dip, offset by a 60–70% drop in RTO on completed orders. Net economics strongly positive.
Post checkout address verification
For orders flagged as risky, Parth calls to confirm the delivery address before dispatch and pushes corrections back to the OMS
Identify the top 30–50 pincodes by absolute RTO volume. Build a risk score from Phase 1:
Refresh scores monthly.
For vague-but-valid addresses, add landmark auto-complete. Require a nearest-landmark field in T2/T3 and let repeat buyers one-tap their last good address.
Automated NDR resolution
Network NDR-to-RTO conversion is 48% and climbs with every failed attempt. ClickPost's top quartile keeps it under 25%.
The first 2 to 4 hours after an NDR are the best chance to resolve it. Ingest in real time and route resolvable NDRs into automation within 30 minutes.
WhatsApp, then SMS if unread, carries the NDR reason and one-tap actions: re-attempt, update address, or switch to prepaid. Resolves 35–45% within 2 hours.
For single call and scalable resolutions, Parth calls the customer, and pushes redelivery instructions to the courier over API.
On confirmation, push the re-attempt to the carrier within 15 minutes with the updated slot and address. Track success by carrier, it feeds Phase 3.
When manually NDR calling doesn't scale, adopt Parth
Parth calls customers, giving a seamless experience and pushes the outcome straight to the courier over API.
Order editing & post-order address fix
Right after checkout, ClickPost validates the address against Google Maps and serviceability. Customers fix address, phone or pincode from a self-serve link; your team can too.
Let customers swap a size or variant, change quantity, or cancel cleanly before dispatch:
The consolidated scorecard to review before moving into Phase 3.
₹50 off on every COD order is a margin leak. Reserve the incentive for the orders that need it, repeat buyers in metros rarely do.
"Your delivery failed" without an actionable next step is noise. The single-tap action (re-attempt / update / switch to prepaid) drives the 35–45% resolution rate.
Let customers edit or cancel up to dispatch, then lock hard. If the window stays open after fulfillment starts, you get mid-pick changes and split shipments that create fresh RTOs.
A pincode high-risk three months ago may have a new carrier serving it well today. Monthly refresh is non-negotiable.
Systemic
What structural changes to our carrier mix, delivery promise, customer communication, and pre-dispatch intelligence drive the rest of the 25% target, and sustain it past Day 90?
Phase 2 was config switches; Phase 3 is the operating model. Together it delivers about 60% of the target (roughly -4 to -5 points)
AI Carrier Allocation
The single largest lever in the playbook
SVP Supply Chain and Procurement, Pilgrim
For every carrier and pincode pair, score RTO, FAD, TAT, NDR and damage over 90 days. Rank carriers per pincode by a weighted composite.
Premium/high-AOV: raise damage weight. Fast-fashion/replenishment: raise TAT weight.
Hard-code the top-2 carriers per pincode for your top 100 pincodes, with a fallback. No ML, yet typically worth 3 to 5 pp in two weeks. Cap each carrier's share and pull volume back if lane RTO spikes.
Switch on the AI engine for all pincodes. It weighs live capacity, per-pincode performance, order and customer attributes, and cost, returning a per-order carrier pick in milliseconds at AWB generation.
Scores refresh daily. A carrier slipping on a lane is deprioritized within days; overperformers gain share automatically. This is the loop PlatinumRx, Pilgrim and Avimee ran.
+ AWB generation
As volume shifts to your best carriers, renegotiate: capacity guarantees with the winners, performance-linked pricing across the board. Close it by Wk 10.
EDD optimization
Replace blanket ranges with a predicted date per pincode and carrier, using live TAT and cutoffs. Show a specific date ("Delivery by Tue, 9 Dec"), not a range.
When a shipment slips past its promised EDD by 24 hours, auto-send a WhatsApp with the revised date and a one-tap accept, reschedule or cancel.
For high-AOV or time-sensitive orders, offer a paid express tier at checkout. A customer paying ₹40 for next-day is 6 to 8× less likely to RTO.
WhatsApp/IVR pre-delivery communication windows
Confirm next-day availability. Four single-tap options: I'll be available, Reschedule, Alternate person, Prepaid pickup point.
For non-responders. AI/IVR calls in the customer's own language (regional-language support is mandatory for T2/T3), checks availability, captures a preferred slot, and sends the revised instruction to the carrier over API.
Final delivery-window WhatsApp with the carrier's actual slot (where available) or a time band, plus the delivery executive's contact number where policy allows.
The achievement criteria for the 25% relative RTO-reduction target.
The performance matrix needs weekly review for the first 8 weeks. New pincodes appear, networks shift, seasonal effects (monsoon, festive, regional events) change the rankings.
Wait for 6 weeks of allocation outcome data (end of Wk 12) before re-cutting major contracts. Earlier, you're negotiating from incomplete information.
95% on-time on a tighter EDD beats 75% on a wider one. If your predictor misses > 12% of the time, widen it by half a day. Trust and accuracy compound.
Shifting all volume to your single best carrier creates capacity risk and kills rate leverage. Cap each carrier's share per pincode and keep a live fallback.
The instinct under volume pressure is to cut non-critical messaging. Don't. Festive spikes are exactly when Bucket 3 balloons and pre-delivery confirmation has the highest marginal value.