Is my RTO rate normal — and what is it actually costing me?

A normal RTO rate for an Indian D2C brand in 2026 sits between 20% and 25% — with COD orders failing at anywhere from 26% to 58% against under 2–15% for prepaid, and each failed delivery costing roughly ₹150–250 at a ₹1,000 order value. Which of those numbers applies to you depends on a denominator most benchmark pages never mention.

That denominator, the real cost of a single returned shipment, and the point at which cash on delivery stops paying for itself — that's what this piece works through, with the calculations shown.

What's a normal RTO rate for an Indian D2C brand right now?#

Between 20% and 25% of orders is the published industry average, spiking to 40% in high-risk categories and pin codes. The largest transaction dataset — Unicommerce, 410 million shipments across 6,000+ brands — measured 39.2% at the festive peak in November 2025, falling to 21.0% by March for brands that fixed their funnels.

RTO — return to origin — is the COD-era failure mode: the order ships, the customer never takes delivery, and the parcel comes home at your expense. Here is what the three platforms that actually see the shipments publish:

Read the last column before comparing yourself to anything. GoKwik's "20–25% average" is a rate over COD orders only — its own page publishes the formula. Unicommerce's 39.2%→21.0% series runs over all shipments. A brand computing RTO over total orders and checking itself against a COD-only benchmark will conclude it's healthy while it bleeds. And the prepaid figures — under 2% versus under 15% — are a 7x spread that denominators alone don't explain; the honest reading is that prepaid RTO is somewhere well under 15% and dramatically below COD, and nobody has published a tighter number.

By category, fashion and apparel run highest at 25–40%. GoKwik's pre-festive 2025 data showed electronics at 27% (down from 31.5%) and footwear at 27.9% (down from 37%). Speed matters as much as category: orders delivered in one to two days show 22% RTO against 35% when delivery slips past five days.

Two things nobody publishes, anywhere: an RTO split by order-value band, and a city-tier RTO table (Unicommerce reports tier-2/3 cities drove 66% of incremental FY26 volume and says metro playbooks fail there — but prints no tier-wise rate). And one number in wide circulation — "Unicommerce: COD share dropped to 18%" — does not exist on the page it's attributed to. Profitbox360 checked.

Sources: Unicommerce India D2C Report 2026 (410M+ shipments, 6,000+ brands) · GoKwik on RTO · Shiprocket on RTO protection

What does one RTO actually cost me?#

About ₹158 per failed delivery on a ₹1,000 COD order, before ad spend — built from published freight and fee mechanics plus named assumptions. GoKwik's published figure is ₹200–250 on the same order value, without itemisation. The number most founders carry in their head — "the shipping charge" — is less than half the real stack.

Here is the stack, line by line. Where a component has no published rate, the assumption is stated — substitute your own panel numbers:

The line most founders get wrong is the zero. The COD handling fee — ₹36 or 2.5% of order value, whichever is higher, on standard aggregator terms — is not a cost of a failed delivery, because it's refunded when the shipment returns: Shiprocket's own passbook glossary states "We reverse the COD amount in case your shipment goes in RTO," while RTO freight is billed on top. Any cost model that stacks the COD fee onto a failed delivery double-counts.

And the stack above still excludes the biggest line for an ad-driven brand: the marketing money that bought the order. A brand paying ₹200 per order in ads loses ₹358 per RTO, not ₹158 — the parcel comes back; the ad spend doesn't.

Sources: Shiprocket COD charge mechanics · Shiprocket passbook glossary · GoKwik RTO cost formula

What does each percentage point of RTO cost me per month?#

For a brand doing 3,000 orders a month at ₹1,000 average order value, one percentage point of RTO is 30 failed deliveries — about ₹4,700 a month, ₹57,000 a year. At the "normal" 21% rate, that's roughly ₹1 lakh a month walking out the door while the benchmark says you're fine.

The working, on the same brand:

3,000 orders/month × 21% RTO = 630 failed deliveries 630 × ₹158 = ₹99,540 a month — about ₹12 lakh a year Each percentage point = 30 RTOs × ₹158 = ₹4,740/month

Now put that against what the delivered orders earn. At a 15% contribution margin — ₹150 per delivered ₹1,000 order — the 2,370 delivered orders contribute ₹3.55 lakh a month. The RTO bill of ₹99,540 consumes 28% of that contribution. And notice the two per-unit numbers sitting next to each other: one RTO costs ₹158; one delivered order earns ₹150. At these economics, every failed delivery cancels the profit of one successful one. A brand at 21% RTO needs its delivered orders to carry a fifth of the book as dead weight before rent, salaries, or ad spend see a rupee.

This is why the festive quarter flatters revenue and starves the bank account: Unicommerce clocked the market at 39.2% RTO in November. At 39%, our worked brand loses ₹1.85 lakh a month to failed deliveries against ₹2.75 lakh of contribution — two-thirds of gross profit, gone in reverse logistics. The calculation is the diagnosis: RTO isn't a logistics KPI, it's a margin line.

At what point does COD stop being worth it?#

When your COD RTO rate crosses roughly 30–35%, sustained. Below ~25%, COD's extra orders earn more than its failures cost and turning it off burns money. The break-even sits where margin lost on vanished orders equals RTO cost saved — and it lands in the high 20s to high 30s depending on your cost per RTO.

COD is both the problem and the revenue. The only measured switching figure comes from Unicommerce: a prepaid incentive converts 20–30% of COD intenders — meaning if COD disappeared, roughly a quarter of those buyers would pay online and the rest are gone. (The "removing COD drops conversion 20–40%" number circulating on vendor blogs has no measurer, no sample, and no source behind it; the equation below doesn't need it.)

So work it per 100 COD-intent orders, with margin per delivered order M, COD RTO rate r, cost per RTO C, and switch rate s:

Keep COD: (1−r) × 100 × M − r × 100 × C Kill COD: s × 100 × M (the switchers, now prepaid, where RTO is a rounding error) Break-even: r* = M(1−s) ÷ (M + C)

Both branches, on the worked brand (M = ₹150, s = 25%):

Lean cost stack, C = ₹158: r* = 112.5 ÷ 308 = 36.5% GoKwik's fuller cost, C = ₹250: r* = 112.5 ÷ 400 = 28%

The decision rule falls out cleanly. A brand whose COD orders fail at 20–26% — the GoKwik-measured norm — keeps COD: it's profitable, just expensive. A brand sustaining 35%+ is paying for the privilege of shipping parcels back to itself. And at the festive quarter's measured 58%, COD is deeply loss-making while it happens — which is why the answer is almost never a permanent switch-off. It's gating: prepaid-only during festive peaks, in the pin codes where your panel shows failure clustering, on the SKUs that attract impulse orders. The equation flips seasonally and by zone; your COD policy should too.

Source: Unicommerce India D2C Report 2026 (FAQ: prepaid incentives convert 20–30% of COD intenders)

What actually brings RTO down — without killing COD?#

Shift orders from COD to prepaid — it's the only lever with a real dataset behind it, and the arithmetic beats every delivery-side fix. Every other published lever number is a vendor's ceiling for its own product: real direction, unreliable size. Here they are, read honestly.

Two honest caps to put on all of it. First, a lever can't fix more than its cause: the only published root-cause split (Shiprocket, unsampled — use as a planning assumption) puts ghost customers at 35% of RTOs, bad addresses at 22%, expectation mismatch at 18%, changed minds at 12%. Address verification cannot touch the 78% of failures that aren't address failures. Second, every vendor number above describes its own paid product with no denominator — direction yes, magnitude no.

Now the prepaid arithmetic, which needs no vendor claim. Move 10 points of order share from COD to prepaid — via incentive, partial COD, or gating — and RTO probability on those orders drops from ~26% to ~2% on GoKwik's numbers:

10 points × (26% − 2%) = 2.4 points off total RTO 2.4 × ₹4,740 = ≈ ₹11,400/month saved on the worked brand — plus the cash arrives in days instead of the 15-day COD remittance cycle, and the incentive's discount cost comes out of a number that's now visible instead of a loss that wasn't.

And the reframe, because sometimes RTO is the symptom, not the disease. If you sell fashion at a low order value to impulse buyers — the 25–40% RTO category — with five-day delivery, no confirmation flow will rescue the economics; your baseline is structural. The equation from the last section says where the effort goes: raise prepaid share, raise AOV, cut delivery days (22% RTO at 1–2 days versus 35% past five). A brand that fixes those three doesn't manage RTO. It stops generating it.

Sources: Unicommerce India D2C Report 2026 · GoKwik × Pepe Jeans · GoKwik on Partial COD · Shiprocket Engage 360 · Shiprocket NDR management

---

FAQ#

Is a 20% RTO rate normal for an Indian D2C brand? Depends on the denominator. Over all orders, 20% sits at the market norm (20–25%); over COD orders only, it's better than GoKwik's measured ~26%. Normal isn't cheap though — at a ₹1,000 AOV and 3,000 orders a month, 20% RTO costs roughly ₹95,000 a month.

How do I calculate what an RTO costs me? Forward freight + reverse freight + packaging + ops time + expected damage + working-capital lock — about ₹158 on a ₹1,000 order — plus your ad cost per order. Don't add the COD handling fee: aggregators reverse it when a shipment goes RTO.

Should I turn off COD completely? Almost never permanently. Break-even is r* = margin(1−switch rate) ÷ (margin + RTO cost) — around 28–37% COD RTO at typical D2C economics. Below that, COD earns its keep; above it, gate COD by season, pin code, and SKU rather than killing it outright.

Why does my RTO rate look worse than the benchmark? Check the denominator first: GoKwik's benchmarks divide by COD orders only, Unicommerce's by all shipments — a COD-heavy brand comparing against the wrong base can be off by half. Then check season and category: festive quarters ran 39% market-wide, and fashion runs 25–40% against 20–25% overall.

---

Profitbox360 builds and runs growth systems for D2C and ecommerce brands — reading the numbers first, so the spend goes where the equation says it should.