How to actually analyse your COD return data (most brands are looking at the wrong things)

Shopify Discussion
, ,
1

Most Indian D2C brands I talk to track RTO rate as
a single number. 28% returns. Okay. But that number
hides everything useful.

Here’s a better way to cut the data:

By pincode: Pull your last 6 months of COD
orders. Map returns by pincode. I’d bet 60% of your
returns come from 15% of your pincodes. Those
pincodes are high-risk by definition — treat them
differently (OTP verification, or block COD there
entirely).

By order time: Look at when high-return orders
were placed. Anecdotally: orders placed 10PM-2AM
have significantly higher RTO rates. The customer
is impulse-buying and has lower intent by morning.

By order composition: COD + first-time buyer +
discount coupon in the same order = highest risk
combination. If you have this pattern in your data,
you’ll see it clearly once you filter for it.

By customer history: How many returns does each
customer have? Most brands don’t track per-customer
RTO rate. The merchants who do find that 20% of
customers cause 80% of returns.

None of this requires any tool. Just a CSV export
from Shopify and a pivot table.

What patterns have you noticed in your own data?
Curious if pincode clustering holds true for
everyone or if it’s category-specific.

2

Hey @LokeshSomaiya69

hope you’re doing well!

This is a much more useful way to look at RTO than one overall percentage. Pincode and customer history especially seem like strong signal for identifying repeat risk

3

This is a great breakdown - the pin code and first-time buyer + coupon + COD combo point especially. Most brands just watch the overall RTO % and never break it down, so they miss where the real risk is coming from.

One thing I’d add : it’s worth pairing this with a quick look at cart value too. In a lot of stores, low-AOV COD orders (especially with a coupon stacked on) get abandoned or returned more casually since there’s less ‘skin in the game’ for the buyer. So a pivot by pin code + order value + first-time buyer flag together usually narrows things down even faster than any one dimension alone.

Also agree that you don’t need fancy tools for this - a monthly CSV export and a pivot table gets you 90% of the way there. Where it gets tedious is doing this by hand every month, so some brands set up a simple dashboard (or use a COD verification/RTO app) once they’ve confirmed the pattern manually, just so they’re not repeating the analysis from scratch each time.