Use cases/ Retail & E-Commerce/Return reason data is too vague to act on
Retail & E-Commerce

Return reason data is too vague to act on

The reason field says "didn't meet expectations" every time.

What it costs you today

You are running a 12% return rate on a specific product line. Your return form gives customers a dropdown and an optional text box. The dropdown is 80% "defective" or "other." You cannot tell if it is a product problem, a content problem, or a packaging problem.

How RadRoot is used here

RadRoot sends returning customers a five-question adaptive interview the moment a return is initiated. Questions adapt to what the customer says — "not what I expected" leads to "what did you expect based on the listing" and "what was different."

The file you would upload

A Shopify - Returns & Exchanges Report export. One row is one product return request. RadRoot reads the whole row, so every question can name something the person already knows is true.

RMA NumberOrder NumberCustomer NameCustomer EmailProduct SKUProduct NameReturn Date
RMA-2026-0412SO-88341Karen Wukaren.wu@silverpine.example.comHPL-2847-GRYHampton Throw Pillow Set (Gray)2026-01-14
RMA-2026-0419SO-88556Michael Torresm.torres@greenfield.example.orgHPL-2847-BLUHampton Throw Pillow Set (Blue)2026-01-16
RMA-2026-0423SO-88602Lisa Patellisa.patel@techcorp.example.netVAS-1129-WHTVintner Storage Basket (White)2026-01-17
RMA-2026-0431SO-88774James O'Brienjames.obrien@consulting.example.comHPL-2847-GRYHampton Throw Pillow Set (Gray)2026-01-19
RMA-2026-0438SO-88821Sarah Chens.chen@bayarea.example.orgCRL-5521-NAVCarmel Knit Throw Blanket (Navy)2026-01-21

First 5 of 20 rows, first 7 of 14 columns. RadRoot interviewed the customer on each row, and 7 of 20 answered.

One interview, exactly as it ran

Six options every time, including the ones that reflect badly on whoever sent it. Each question was written after the answer above it landed.

Q1You are returning the Hampton Throw Pillow Set in Gray. What happened?
Q2Were the pillows themselves damaged, or just the shipping box?
Q3Do you know which delivery service brought your package?
Q4Before you noticed the damage, did the color look accurate compared to the website?
Q5If the product had arrived undamaged, would the size have matched what you expected?

The verdict it produced

Written for whoever ran the campaign. The respondent never sees it.

Rachel returned the Hampton pillows in Gray because the box arrived completely crushed on one side and one of the pillows inside was flattened with lumpy stuffing. FedEx Ground made the delivery to ZIP 94104 in San Francisco. The gray color looked accurate and the size did not appear obviously wrong. This is the fourth return out of seven where FedEx Ground delivered to a 941xx ZIP with significant packaging damage. That concentration is now definitive and requires immediate follow-up with FedEx about their San Francisco ground routes.

photography gap mixed low confidence

The interview did not gather strong evidence about whether the product photos matched reality.

scale perception mixed low confidence

Rachel indicated the size did not look obviously wrong, but she could not fully evaluate it due to the damage.

packaging damage negative high confidence

The box was crushed on one side and one pillow inside was flattened with lumpy stuffing.

carrier concentration negative high confidence

FedEx Ground delivered a crushed box to ZIP 94104 in San Francisco.

color accuracy positive medium confidence

The gray color appeared accurate to the customer before she focused on the damage.

See this on your own data

Fifteen minutes, screen to screen. We run your real question through RadRoot while you watch, and you leave with the follow-up it wrote and a price.

Book 15 minutes Answer one yourself