Where to find customer feedback beyond reviews: tickets, returns and questions
Beyond reviews, customer feedback sits in five places most stores already have: support tickets and emails, return and refund reasons, pre-purchase questions, post-purchase surveys, and marketplace reports such as Amazon’s Voice of the Customer. These sources are larger and more specific than reviews, because they come from customers who wrote to you instead of writing in public.
What this comes down to
- Most customers never leave a review, so reviews alone are a small and lopsided sample of what people think.
- Return reasons and support tickets explain a problem in the customer’s words, at the moment it happened.
- Questions asked before buying show what your product page fails to say, which reviews rarely reveal.
- Read every source with the same set of topics, so a complaint in a ticket and one in a review count together.
Why reviews are not enough
Reviews feel like the voice of the customer, but they come from a small group. Research on online ratings by Hu, Zhang and Pavlou found that review scores tend to pile up at five stars with a smaller cluster at one, and few in between. They put it down to two biases: people who buy are already inclined to like the product, and people with moderate views mostly do not bother to write.
The customer who got a slightly wrong size, exchanged it and moved on rarely writes a review. They did leave a trace, though: a return reason, an email, maybe a question before they bought. Those traces are where the quiet majority shows up.
Reviews are what customers say in public. Tickets are what they say to you.
Five places feedback is already waiting
| Source | What it tells you | Where to look |
|---|---|---|
| Support tickets and emails | What went wrong, in detail, and how urgent it felt | Your helpdesk or shared inbox; most helpdesks can export tickets |
| Return and refund reasons | Fit, quality and description problems, tied to a product | Your store’s returns records, if a reason or comment is captured |
| Pre-purchase questions | What the product page leaves unclear | Chat logs, contact forms, marketplace question sections |
| Post-purchase surveys | Answers to the question you chose to ask | Whatever survey tool sends them; export the free-text answers |
| Marketplace reports | Defect and return comments per product | On Amazon, the Voice of the Customer download |
Returns: the most honest feedback you own
A return costs the customer effort, so the reason they give is usually real. If your returns process asks for a reason, you already have a labelled dataset: every return tied to a product, with a category of cause and often a comment.
If it does not ask, that is the cheapest fix on this page. Add one required choice and one optional comment box to the return form. In a month you will know whether a product comes back for size, for quality, or because the photos oversold it.
On Amazon, sellers cannot download product review text, but Amazon names Voice of the Customer as an alternative, and its download carries the defect and return comments per product. The Amazon page has the details.
Questions before buying: the gaps in your product page
Every question a shopper asks before buying is something your product page did not answer. Does it fit a 15-inch laptop?
Is the lid dishwasher safe?
If the same question comes in three times a week, the answer belongs in the description, and probably in the feed as well.
Reviews rarely surface this, because people who could not find the answer often did not buy. Questions are the only record of the customers you lost before checkout.
Reading them together
Five sources are only useful if they are read with the same lens. A complaint about a zip in a review, a return marked defective and a ticket about the zip breaking are one problem, not three. Sort all of them into the same short list of topics and count across sources.
- Export a month of each source you have. Remove names, emails and order numbers.
- Keep only the free text and the product it refers to.
- Tag each item with a topic and a severity, using one list of topics for every source.
- Rank topics by count across sources, not within one.
The Review Analyzer is built for reviews, but a return comment or a ticket summary is the same kind of short customer text, and you can paste up to 30 at a time free without an account. The needs-a-reply flag matters less for those; the topics and severity carry over.
Questions
What are the main sources of customer feedback for an online store?
Product reviews, support tickets and emails, return and refund reasons, questions asked before buying, post-purchase surveys, and marketplace reports such as Amazon’s Voice of the Customer.
Why do so few customers leave reviews?
Research on online ratings points to under-reporting bias: customers with moderate experiences rarely bother to write, so reviews over-represent the very happy and the very unhappy.
Can I analyse support tickets the same way as reviews?
Yes. Strip personal details, keep the free text and the product it refers to, and tag it with the same topics you use for reviews, so problems can be counted across every source.
Read next
Who actually writes product reviews, and why it skews your average
Reviews come mostly from delighted and disappointed customers, not the quiet middle. What the J-shaped rating pattern means for how you read your reviews.
What is customer review analysis? What it finds that star ratings hide
Review analysis turns reviews into counts: topics, sentiment, severity and which need a reply. What it produces, and what a star average cannot tell you.
Whose data is in your reviews? What to strip before you analyze them
Review exports carry names, emails and order numbers the analysis never needs. What to remove, what to ask any tool, and how SellerSort handles your text.