A sale is a promise. A return is a chance to check whether the product, the page and the delivery experience kept it.
When returns increase, the first reaction is often to change the listing or reduce advertising. Either might eventually be appropriate. Neither should be the starting assumption. A product that arrives damaged needs a different response from a product that arrives exactly as described but does not fit the customer's intended use.
For U.S. marketplace and Shopify teams, a useful returns review ends with an owner and a testable correction. It should not end with a spreadsheet labeled “customer issues.”
Start with the promise the customer saw
Choose one SKU with a recurring return reason. Capture the actual product page, selected variant, main images, dimensions, compatibility language and package contents. If the customer ordered weeks ago, today's page may not be the page they saw. Keep dated screenshots when you change important claims.
Ask a specific question: what could a reasonable shopper have expected from this information?
A close-up can make a small accessory appear substantial. A lifestyle image can imply that extra pieces are included. A broad compatibility statement can hide an important exception. None of these automatically proves that the page caused a return. They identify a claim worth checking against customer comments and the actual item.
Sort evidence into three working categories
Expectation gap: The item works, but its size, finish, quantity or intended use differs from what the buyer understood. Review copy, images, variant selection and instructions. The correction should clarify the purchase before checkout.
Product problem: The item fails, varies from its specification or repeatedly disappoints in the same way under its intended use. Inspect samples and batch information. A new image cannot repair a defective component.
Fulfillment problem: The wrong item arrives, parts are missing or packaging does not protect the shipment. Check picking, packing and the delivery evidence. Changing ad targeting is unlikely to fix a recurring packing error.
Allow an “unknown” category. A short reason code is not enough to confidently assign a cause, and one return can involve more than one problem. Use available customer comments and inspection notes to refine the classification. Keep personal customer information out of general marketing documents.
Match the measurement to the question
Shopify's ecommerce reporting guidance identifies SKU-level return patterns and return reasons as useful inputs for connecting post-purchase problems to content, sizing and quality. That is a useful starting point, not a diagnosis of an individual product.
For an operational workload view, count returns processed during a period. For evaluating whether a page change helped, try to connect returned units to their original order group. Those are different questions.
Do not divide this week's returns from older purchases by this week's new sales and assume the result measures the quality of this week's product page. Record the definition, the sales window and how long customers have had to return the product. Use comparable observation windows before and after a change. If the data cannot support that comparison, state the limitation.
Keep counts beside percentages. Two returns out of twenty units deserve investigation, but they do not establish the same pattern as two hundred returns out of two thousand. Also distinguish returned units, refunded orders and partial refunds. Combining them casually can obscure what actually happened.
A hypothetical diagnosis
Imagine a storage accessory with repeated comments that it is “too small.” The team initially proposes a new hero image.
A review finds three different situations. Some shoppers expected a larger internal opening because the page showed only exterior dimensions. Others selected the wrong size variant. A small number received the wrong SKU.
These findings suggest three separate corrections: show usable interior dimensions, make size selection easier to understand and investigate the picking error. Replacing every image at once would make it harder to learn which correction addressed which problem.
Start with the clearest supported issue. Save the original page, document the exact change and note the effective date. Check that the new information is visible on mobile and on the selected variant. Keep other major changes stable where practical, and record promotions or traffic shifts that could affect the comparison.
This is an illustration of a decision process, not a client result or a promise that returns will fall.
Give each finding a destination
A weekly review can stay small. Select a recurring issue, attach the evidence, assign the person responsible and set a review date that allows enough orders and returns to accumulate.
Content findings go to the listing owner. Repeated defects go to the product or supplier team. Packing errors go to operations. Unknown cases need further evidence rather than a confident explanation invented to fill a report.
Evaluate the correction using more than conversion alone. A clearer page may discourage an unsuitable purchase while helping the right customer buy confidently. Review orders, return reasons, customer questions and contribution after relevant costs together before deciding whether the change helped.
DMM connects product content, advertising and account operations so each problem reaches the right part of the business. Explore our marketplace support approach, or book a discovery call to discuss your current priorities. Follow Digital Marketing Management on LinkedIn for practical marketplace reviews.
Which recurring return reason would you investigate first: unclear expectations, product performance or delivery execution?
