A SKU can have a high return rate and still create little financial exposure because it sells in low volume. Another SKU can have an ordinary rate but produce far more returned units and operational work because it is a bestseller.

Rank every SKU through three separate lenses:

• Observed exposure: attributable refunded merchandise value and any known shipping, label, or return fees.
• Returned-unit volume: the number of units that came back from the selected sale cohorts.
• Cohort return rate: returned units from the sale cohort divided by units sold in that same cohort.

SKU B deserves investigation, especially if it is new or strategically important. SKU A may still be the first commercial priority because improving it by only a few percentage points could prevent more returned units.

Start with observed values and add estimates only when their source is reliable. This keeps a useful prioritization model from pretending to be a perfect profit calculation.

• Level 1, merchandise exposure: attributable refunded merchandise value.
• Level 2, known return spend: add labels, carrier charges, platform fees, and other recorded costs.
• Level 3, handling estimate: add inspection, processing, and restocking labor only when the assumption is documented.
• Level 4, inventory loss: add non-restocked unit cost, markdown, or write-off only when reliable COGS and disposition data exist.

Keep observed and estimated columns separate. A merchant should be able to see which part came directly from records and which part came from an operating assumption.

A ranked list identifies where to look. Reasons, variants, and channels suggest what to change. Shopify exposes return reasons and product variant SKU at the time of sale, which makes more specific investigation possible when the underlying data is complete.

Generic reason lists can conceal the real cause. Where volume supports it, break a priority SKU down by size, color, category, channel, supplier, and fulfillment path.

• Start with the SKUs creating the most observed exposure.
• Remove products whose high total is explained entirely by enormous sales volume and an ordinary rate, unless a small improvement would still be valuable.
• Look for a rate that is meaningfully above comparable products or the SKU's own earlier mature cohorts.
• Prefer a SKU with a concentrated, actionable reason over one with scattered or missing reasons.
• Record one change, an owner, and the affected sale date.
• Re-measure the next mature cohort rather than judging the newest sales too early.

Work through the Shopify Return Leakage Checklist before making the decision, or review the cohort return-rate formula if the monthly numbers are not yet trustworthy.

Further reading: Shopify's ecommerce returns management guide discusses shipping, inspection, handling, restocking, and return reasons by SKU and category.