Returns margin engine · free diagnostic
Your returns are eating your margin. Find out how much.
Nearly one in five online orders comes back — and each one costs far more than the refund. SmartBitReturn models the seven costs merchants never add up, scores your operation against your category, and hands you a ranked recovery plan in about ninety seconds.
No signup to run it · no card · nothing leaves your browser
US retail returns · accruing this year
$0
Retailers are on track to absorb $849.9 billion in returned merchandise this year — 15.8% of all retail sales, and 19.3% of everything sold online. Source: NRF / Happy Returns.
The margin leak scan
Eight inputs. Everything below recalculates live. Defaults are seeded from your category's published benchmarks — override any of them with your real numbers for a sharper read.
Annual margin lost to returns.
The keep-it threshold
Below a certain price, physically taking an item back destroys more value than writing it off. Amazon runs this calculation at scale and refunds without return under $75. Almost no independent merchant knows their own number. Here is yours.
Modelled on your logistics cost, margin and restock rate. Exclude hazardous, serialised, high-fraud and heavy goods before you switch it on — and cap it per customer per year.
What to fix, in order of money
Seven levers, each sized against your own inputs and ranked by annual recovery. The top three are almost always worth more than everything below them combined.
Take it with you
Get this scan as a shareable report
We'll send your full breakdown — leak ledger, keep-it threshold, ranked plan, category benchmarks — plus the implementation checklist for the top three levers. One email. No drip.
The scan is free. The fix is where we earn.
Most merchants can act on the plan above on their own. If you'd rather not spend a quarter figuring out sequencing, vendors and policy language, we've already done it.
The diagnostic on this page, unlimited runs, forever.
- Seven-line leak ledger
- Returns health score
- Category benchmarking
- Keep-it threshold
- Ranked recovery plan
- PDF export
The implementation manual behind every lever the engine ranks.
- Step-by-step build for all seven levers
- Returnless-refund policy templates + SKU exclusion rules
- Return-reason taxonomy that actually diagnoses
- Exchange-first flow copy that converts refunds
- Vendor comparison: returns platforms, 3PLs, fraud screening
- The spreadsheet model this engine runs on
- Free updates for 12 months
We run it on your real export data and hand back a board-ready plan.
- Full analysis of 12 months of order + return data
- SKU-level keep-it thresholds, not one blended number
- Return-reason clustering to find the true root causes
- Your worst 20 SKUs by margin destroyed
- 90-day sequenced rollout with owners
- 60-minute walkthrough call
Returns benchmarks by category
Published rates and modelled cost stacks the engine uses as defaults. Your own numbers always beat a benchmark — these exist so you have somewhere to start.
| Category | Return rate | Reverse ship | Processing | Support | Resold as new | Est. cost / return |
|---|
Return rates from NRF, Happy Returns and category studies. Cost stacks modelled from published reverse-logistics ranges of $10–$30 for parcel goods and $40–$120+ for bulky items. Cost per return assumes an $80 order at 55% margin.
How the number is built
No black box. Every figure on this page comes from arithmetic you can check, and the assumptions are stated rather than buried.
Eight inputs, category-seeded
Your category sets the reverse-shipping, processing and support cost stack. Everything else — volume, order value, margin, return rate, restock rate, who pays shipping, refund mix — is yours.
Seven costs, no double-counting
Reverse shipping, inbound processing, sunk outbound freight, unrecovered payment fees, recovery loss on units that can't go back to A-stock, support handling, and fraud. Fraud is modelled on 9% of returns as a total loss of goods; recovery loss is applied only to the remaining 91%, so nothing is counted twice.
Scored against your own category
The health score weights return rate against your category benchmark (40 points), leak as a share of gross profit (35) and A-stock recovery rate (25). Comparing a jewellery store to an apparel brand tells you nothing; comparing it to jewellery tells you everything.
Questions people ask
Why is your number so much bigger than my refund total?
Because the refund is the smallest part. When a $78 order comes back you don't just return $78 — you've already paid outbound freight, you pay reverse freight, someone opens, inspects, grades and restocks it, your processor usually keeps its percentage on the original sale, a share of units never make it back to full price, and support spends time on the ticket. The engine adds all of that up. Merchants routinely find the true cost is two to four times the sticker.
Do I need to connect my store or upload data?
No. The scan runs entirely in your browser — nothing is uploaded, and nothing is stored on a server. That's a deliberate constraint: you should be able to get a real answer before you trust anyone with your order data. The paid teardown is where real exports come in, and that's opt-in.
Is a returnless refund really cheaper than getting the item back?
Frequently, yes. If reverse freight plus handling costs $16 and the unit's recoverable value is $11, insisting on the return burns $5 and adds a week of friction. The threshold on this page is where those two lines cross for your economics. Amazon reached the same conclusion at $75 and now extends it to third-party sellers. The discipline is in the exclusions — cap it per customer, exclude high-value and high-fraud SKUs, and monitor abuse.
My return rate is above my category benchmark. How bad is that?
It depends entirely on what's driving it. A high rate caused by generous sizing bracketing on a brand with strong repeat purchase can be healthy. A high rate driven by product photography that oversells, sizing that runs inconsistent, or quality failures is margin on fire. The recovery plan sizes each of those paths separately so you can tell which one you're in.
Where do the benchmark numbers come from?
Return rates come from the NRF's returns research with Happy Returns and published category studies. Cost stacks are modelled from reported reverse-logistics ranges — roughly $10–$30 all-in for parcel goods, $40–$120+ for bulky and furniture. Fraud is the NRF's 9% figure. Sources are listed below and the model is included with the playbook if you want to audit or fork it.
Sources
NRF & Happy Returns, 2025 Retail Returns Landscape — $849.9B returned, 15.8% of retail sales, 19.3% of online sales, 9% of returns fraudulent.
NRF, Consumers expected to return nearly $850 billion in merchandise — merchant cost drivers and 2026 priorities.
Shopify, Returnless refunds: how they work for retailers — eligibility filters and the sub-$75 threshold.
Category return rates and reverse-logistics cost ranges compiled from published 2026 industry benchmarks.