Returns are not a shipping line. They are specific customers.
See who is driving your returns, what each one really costs, and which high-returners are still worth keeping.
Returns
168,905
refunded, 14.3% of sales
23%
of customers have ever returned anything
9%
of customers are behind half of all refunded dollars
9.1%
of first orders came back in full
Who the refunded dollars come from
9% of customers, half of all refunded dollars
the next 14%, the other half
77%, never returned a dollar
Half of the refunded dollars come from 9% of customers. A store-wide return rate would never show it.
The industry number cannot hold still. Yours can.
Returns fraud is the most talked-about line in retail and the least measured. The headline figure moves with the method that made it. The only returns number you can act on is your own, read customer by customer.
$103B
Retail’s most-cited returns-fraud figure for 2024, published twice in the same year from bases $205 billion apart.
93%
of retailers call returns fraud and abuse a significant problem, in the same report whose headline rate fell.
47%
of merchants who added return fees saw complaints rise. 37% lost customers, and 34% saw order value fall.
NRF and Happy Returns, 2025 Retail Returns Landscape; Appriss Retail and Deloitte, 2024. The whole series, year by year, is in the returns-fraud number cannot hold still.
Hard fraud
A crime. Block it at the payment.
- Empty box, or the wrong item returned
- Overstated quantity on a claim
- Counterfeit swapped for the original
- Never-arrived claims on delivered orders
Belongs to your payments and fraud stack, which sees the transaction. Tacet never blocks a return and never flags a person.
Soft abuse
A paying customer exploiting a lenient policy.
- Wardrobing: worn once, returned
- Bracketing: many sizes, low keep rate
- Serial returning, order after order
- Chargebacks filed instead of returns
Invisible to fraud tools, because every event is a legitimate customer making a legitimate return. The pattern exists only at the customer level, across orders. That is where Tacet reads.
Who is driving your returns.
Two reads on every customer, side by side: the share of their revenue that has already come back, and the share expected to come back from here. Forward CLV deducts the second, so a loyal high-returner and a costly one stop looking alike.
Ask Tacet
In Claude
Who is driving my returns?
Read Returns concentration · scored 23 Oct
23% of your customers have ever returned anything. 9% of them account for half of all refunded dollars, and 77% have never returned a dollar. The five with the highest expected returns are listed; two of them hold more than $140 of Forward CLV each and are worth keeping.
Customers
| Customer | Orders | Refunded to date | Expected returns | Forward CLV |
|---|---|---|---|---|
| C-48112 | 7 | 61% | 54% | $6 |
| C-31907 | 5 | 48% | 45% | $19 |
| C-09344 | 9 | 44% | 31% | $212 |
| C-27761 | 3 | 42% | 40% | $8 |
| C-15520 | 6 | 39% | 28% | $147 |
Refunded to date is observed: refunds over revenue, so far. Expected returns is the share of future revenue expected to come back, re-scored every night and blank for a one-time buyer. Forward CLV deducts it.
Two of the five are worth keeping anyway, at more than $100 of Forward CLV each. A return rate would have taxed all five.
See who is driving your returns, free, by tomorrow morning.
What a return costs, after everything.
The refund is the small part. Tell Tacet what one return costs you to handle, and every return is priced that way: the handling, and the margin on what was refunded, landing on the customer who sent it. Priced that way, two customers with the same revenue are not the same customer.
Two customers, one revenue line
Worked example
Twelve months, a 50% gross margin and $18.00 of handling for each returned unit.
| Twelve months | Customer A | Customer B |
|---|---|---|
| Orders | 3 | 4 |
| Units ordered | 8 | 14 |
| Units returned | 1 | 9 |
| Gross revenue | $560.00 | $560.00 |
| Refunds | $70.00 | $360.00 |
| Return handling | $18.00 | $162.00 |
| Contribution after returns | +$227.00 | −$62.00 |
Customer B brackets: three or four items in every cart, two or three sent back each time. On a revenue leaderboard, A and B are the same customer.
Refunds count when they are issued.
On any order, in the period the money went back, and restated when a late one arrives. Never frozen at the number you first saw.
Exchanges are not refunds.
Shopify writes an exchange as a $0 order. 412 of them this period, and none reached the refunds line.
First orders that came back are followed.
9.1% of first orders came back whole. Refunded, exchanged or kept, each first order is followed to see whether the customer came back at 90 and 180 days, and what they are worth now.
Loss prevention without a blanket fee.
A blanket return fee taxes the loyal high-returners you most need to keep, and the customers abusing the policy barely notice it. Score the customer, then aim the lever at the segment.
Positive net, high returns
Protect. Free, easy returns.
The loyal high-returner. A generous return experience is what keeps them.
Negative net, low keep rate
Add friction: a fee, final-sale nudges, size guidance.
The bracketer. Recovers margin from the customers driving two-way shipping.
Positive net, low returns
Grow share of wallet. Cross-sell, replenish.
The cheapest customers to serve, with the most headroom.
Negative net, discount-acquired
Stop re-acquiring them on paid.
You lose money each time you buy them back.
How Tacet aims it.
Tacet does not run your returns desk. It gives you the customer-level read a returns platform cannot, puts it where your team already works, and reads every playbook in profit after refunds, so a send that reached the bracketers cannot look like it worked.
The segment, by name.
Ask Tacet for the customers with the highest expected returns, with Forward CLV beside each and emails when you need them. Take that list to your returns desk and aim the policy there.
A return band on every Klaviyo profile.
Two properties Tacet writes to your Klaviyo profiles: whether the customer has ever sent something back, and a band, light, moderate or heavy, on the ones who have. Your own segments can route a heavy returner to size guidance instead of a sale.
The Serial Returner, named.
One of the customer archetypes in your overview. Gross order value looks strong, but much of it never becomes Forward CLV. Tacet shows what these customers hold and what they refunded, never a lift from cutting them off.
Recovery that knows a swap from an exit.
After a return, only the customers who wanted the product hear from you again. An exit rests. See post-return recovery.
For the deeper read, the returns audit.
Send a returns export from Loop, Redo or Shopify and Tacet reads a fixed window of it once, behaviour by behaviour, ranked by what each cost you. Behaviour is described. A person is never flagged.
Returns across separate purchases, the pattern a single refund hides
Failed exchange chains on one purchase, a fit failure rather than a customer problem
Strict size bracketing: the same item in several sizes, most sent back
Late returns, well past the window
Repeat returners who always take the refund, never the exchange
The loyal high-returners, set apart from all of the above
What it reads, and what it never does.
Reads refunds, orders and exchanges from Shopify.
And the return reason where Shopify carries one: too small, too large, no longer wanted, not as described, defective. Read-only, every night.
Never blocks a return or flags a person.
The read is a number on a customer, for you to act on. Tacet writes nothing to your store and nothing to your returns desk.
Never a figure from an industry report.
Every number in the returns view is your own, from your own refunds, with the date it was scored.
Know which returns are worth it, and which customers are not.
Your customer P&L, free, the morning after you connect Shopify.