← All posts

June 4, 2026 · Marc Hendricks

Failed Delivery Costs 2026: We Did the Math

A failed delivery costs $15 to $40 before churn. See the 2026 UPS, FedEx, and USPS fees, why addresses fail, and how Shopify merchants prevent it.

Failed Delivery Costs 2026: We Did the Math

A failed delivery costs a US merchant between $15 and $40 in direct expenses, and the meter starts before the package even comes back. UPS charges $25.25 to correct a bad address in transit (2026 rate guide), FedEx charges $25.50, and transportation alone on a failed attempt runs $17 to $18 per parcel (Loqate data, 2025). Then come the costs nobody invoices you for: a replacement label, fifteen minutes of support time, and a customer you probably lose. Around 70% of shoppers say they’re unlikely to order again after a delivery fails.

Most of these failures start at checkout, not in the truck. Address problems account for roughly 25 to 30% of failed deliveries, and they’re the one cause a merchant can prevent. Here is where the money goes, why addresses go bad, and what you can still do about it at each stage.

How much does a failed delivery cost?

The direct cost of one failed delivery is $15 to $40: the carrier’s address fee or return shipping, plus the original label you already paid for. The true cost is higher once you add the replacement shipment, the support time spent on it, and the customers who don’t order again. The fees break down by carrier:

Fee (2026, US domestic)UPSFedExUSPS
Address correction in transit$25.25$25.50No per-package fee
Redelivery attemptUp to 3 attempts, no chargeUp to 3 attempts, no charge1 attempt, then held for pickup
Hold window before return~7 days5 business days15 days
Return to senderReturn shipping billed to you, original label not refundedReturn shipping plus the correction feeOriginal postage; some classes return free unopened
Intercept before deliveryDelivery Intercept $21, $15 small businessReroute $5.55–$33.50 by distance$19.45 Package Intercept

Run the math at a typical store’s scale. At 1,000 orders a month, if 2% of addresses have a problem, that’s 20 orders. Say half get corrected in transit at $25.25 and half come back as returns where you pay return shipping and reship at roughly $10 to $18 a label. You’re out $400 to $600 that month in carrier fees alone, before a single support ticket or lost customer. The scale of the problem nationally: USPS handled 4.37 billion undeliverable-as-addressed mailpieces in FY2023, which cost mailers $1.325 billion.

The churn is the expensive part of the math. The product, the label, and the fee are one-time losses. The 70% of customers who don’t come back after a failed delivery take their repeat orders with them.

Why do shipping addresses fail?

Shipping addresses fail in five recurring ways: typos, missing unit numbers, stale saved addresses, addresses that don’t exist, and valid addresses the carrier can’t serve. Most of these you can prevent at checkout. One of them you can’t, and it’s worth knowing which.

1. Typos. A transposed ZIP digit or a misspelled street sends the package to a sorting facility that can’t match it to a route. Carriers catch some of these with their own data, and every catch is a $25.25 correction fee on your invoice.

2. Missing apartment or unit numbers. The most common preventable failure. The street address is real, the building exists, but “482 Maple St” in a 40-unit building is not a deliverable address. Carriers won’t guess. UPS and FedEx raise a delivery exception, hold the package about 5 to 7 days waiting for the unit number, then return it, usually with the correction fee attached. There’s a subtlety here that matters for prevention: postal data distinguishes between “unit number required but missing” and “unit number provided but wrong.” The first is fixable with a prompt at checkout. The second means the customer typed something, so only verification against carrier data catches it.

3. Stale saved addresses. The customer moved and their wallet didn’t. Express checkouts (Apple Pay, Google Pay, Shop Pay) inject a saved address with one tap, which means the customer never re-reads what they’re shipping to. These addresses are perfectly formatted and point to where the customer used to live. There’s no typo to catch, just a stale record.

4. Addresses that don’t exist. New construction that isn’t in the carrier’s database yet, or a customer guessing at a vacation rental’s address. The reverse also happens: USPS data marks an address “not found” or “vacant” with a lag of two months or more, so a brand-new but real address can look fake. Good validation treats “not in the database” as a warning, not a hard stop.

5. Valid addresses the carrier can’t serve. A PO Box is a real address that UPS and FedEx physically cannot deliver to; only USPS can. Mailbox-store addresses (like a UPS Store) need the private mailbox number or the package arrives at a counter with no idea whose it is. Military APO/FPO addresses follow their own format and only ship via USPS. Each of these passes a “does this look like an address” check and still fails.

Why do packages get returned to sender?

Packages come back to you for a few specific reasons, and the carrier’s exception code tells you which:

For a merchant, every one of these codes means the original shipping cost is gone, the return freight is usually billed to you too, and the product sits in transit limbo for one to three weeks while the customer waits.

Who pays when a customer enters the wrong address?

You do, in practice. The carrier bills the shipper for corrections and returns, not the buyer, and Shopify doesn’t make the customer cover reshipment. Merchants on the Shopify forums describe the same arc over and over: the package comes back, the customer expects a free reshipment, and the merchant either eats the cost or tells the customer no and hopes they don’t dispute it.

None of the options are clean:

Whatever you decide, write it into your shipping policy page before a return lands. It’s easier to point a customer at an existing policy than to invent one in the middle of a support thread.

Can you fix an address after the order is placed?

Yes, if you catch it before fulfillment. In Shopify admin, open the order, find the customer section, and choose Edit shipping address. That’s free and takes thirty seconds. Customers can’t edit the address themselves after ordering, so this only happens if they email you in time, and “in time” is doing a lot of work in that sentence: stores that fulfill same-day have a window of hours.

Once the package ships, your options get expensive: USPS Package Intercept costs $19.45 and only works before the package is out for delivery. UPS Delivery Intercept runs $21 online ($15 for small business), and FedEx reroutes run $5.55 to $33.50 by distance. And once the carrier flags the address themselves, the correction fee has already landed.

The cost curve is steep: correcting an address at checkout costs nothing, doing it before the label prints costs thirty seconds, doing it in transit costs $19 to $33, and not doing it at all costs $15 to $40 and probably the customer.

How do you prevent bad addresses at checkout?

Address validation at checkout is the only fix that catches every order automatically, before anyone has to notice the problem.

Shopify’s built-in checkout helps but doesn’t finish the job. Autocomplete suggests addresses as the customer types in 20+ supported countries, and the admin flags suspect addresses on orders after they’re placed. But the flags are advisory: nothing stops a customer from submitting a missing unit number, and 55% of e-commerce checkouts still don’t auto-verify addresses at all (Baymard, 2024). Express checkout makes it harder still, since a stale saved address skips the address form entirely.

A validation app closes that gap by checking every order against carrier delivery data and acting on what it finds. This is what we built Address Verifier to do: it verifies each address down to the unit number, asks the customer to confirm a correction when one exists (a corrected ZIP, a standardized street), prompts for the apartment number when the building needs one, and blocks the genuinely undeliverable cases like a PO Box on a carrier that can’t serve it. Two design choices matter more than the feature list. It fails open, so if verification is ever unreachable your checkout keeps selling and the order is just flagged for review. And it’s deliberately conservative about blocking, because a validator that flags good addresses costs you conversions faster than bad addresses cost you fees. Pricing is $0.04 per order, first 100 orders free.

Be realistic about what validation can’t do: it can’t know your customer moved last month if their old address is still deliverable. Someone else lives there now, and that package is gone. Validation removes the large, boring majority of failures (the typos, the missing units, the impossible addresses) so those are the only ones left.

FAQ

What does “insufficient address” mean?

It means the carrier doesn’t have enough detail to deliver, almost always a missing apartment, suite, or unit number on a multi-unit building. USPS holds the package for pickup, while UPS and FedEx hold it about 5 to 7 days for a correction (with their $25-plus fee), then return it.

Can UPS or FedEx deliver to a PO Box?

No. PO Boxes sit inside post offices, and only USPS can deliver there. An order shipped to a PO Box via UPS or FedEx will be corrected to a street address if the carrier can find one (for a fee) or returned to you.

Can a customer change their shipping address after ordering?

Not by themselves. On Shopify, only the merchant can edit the address, and only realistically before fulfillment. After the package ships, the options are paid intercepts and reroutes through the carrier.

Does USPS charge an address correction fee like UPS and FedEx?

Not per package. USPS will forward, hold, or return mail under its own rules instead of charging a correction surcharge. The cost shows up differently: lost postage, return handling on some mail classes, and the $19.45 Package Intercept if you need to pull a shipment back actively.


Most of this is preventable at the one moment the customer is still looking at their address. Verify it there, and the $25 fees, the support threads, and the reshipment arguments largely never happen. See how Address Verifier works.