Day eleven, 9:14am, a one line email: hi, any update on my order? You open the order and everything looks finished. Fulfilled. Tracking number right there. The sale is in last week's numbers, the stock left your count a week and a half ago, and the carrier page, when you finally click through, has been saying the same thing since day three. Returning to sender.
Then somebody finds the box. It has been on the floor behind the printer for two days, which is roughly how long it has been back in the building, unopened, because no part of your store ever said it was coming.
The short answer
Shopify has no status that means a parcel is coming back to you. The 2026-10 Admin GraphQL FulfillmentEventStatus enum has eleven values and none of them is returning to sender, the value exists only in the unstable version of the API, and the order stays Fulfilled while the box sits in your own receiving area.
Flow has no trigger that fires on a carrier turnaround and no condition to test against, voiding the original label is almost always impossible because voiding requires a label under 30 days old with no tracking events, and the community reship workarounds all distort either revenue or inventory. Shopify also files a return to origin as a customer return, which is why return rates in RTS-heavy markets are mostly couriers rather than products.
The playbook reads the carrier feed hourly, flags the turnaround the hour it happens, names the cause, emails the customer for a corrected address before they ask, applies one reship or refund rule, restocks on physical arrival, and reports RTS separately from returns.
The word Shopify has not shipped
Here is the whole problem in one list. The
2026-10
Admin GraphQL FulfillmentEventStatus
enum, the thing that describes where a shipment is, has
eleven values: attempted delivery, carrier picked up,
confirmed, delayed, delivered, failure, in transit, label
printed, label purchased, out for delivery, ready for
pickup. Read them again and look for the one that means
the parcel is coming back to you. It is not there.
The closest candidate, failure, is documented as the fulfillment request failing, which is a different event entirely: that is your warehouse or your 3PL refusing the job, not a courier reversing a van. Delayed is for a parcel that is still trying. Attempted delivery fires and then fires again and tells you nothing about what happens after the third attempt.
The genuinely odd part is that Shopify has written the
word down. In the unstable version of the same enum there
is a RETURNING_TO_SENDER value, described as
the fulfillment being returned to the sender, sitting
beside a BUYER_ACTION_REQUIRED that would
also be useful. Neither is in a stable version. So the
concept exists in Shopify's vocabulary and not in
anything you, your developer, or any app in the store can
build on today. Everyone in this category is working
around a value that has been drafted and not released.
Meanwhile the order object does what it was told. It went to Fulfilled when the label printed, and nothing after that is allowed to change its mind. An order whose contents are physically in your own receiving area reads exactly like an order sitting happily on a customer's kitchen table.
Flow has nothing to fire on
The reasonable next thought is to automate around it with Shopify Flow, and the reasonable next thought runs into the same hole. In August 2024 a merchant asked the simplest version of the question: he wanted an automation to run once an order is delivered, and asked which parameter to use. The community answer was the fulfillment event created trigger. Four days later he reported back that it does not work, because delivery happens around 48 hours after fulfillment, so by the time the carrier says anything the workflow has already run and finished. The person who suggested it asked whether a delivered event really fails to start the trigger. Nobody from Shopify answered. The thread stops there.
That is the trigger side. The condition side is worse, because even a trigger that fired on every scan would have nothing to compare against: there is no returning to sender status in the stable API for a Flow condition to test. This is the same shape as the parcel that stops moving and never gets flagged, and it has the same root cause. Shopify models the moment you hand the box over. It does not model the days afterwards.
Shopify calls it a return, and it is not one
When the box does come back, the admin offers you the returns flow, which is the wrong drawer. A customer return is a statement about a product. A return to origin is a statement about an address, a delivery window or a courier. Filing them together quietly poisons the only number anyone looks at.
A merchant made this case to Shopify in a May 2026 feature request, asking for a dedicated RTO status separate from customer-initiated returns, on the grounds that mixing them creates misleading return-rate data and makes performance analysis inaccurate. He asked for four things: a separate status, separate reporting, its own webhooks, and filters in Orders and Analytics. The replies are the useful part. One operations lead wrote that an RTO and a customer-initiated return are completely different events operationally, that reporting gets noisy and people start making wrong decisions off the same dashboard, and that manual tags work for a while but break down once order volume increases or multiple people touch fulfillment.
Then the number that should make you check your own dashboard. Another reply described a brand whose return rate looked alarming until they split the data, at which point about 60 percent of it turned out to be RTOs from bad pin codes rather than product issues. Splitting it stopped them discontinuing SKUs that were selling perfectly well. Three months later an app builder in the COD-heavy Indian market added the cleanest summary of the gap: RTO is not a returns problem, it is a fulfillment risk problem, and every app in this space, including his own, has to reconstruct the RTO signal from fulfillment and tracking events instead of getting a clean, structured status.
The workarounds, and what each one breaks
There is no reship button, so merchants have built their own, and the thread where they did it is six years long. January 2020: a package came back because the customer typed the wrong address, and the merchant wanted a new label without her reports showing two items sold where one was. Shopify Support answered twice. First, void the original label if you still can, fix the address on the order and buy a new one. Second, if the void window has closed, duplicate the order and apply a 100 percent discount, which keeps revenue honest and then needs a manual inventory adjustment, because the duplicate takes the stock again.
The void route mostly fails, and the Help Center says why. A label can be voided only if it is under 30 days old, has no tracking events at all, was not bought on USPS First Class Mail, and was not added to a manifest in the last 15 minutes. A parcel that traveled far enough for a courier to give up on it has tracking events by definition. You are paying for that label whatever happens, and you will pay for a second one.
So the community built the third way. Edit the original order, add a custom line item called something like Corrected Shipping Address, mark it as requiring shipping, fix the address, and buy a label against that item at the real weight. No zero dollar order, a fresh tracking number, reports roughly intact. It works, people in the thread confirm it still works years later, and the cost is that your customer receives a shipping confirmation for a line item that is not the thing they bought. In August 2024 a merchant looked at all of this and asked the question that belongs in the title of the feature request: why is this not a standard button.
One more wrinkle, and it is the one that catches accounting. A merchant in March 2024 described his process, mark the order returned, create a return with no label, restock, then either refund or reship free, and reported that Shopify had started deducting returned products from daily sales automatically even when they are not refunded, and flagging the order refund owed, which he said is untrue about 90 percent of the time because the goods went back out at no charge. The only reply, four months later, was another merchant saying he wished he had a solution and wished Shopify supported this natively.
What the silence costs
Price one RTS honestly. Outbound freight, already spent, unvoidable. Return freight, charged by most carriers. A second label if you reship. Eleven days of a customer who has paid and has nothing, which is the part that becomes a chargeback or a public review. Units missing from your sellable count the whole time, so the reorder math reads a shelf that is wrong and the storefront may sell stock that is sitting in a sealed box twelve feet from the packing bench. Then a refund request two months later, from a merchant's own numbers in the community: a 165 dollar order that cost 17.40 to ship, returned after two failed attempts, with the refund conversation starting nine weeks after the parcel came home.
None of those line items is the real cost. The real cost is the eleven days, because every single thing you would do about an RTS gets cheaper the earlier you do it. On day three the customer is still thinking about the order and will happily send you a corrected address. On day eleven they have asked for a refund and your reship has become an argument. The event is the same; the delay is what prices it.
What the automation actually has to do
As a Dugong playbook, in plain prose:
# trigger
Every hour, every open
shipment's tracking feed,
plus a daily age sweep
# steps
1. Read the scans Shopify
does not model, every
carrier, every wording
2. Flag the turnaround the
hour it happens: tag the
order, note the timeline
3. Name the cause: bad
address, refused,
uncollected, customs
4. Email the customer first
and ask for the corrected
address while it matters
5. Apply your rule: reship
at whose cost, or refund
less the shipping
6. Reconcile the box on
arrival, restock what is
sellable, close the order
7. Report RTS on its own,
by courier, by region, by
repeat customer
Step one is the whole trick and it is unglamorous. Nobody can give you a status Shopify has not released, so you read the thing that does know: the carrier. The scan text is inconsistent across couriers and that is fine, because recognizing returning to sender, return to shipper, refused by recipient, unclaimed, undeliverable as addressed and a dozen regional variants as the same event is exactly the sort of judgment a language model is better at than a regular expression. The output is one boolean and one reason code, written back onto the order where the rest of your automation can see it.
Steps two and three turn that into something a person can act on. The tag lands on the order within the hour, the timeline gets a line saying what the carrier said and when, and the cause decides the route. A wrong address goes to the customer, immediately, because the address was probably wrong at checkout and you now have a second chance to fix it. A refused delivery goes to support, because refusal often means a dispute you have not heard about yet. An uncollected parcel at a pickup point is the same failure as a local pickup nobody comes for, and wants a reminder, not a refund. A customs rejection goes to whoever owns paperwork, because it will happen again on the next order to that country.
Step four is where the money actually is. One email, on the day of the turnaround: your parcel is on its way back to us, here is the address we have, reply with a correction and we will send it straight out. Customers answer that email, because it arrives while they are still waiting for a box. The same email on day eleven gets a refund request, and it was the identical email. This is the same economics as telling people about a delay before they ask, and it is the main reason to automate the detection rather than the paperwork.
Steps five and six keep the books straight. The reship or refund rule is yours, applied the same way every time, with the second label priced to your policy rather than to whoever is on chat that afternoon. Then the physical reconciliation, which matters more than it sounds: restocking should happen when somebody has opened the carton and looked, not when the scan appeared, because restocking on faith is how you promise stock you do not have. What comes back damaged or opened gets written off rather than resold. The order closes with a real outcome attached, not a dangling fulfilled.
Why this is a compiler problem, not an app problem
Everything above sounds like a feature, and that is exactly why it keeps not existing. Look at what step five requires: one store reships free once and refuses twice, one deducts the original shipping, one is COD-heavy and blocks the customer's third attempt entirely, one sells perishables and never reships at all. Step three has the same spread: in India the dominant cause is a bad pin code and the fix is address intelligence at checkout, in the US it is apartment numbers and porch refusals, in the EU it is uncollected parcels at pickup points, and for cross-border it is paperwork. Those are not settings on a toggle. They are paragraphs about your shop, and the paragraph is different in every building.
Which is why an app sells you its paragraph and you spend a quarter bending your operation to fit it. A compiler takes yours as written, the same principle behind every automation worth building: describe the RTS policy the way you would explain it to a new hire on their first morning, and let the automation be that explanation, running hourly, in the one place where the carrier is still telling the truth and your store has stopped listening.
Do this before you close the laptop
Two things, both quick. First, pull every order from the last 60 days that is marked fulfilled and has no delivery scan, then open the tracking for each one. Most stores find at least one parcel already home and unrecorded, and a couple still in motion in the wrong direction. That list is your actual RTS rate, which you have probably never seen, because it has been hiding inside your return rate.
Second, write the policy down in four sentences before you automate anything. When we reship free. When the customer pays for the second label. When we refund and what we deduct. Who we will not ship to a second time without a verified address. If you cannot write those four without a meeting, that is the finding, and the first hour of RTS detection will force the meeting anyway.
Then say the brief out loud the way you would say it to someone who starts on Monday. Watch every shipment. The hour a carrier turns one around, tag the order, tell me why, and email the customer for a corrected address. If they answer in two days, reship on our rule; if they do not, refund less shipping and close it. When the box lands, open it, restock what is good, write off what is not. And never, ever put these in the return numbers. The compiler turns that into the workflow, and day eleven stops being when you find out.
If you are a Shopify merchant with a return to sender
story, the box that sat behind the printer, the reship
that broke your reports, the return rate that turned out
to be couriers, the inbox is open:
field-notes@dugong.live. We are collecting
case studies for the next issue.