An ASG warehouse bench. This photo illustrates fulfillment operations in general — it is not evidence for any specific order, review, or dispute discussed in this article. By Janson, CEO of ASG Dropshipping · Updated 2026-09-16 · 27 min read
You changed the lane, and your Shopify fulfillment is finally behaving. Your dashboard still reads like the stretch you are trying to put behind you. That is not a lag you can wait out. Those orders were judged on what happened to them, and a better lane does not go back and re-judge them.
I’m Janson. I run fulfillment out of Shenzhen and Dongguan, and this is the part of a recovery sellers ask me about most.
Fixing your Shopify fulfillment changes what happens to your next order. It does not rewrite the records already written against earlier ones.
What a bad stretch leaves behind is not a pile of orders. It is a pile of work items — each waiting on a next decision — living in four places: your reviews, your payments back office, your fulfillment records, and the promise on your product page. One order number can be carrying several at once.
Sort on two things only: what has somebody else’s deadline printed on it, and what will never come find you. Know where your leverage stops: a chargeback is decided by the customer’s card-issuing bank, on that order’s evidence, not on how your lane runs today (Shopify Help Center , retrieved 2026-09-14).
You did the hard part already. You found the bad lane, moved off it, and your recent orders go out clean. And the store still reads like nothing changed — like the fix didn’t take.
Your Shopify fulfillment fix applies to orders you haven’t shipped yet. The review a customer posted and the dispute they filed were both written against the order they actually received, and a new lane does not reach backwards into those records. That is why the store can be healthy and the dashboard still look sick.
A review is a public note attached to a moment that already happened. A dispute is the same: once filed, it reads the evidence from that order, not your current lane.
That boundary decides where your effort goes — but it gets stretched into something it does not say. It does not say a review can only be replied to. The person who wrote it can change the rating, rewrite the text, or delete it outright — Shop’s help pages give buyers the steps (Shop Help Center , retrieved 2026-09-16). What is fixed is that you cannot do it for them.
A 2025 Sifted survey of 500 US shoppers reports that nearly half of consumers stop buying from a brand after a poor delivery or packaging experience (Sifted , retrieved 2026-09-14).
Read that for what it is: what people told a survey, not churn measured in an order table — and Sifted sells delivery intelligence, so they are not neutral. Read down, it still points somewhere useful: the damage that worries me is the part that never files a ticket. Sellers feel it the moment the reviews land:
“So I got a couple of complaints from long shipping times and I’m not sure whether to display those reviews or hide them.”
— Davis-Hend, Shopify Community thread #413168 , 2025-05-08
For how the break happens in the first place, see why fulfillment breaks when Shopify orders start growing .
Do this next: decide what counts as your bad period, in writing. Pick a start and end date, and anchor every record to the date the order was placed . Then pull those orders, reviews and disputes into one list — and don’t fix anything yet.
I get it — sorting feels like stalling. But you are about to spend real money, and the sort stops you spending it where it cannot work.
To rebuild customer trust after shipping delays, stop counting orders and start counting work items. A work item is one thing waiting on a next decision, and one order number can carry several. They sit in four queues: reviews and replies owed, payment disputes, fulfillment records, and the promise on your own site. Work the ones with a printed deadline first.
Illustrative diagram, not a screenshot of any real order or dashboard. It shows why one order number can carry work in Queue A, B and C simultaneously; Queue D never carries an order number. The four queues
The four queues
Queue
Where it lives, and what you see on screen
What you can press right now
A deadline someone else printed?
A · Reviews, and replies you owe Your public review listing; and on a single order, the Contact section or the Timeline
Post a public reply; report a review that breaks platform policy; answer the message still sitting there
No
B · Payment disputes Your payments back office, each one showing a status
Refund to stop an inquiry escalating; submit evidence on an open chargeback; on a decided one, nothing this round
Yes — printed on the dispute itself
C · Fulfillment records Shopify’s fulfillment status field; and, separately, your carrier’s scans
Chase it, re-source it, upgrade the lane, cancel cleanly; or send the documents that clear customs
No
D · The promise on your own site Your product page, notification templates, shipping policy. There is no order number here
Edit the text, before it becomes the promise on the next order
No
Queue A · Reviews, and replies you owe Where it lives, and what you see on screen Your public review listing; and on a single order, the Contact section or the Timeline What you can press right now Post a public reply; report a review that breaks platform policy; answer the message still sitting there A deadline someone else printed? No Queue B · Payment disputes Where it lives, and what you see on screen Your payments back office, each one showing a status What you can press right now Refund to stop an inquiry escalating; submit evidence on an open chargeback; on a decided one, nothing this round A deadline someone else printed? Yes — printed on the dispute itself Queue C · Fulfillment records Where it lives, and what you see on screen Shopify’s fulfillment status field; and, separately, your carrier’s scans What you can press right now Chase it, re-source it, upgrade the lane, cancel cleanly; or send the documents that clear customs A deadline someone else printed? No Queue D · The promise on your own site Where it lives, and what you see on screen Your product page, notification templates, shipping policy. There is no order number here What you can press right now Edit the text, before it becomes the promise on the next order A deadline someone else printed? No
These four do not carve up your damage
Four places to look, not four boxes every record drops into. Two things keep it honest:
One order number can be carrying several items at once, with no cap. The same queue can hold more than one item from one order. Ship in two parcels and one unit can sit in customs while the other never left — two items in Queue C, one order number.
That “several at once” is an inference , drawn from three systems keeping independent records; the chargeback section shows one merchant carrying two.
The triage question
For any item, ask: in the system it lives in, is there still a next step I can press, and has somebody else printed a deadline on it?
That sorts your Shopify fulfillment backlog, but the answers flip. An order released from hold drops back to its previous status (Shopify Help Center , retrieved 2026-09-16); a chargeback you won can be reopened; a buyer can delete their own review tonight. Re-walk this list, don’t file it.
What a single work item looks like on paper
What a single work item looks like on paper
Field
What goes in it
Order number The order it came off, or blank for a Queue D item Queue A, B, C or D Current status Copied from the screen it lives on, in that system’s own words The promise it was sold on What your page said when they bought Last verified event The most recent thing you actually confirmed, not assumed Dispute status and deadline Copy the date and time zone off your own back office. Not a range from this articleNext action One action, not a plan Owner A name Next time you’ll speak to them A date you will hit
How to count each queue
Do not add these up. They are different units; a single total would be a number about nothing.
Queue A — two counts, reported separately: the reviews you can tie to an order, and the conversations awaiting your reply in that order’s Contact section or Timeline (Shopify Help Center , retrieved 2026-09-16). Never merged into one figure. Reviews you cannot tie to an order get their own line.
Queue B — three counts, one per state: inquiries, open chargebacks, decided ones. An inquiry can escalate into a full chargeback (Shopify Help Center , retrieved 2026-09-16), and a won one can be reopened — so your B count can legitimately come out higher than your order count.
Queue C — count work items, not unique orders. A split shipment gives you two.
Queue D — a separate list of text to change , which never joins the other three counts.
What is in no queue at all
Five kinds of damage in no count you just made:
The buyer who quietly stopped buying — no ticket, no review, no dispute. Write down that you cannot count them.Reviews on sites you do not control. A buyer deleting their own review , which moves your numbers without you.The bank’s internal deliberation on an open dispute: you submit into it, you don’t see it.Sending a delay notice in bulk , for which I could not verify a sourced method.
The next three sections take A, B and C in order; D follows the notification build.
Do this next: register every item with a deadline first, using the fields above. Deadlines are the only part of this list that expires while you think about it.
This is the section nobody wants. The honest answer about Queue A is that you do not get to take anything down — but you have more moves than “reply and hope”, and one is regularly missed.
The first everybody sees: the public review, already working on the next visitor. The second is quieter — open a single order and there is a Contact section (email, phone or text) and a Timeline where you can resend an order email or leave an internal note (Shopify Help Center , retrieved 2026-09-16). Every unanswered message in there is a work item with your name on it.
Who can do what to a posted review
You cannot take it down — not by improving your fulfillment, not by asking a partner to. Anyone selling you review removal is selling something they do not have. What you can do is reply in public, or report a policy breach, which is a request to the platform rather than a decision of yours. The buyer can do what you can’t: Shop’s help pages walk the reviewer through changing their rating, rewriting their text, or deleting the review (Shop Help Center , retrieved 2026-09-16).
So the record is not frozen — just not yours to move, and a customer whose problem you actually resolve is holding an edit button you do not have.
When a reply is worth writing
Write it when undecided buyers will see the review, and when the complaint is specific enough that you can name what changed without promising a timeline.
Skip it when it would be your fourth near-identical paragraph in a row — copy-pasted apologies read as a bad stretch twice over — and skip it when what you are really doing is arguing. The next customer is reading.
Back to Davis-Hend’s question: show the 1-star reviews or hide them? Three merchants answered; not one said hide. One put it plainly:
“It’s usually better to show the 1-star reviews — but respond to them professionally. Hiding all negative feedback can damage trust more than a few poor reviews ever would.”
— goldi07, Shopify Community thread #413168 , 2025-05-08
A second, Kudosi-Carlos, wrote on the same thread that he would keep honest reviews up because real social proof is stronger. The third went further, and I come back to him in the delivery-promise section. That is one thread, not a study.
Do this next: work the unanswered conversations before the public replies. Those are the ones where somebody is still waiting on you personally.
This is Queue B, the queue with a clock on it , and the worst mistake is treating every dispute the same. Three states, and the button differs in each.
A dispute over a late delivery passes through three states, and your options change with each. During an inquiry no money has been taken yet, so a full refund is still available. Once a chargeback is open, the amount and fee are already withdrawn and an ordinary refund cannot be issued. Once the bank decides, that round of evidence is over — though the transaction can be disputed again.
Illustrative diagram, not a screenshot of Shopify’s dashboard. The date ranges shown are Shopify’s published figures, not any specific order’s real deadline — check your own back office for that. The three states
One: an inquiry. No money is taken during the investigation (Shopify Help Center , retrieved 2026-09-16). That makes a refund cheap here and nowhere else: if you judge the reason behind an inquiry justified, you can issue a full refund while the card company is still holding nothing (Shopify Help Center , retrieved 2026-09-14). It is the one state where refunding does not mean refunding money already withdrawn.
Two: an open chargeback. The disputed amount and the fee are withdrawn immediately, and refunds cannot be issued while a chargeback is open. The move that worked one state ago is gone; what is left is the evidence file.
Three: decided. If it went your way the amount and fee come back; if not, they are gone. Either way this round of evidence is over — the correct scope for the word. A won chargeback can be disputed again on the same transaction, and a lost inquiry can escalate into a full one. The state moves; it does not retire.
Shopify is blunt about who decides all this: it “isn’t involved in the decision making of chargeback outcomes,” and strong evidence carries no guarantee of a favourable result (Shopify Help Center , retrieved 2026-09-14). The card-issuing bank decides.
Two clocks, and only one of them is yours
The first is your evidence window: typically 7 to 21 days after the chargeback is filed (Shopify Help Center , retrieved 2026-09-14).
The second is the bank’s review, which runs after you submit. That can take up to 75 days on a chargeback; on an inquiry it usually runs 65 to 75 days and may take longer (Shopify Help Center , retrieved 2026-09-16).
One you can miss; the other you cannot shorten. And those are Shopify’s published figures — the date that governs your case is the one printed on that dispute in your own back office.
What actually goes in the file
Submit the evidence that relates to that order: the label, the tracking history, the delivery scan, messages you exchanged with that customer.
This is about relevance , not age: a delivery that completed after the dispute was filed still relates to that order, and so does a written withdrawal from the customer.
What does not belong is your improvement story. The new lane, the new partner, the better handoff — none of it was in effect on the order under review.
Being organised does not guarantee the outcome
Here is the case that made me stop softening this section. A merchant submitted the label and the tracking, showing the parcel still moving, and lost anyway.
“I have an issue with chargebacks. I received chargebacks for order not received, and I provided shipping label and tracking as my response to the chargeback which showed that it was still in transit, and for some reason I still lost the chargeback. This is unfair, as the product was still delivered and I received no profit, plus a chargeback fee… Please help me out as I’m a small business and I’m struggling due to this.”
— heavensent, Shopify Community thread #96158 , 2022-02-01
Another merchant replied the next day, then said the part nobody puts in a sales page:
“Other than that, you’re going to lose even when you’ve done nothing wrong.”
— happymug, Shopify Community thread #96158 , 2022-02-02
Two merchants on one thread is not a loss rate.
But look at that first situation again — it is the clearest example of what the queue table is telling you. A parcel physically moving, a Queue C item, and a dispute counting down in the payments back office. One order number, two queues, two different next actions.
There is a single-order version of this, one step down: when supplier delay becomes a refund or cancellation risk .
Do this next: write down two things per dispute — which of the three states it is in, and the deadline printed on it. State first, because it decides which button you even have.
Queue C is where your Shopify fulfillment leverage actually is , and the one most often worked as a single pile. It holds two kinds of item that look alike on a dashboard and need different hands.
C1 — it has not gone out yet. This is Shopify’s fulfillment status field: unfulfilled, in progress, on hold, partially fulfilled (Shopify Help Center , retrieved 2026-09-16).
C2 — it has gone out, and stopped. A tracking number exists, no delivery scan has landed, and it shows on your carrier’s side, not in that status field.
Why these are two screens: the values Shopify documents for that field include none meaning “shipped but not delivered.” Transit is a tracking concept, not a fulfillment-status value. Build one filter for both and you will silently lose a kind.
The orders still sitting with you
This kind has the most leverage, because everything is still in reach: chase the supplier, re-source, upgrade the lane, or cancel cleanly before it becomes somebody’s disappointment.
And on hold deserves a warning. A hold reserves inventory but blocks the shipment until released — and orders on hold are excluded from your list of unfulfilled orders (Shopify Help Center , retrieved 2026-09-16).
Read that twice: your most actionable orders quietly dropped out of the list you check every morning. They do not chase you.
The orders that stopped moving
Here you are mostly not the actor, and pretending otherwise wastes the week.
If it is held at customs, the fix is documents. Shipments are held to satisfy import or export rules, typically over missing paperwork, and the carrier contacts the shipper — nothing clears until the authorities have what they asked for, unless you run your own broker (FedEx , retrieved 2026-09-16).
No days figure and no duty figure from me: it depends on the goods, the route and the office.
And no email to your customer moves that parcel. Send the update because they deserve to know; do not mistake sending it for having done the work.
Where the standard advice is right
The common advice here is about talking to customers during a delay, and where somebody is genuinely still waiting, it is correct. The sharpest version is a competitor’s, so I’ll quote it as written.
“Delays are not what damage customer trust. Silence is.”
— OrderStatus Pro, How to Communicate Shopify Order Delays Without Losing Customer Trust (published 2026-04-12; retrieved 2026-09-14)
A merchant, order_ops_guy, made the same point in operator language on the Shopify Community: customers handle delays far better while the order still feels alive, and the damage starts when tracking stops updating and nothing is said (thread #28144 , 2026-05-14). Two others are worth reading directly — GlowLoyalty on ranges rather than hard dates, and Shopify’s own guide to shipping delays on causes and handling.
And here is the boundary: a message is not a shipment. For an order on hold in your building, or a parcel behind a customs form, communication runs alongside the work. It is not the work.
For why tracking goes quiet, see why tracking updates become a fulfillment bottleneck .
Do this next: filter for orders on hold today, separately from everything else. They are the ones your normal view is hiding from you.
You have probably gone looking for this button. Most sellers do, and come back confused — because what you can send today and what fires on its own are two different things, and hearing only about the second is what makes people give up.
On any single order you can contact that customer directly from the order’s Contact section, by email, phone or text. What you cannot do is set it to fire by itself: Shopify’s notifications documentation lists customer emails tied to specific events — order placed, fulfilled, refunded, exchanged — and a proactive “your shipment is delayed” trigger is not among them.
Start with what already works. Open the order, go to the Contact section, and reach that customer by email, phone or text; from the Timeline you can resend an order email or leave an internal note (Shopify Help Center , retrieved 2026-09-16). For one order and one customer you have a path right now. What you are assembling is the habit and the wording, not a missing feature.
What is missing is the automatic version. The notifications documentation covers emails fired by specific events (Shopify Help Center , retrieved 2026-09-14). A delay trigger is not among the types that page lists, and I could not verify a sourced method for sending one across a batch. That is a statement about what is documented there, not me telling you the platform cannot do it.
Merchants have been asking for the built-in version for years:
“I am not a developer. Is there an easy way to send a custom message to my customers to inform them of an extended delay in delivery?”
— ktaylor-2, Shopify Community thread #28144 , 2020-12-13
An older 2019 thread carries the same request (Shopify Community , posted 2019-10-23), and its replies pointed at banners, checkout notes and third-party apps.
The three things a delay notice has to carry
Where it is now. Not a status code. A sentence a person can picture.Why it is late. One clause, in plain language, no supply-chain vocabulary.When you will speak next. A specific date you will actually hit.
That third one is the hard one, and it earns the message: a scan tells the customer where a parcel was, not what you are doing about it. And it is a commitment — name a date and you send something on that date, even with nothing new to say.
For what your customer sees on the tracking side, especially with more than one supplier, see multi-supplier Shopify tracking confusion .
An ASG packing bench. Shown to illustrate a general fulfillment workspace — not the specific notification workflow described in this section, and not evidence of any individual order’s handling. Do this next: write the delay email once, with a blank for the next-update date. The template is not the hard part. Naming the date is.
Your product page is still making the old promise. It was written when the old lane worked, and nobody went back to it. That page is Queue D: no order number, nothing to count — just text you can edit before it becomes the promise on somebody’s next order.
A delivery promise holds when something real is holding it up: the lane you ship on, the stock you position ahead of demand, whether you have inventory near the customer. Change those and you change the promise. Change only the wording and you have written a new claim against the same capability.
Our own order-processing procedure says it in boring language: priority lanes, pre-stocking and overseas warehousing set a delivery target, they do not guarantee an outcome.
The mistake is quoting a number without its conditions
I’ll use our own public site , since it is the one I can send you to. Our homepage FAQ says orders to the USA “typically arrive in 5–8 days via our dedicated express line ,” with processing “typically 1–3 business days from payment confirmation .” The same page’s delivery-time tabs, split by warehouse, list the China Warehouse at a typical 4-12 days to the USA.
Those are not two claims about one route. The 5–8 is a named express line, sitting inside the tab’s broader range rather than arguing with it. A figure quoted without its conditions will not match your own other figures — and the customer holding the slower one writes the review.
So before you promise a date on a specific order, get four things straight: where it is going, which warehouse it ships from, whether it has actually left, and which service line it is on.
A hard date is a promise you handed a carrier to keep for you; a range with its conditions attached is one you can hold. GlowLoyalty argues the same (retrieved 2026-09-14), and I would go further: the conditions belong in the same sentence as the range, not three clicks away in a policy page.
Say where the clock starts and stops, too — ours starts at payment confirmation, not checkout. Decide whether your window ends at the first carrier scan or at delivery, because your customer has already assumed one.
Delivery promise comparison
The promise on your page now
What would have to hold it up
The version that can hold
A hard delivery date
A carrier that never varies, on a lane you don’t control
A range, with the factors that move it named in the same sentence
“Fast shipping” with no definition
Nothing. It means whatever the angry customer decides it meant
A stated window per destination, phrased as typical rather than fixed
A blanket cutoff time
Pick-and-pack capacity that holds on your worst day, not your average one
A cutoff you meet on your worst day, stated as typical
Silence on delays
Someone noticing before the customer does
A named next-update commitment, per the section above
“Typically” does real work in that table, but not all of it: it marks a figure as typical rather than guaranteed, and says nothing about what it depends on. You need both.
One more thing about the old reviews — the third merchant on that 1-star thread surprised me.
“negative reviews complaining about shipping times are actually good reviews, because they help set expectations for future customers.”
— JoesIdeas, Shopify Community thread #413168 , 2025-05-09
I would not go that far, but he is pointing at something real: promise and reviews get read together, and a promise the reviews contradict is worse than none.
Do this next: open your product page and read the delivery line out loud. If you cannot name what holds it up, it is a claim, not a promise.
You want to know what you can hand off. A Shopify fulfillment partner owns a real slice of this, narrower than most pitches imply.
What we take on vs. what we do not
What we take on
What we do not
Who decides in the end
Lane choice, pre-stocking, overseas warehousing to set a delivery target
Commit to a fixed transit time that ignores destination, warehouse and service line
Your lane and destination, quoted per order
Free reshipment or refund where the fault is ours — quality, logistics, wrong item
Absorb the procurement cost when a buyer cancels after we have bought the goods
The responsibility split, agreed before the order
Proactive tracking of exceptions against internal response targets
Publish those targets as a blanket public service level
Us, on our own performance, in writing to you
Assemble the process record and evidence material for a dispute
Make the appeal decision, or promise you will win it
The card-issuing bank
One line from our internal rules belongs under that table, because a cell would swallow it: ASG cannot make appeal, compensation, refund or chargeback decisions on a seller’s behalf. We build the evidence trail; we are not allowed to write that as a promise of a successful appeal, and we do not. Shopify’s documentation lands in the same place from a different direction (Shopify Help Center , retrieved 2026-09-14) — two publishers, one boundary.
Three things I will not promise, and what you get instead
We cannot erase a historical negative review. We can ship your orders on a lane whose exceptions we track and report back to you, with a dated record of how each went — so your reply to an old review can name something specific.
We cannot restore a payment provider’s confidence. We can give you a clean, dated record of how each order was handled, so your side of the conversation is documented rather than remembered.
We cannot promise you an appeal outcome. We can hand you the evidence pack for a disputed order — label, tracking history, handling record — within an agreed working window, on one precondition: you forward us the dispute and its deadline. We do not see your payments back office.
Each is a “no” with a specific “here’s what you get instead” attached. A merchant put the point better than I can:
“As a business owner, your merchant-customer relations is going to be important and ultimately, it’s a decision you need to make for what you believe is best.”
— Dirk, Shopify Community thread #92586 , 2022-01-19
That applies to us too: we execute the part that is ours, and the judgement calls stay yours. If your conclusion is that the partner is the problem, the criteria are in when to replace your dropshipping supplier or agent .
ASG’s dispatch area, with a wall sign identifying it as the outbound sealing zone. Shown to illustrate the outbound handoff point in general — not evidence of any specific order’s fulfillment record. Want a second pair of eyes on the sort? Send your three oldest still-open disputes, one line each straight off the work-item table, plus the lanes you were running. We’ll tell you which a fulfillment change can still reach and which it cannot — including where the answer is that we can’t help. Form and contact: asgdropshipping.com/contact-us .
Do this next: take any promise a partner makes you and ask who decides the outcome. If the answer is a bank, a platform or a customer, the promise is not theirs to make.
Let me argue the other side properly , because there is a serious version of it. When the numbers stall, the blanket discount is the lever I watch sellers reach for first, and the one here I can price honestly.
The case goes like this: sorting takes hours you do not have while the store is bleeding, so refund everyone from the bad period, attach a discount code, and you have stopped it today with no judgement calls.
This is not a straw man. An experienced merchant recommends it in public, answering a question close to yours.
“you can offer a complimentary discount code for a future order (say 20% off). You could even offer a partial refund and give him back the amount he paid in shipping costs.”
— Dirk, Shopify Community thread #92586 , 2022-01-19
Note what he is actually advising: a gesture to one customer who came to him with a problem, not a refund of a whole period. I am the one extending his logic into the opposing case, and you should see that. Another merchant, TravelingBags, went further, writing that he was trying to automate the apology email and its discount code for whenever an order went wrong (Shopify Community thread #566985 , 2025-09-22).
Where the fast-refund case genuinely wins
It wins on speed, and speed is a real asset when your attention, not your Shopify fulfillment, is the bottleneck — and when the affected group is small enough that the refunds cost less than the hours the sort would take.
It also wins on inquiries, not as a judgement call but as the platform mechanism from the chargeback section: in the inquiry state no funds have been taken yet, so a full refund gives back nothing you have already lost.
If that describes your situation, do it. I would rather you act than optimise.
Where I would still sort first
I have no evidence that discounting slows a recovery down — nobody I found has measured it. My objection is mechanical, not statistical.
Start with the part that cannot be executed at all. Once a chargeback is open, an ordinary refund cannot be issued on that order — so “refund everyone” is not a policy you can apply to your whole list. The platform stops you on exactly the items with a deadline running: a blanket action quietly skips the queue costing you the most.
Then look at what you are giving back. These orders were delivered; the goods are with the customer and are not coming back. What you hand over is the landed cost, the outbound shipping you already paid, and the money you collected — plus the chargeback fee if that order later loses a dispute. The merchant quoted earlier lost all three together. Calling that “margin” makes the fast option sound cheaper than it is.
And notice the pricing. A blanket action prices every item as though they were in the same state: the customer whose dispute the bank already decided gets the same code as the one whose parcel is still moving. Only one of those endings you can influence — and you are spending when your cash is thinnest. A sort costs an afternoon; a blanket refund costs the afternoon and the cash.
Does fixing your shipping lane change the reviews you already have?
What is the unit you should be counting?
Which items should be worked first?
Can anyone guarantee you will win a late-delivery chargeback?
Is proactive delay communication enough on its own?
Can a customer still file a chargeback after the order was delivered?
Yes. One merchant described receiving disputes for “order not received” on parcels still in transit that arrived afterwards (Shopify Community ). Delivery is evidence you can submit, and one completed after filing still relates to that order — but it does not close the case, because the bank weighs the whole file.
How long do I have on a late-delivery chargeback — and does that deadline decide the outcome?
Two different clocks. Your evidence window is typically 7 to 21 days after the chargeback is filed (Shopify Help Center , retrieved 2026-09-14); the bank’s review then runs separately, up to 75 days on a chargeback and usually 65 to 75 on an inquiry (Shopify Help Center , retrieved 2026-09-16).
Missing the first costs you your say; meeting it buys no result. The date that governs your case is the one printed on the dispute in your own back office.
If I already won a chargeback, is that money safe?
No. Shopify’s documentation states that winning the first chargeback does not guarantee you keep the amount if the transaction is disputed again (Shopify Help Center , retrieved 2026-09-14). Treat a win as the end of that round of evidence, not of the matter, and keep the file.
Can I remove a negative review if the delay was the carrier’s fault?
No — whose fault it was has no bearing on who holds the delete button. That is the buyer, and Shop’s help pages give them the steps (Shop Help Center , retrieved 2026-09-16). Your side is a public reply, or a policy report the platform rules on. No fulfillment partner can promise removal.
When does holding partial local stock actually pay for itself?
When demand repeats predictably, the goods are cheap to hold, and your orders cluster in one region — any one failing is reason enough not to stock. I am not going to give you an order-count threshold , and be wary of anyone who does: it depends on your catalogue in ways I have not seen measured. The three are a filter for when to run the numbers, not a formula.
Rebuilding trust is not a stronger version of fixing your Shopify fulfillment. It is a different job on a different set of records, and only one of the two runs forward in time.
I have watched sellers spend a lot of margin on items where nothing they did could change the ending — not because they were careless, but because nobody had told them the work divided at all.
You do not need a recovery campaign. You need each queue counted in its own unit and never summed, an action list ordered by the deadlines printed in your own back office, and a short list of the copy still making a promise you cannot hold.
This article carries no ASG outcome metrics — no defect rate, no order volume, no on-time percentage. A piece arguing that results are not promises should not open with our own results.
The delivery-day figures in the promise section are a deliberate exception, not a performance claim. They are quoted from our own public homepage as a worked example of that section’s point: a delivery figure means nothing until you say which destination, warehouse and service line it describes. You can open that page and check them.
The capability statements come from four internal ASG documents, one per row of that responsibility table: our order-processing SOP, which treats lane choice, pre-stocking and overseas warehousing as ways to set a delivery target rather than guarantee one; our payment and settlement policy, which sets who carries the cost of a reshipment, refund or post-procurement cancellation; our logistics exception-handling standard, which makes response times internal targets rather than a public service level; and our boundary rule that ASG cannot make appeal, compensation, refund or chargeback decisions for a seller.
You cannot open any of the four — the one place here where you take my word for it.
Company scale figures appear once, in the author note below.
I’m Janson, CEO of ASG Dropshipping. We have been running fulfillment for Shopify sellers since 2019, with a 200+ team and 4 warehouses in Shenzhen and Dongguan.
Most of what I write comes out of the same place this piece did. A seller arrives after something has already gone wrong.
The first useful thing anyone can do is separate the items that still have a next step from the ones that do not.
