The Chargeback Math of Slow Shipping: What Every Dispute Actually Costs You

Dropshipping chargebacks cost far more than the order value. On a $60 order lost to an item-not-received dispute, you refund $60, forfeit $15 in non-refundable Stripe dispute fees plus $15 more to counter it, and eat the goods and the postage already spent. The reversal is the smallest line. Here is the full arithmetic.

The two clocks running on every item-not-received dispute

A cardholder who says the parcel never came does not file a vague complaint. They file into a specific slot. On Visa that slot is Condition 13.1, Merchandise/Services Not Received, and Visa's own merchant guidelines describe the trigger precisely: the cardholder claims the goods "were not received by the expected date." Stripe files the same family of reason codes under the category product_not_received.

Note what the rule turns on. Not whether you shipped. Whether the customer had it by the date they expected it.

The first clock is the customer's. Card networks typically allow cardholders 120 days from the original payment to open a dispute, and longer in some cases. That clock starts at checkout and it runs against a delivery promise the customer usually invented themselves from your shipping page.

The second clock is the money clock, and it is far less forgiving. The moment the issuer files, the network pulls the funds out of your processor and your processor pulls them out of your balance. Those funds are held for the entire duration. Stripe gives you 7 to 21 days to respond depending on the network, then the issuer takes 60 to 75 days to decide. Shopify tells merchants the same thing in plainer terms: the credit card company's review "can take up to 75 days after you submit." The full lifecycle runs two to three months, and you cannot speed it up except by giving up.

So a dispute filed on day 45 of a $60 order removes $60 plus fees from your balance on day 45, and the earliest realistic resolution is somewhere past day 120. If you are running lean on working capital, that hold is the part that actually hurts. We went through the mechanics of that squeeze in the cash-flow math of dropshipping.

The fee stack: what a dispute costs before you count the goods

Most stores know there is "a chargeback fee." Very few know the shape of it, and the shape differs by processor in a way that changes whether fighting is worth it.

Sources: Stripe pricing and June 2025 dispute fee update, Shopify Help Center on responding to chargebacks.

Read the Stripe column again. You pay $15 to be accused. You pay another $15 to defend yourself. Stripe states it directly: "Unless otherwise stated in your Stripe contract, we never return the dispute received fee." A perfect win, on a dispute that was nonsense from the start, still leaves you $15 down.

Outside the US the numbers move. Shopify Payments charges £10 GBP in the UK, €15 EUR in Germany, $15 CAD in Canada and $25 AUD in Australia. Stripe's fees run €20 in most of the eurozone and $25 AUD in Australia, applied twice on a countered dispute. Nothing about a "worldwide" store is uniform, which is part of why we run country-level shipping pages.

The full cost of one lost dispute, line by line

Here is a single order taken all the way through: $60.00 retail, one parcel, one lost dispute after a fight. Every input below is illustrative except the two processor fees, which are Stripe's published US figures. Swap your own numbers in.

That is the number worth sitting with. A $60 dispute is a $134.54 event. At an illustrative $15 contribution margin per order, you need to ship nine more orders just to get back to zero. If you have not built a per-order model that carries these lines, our dropshipping profit calculator and the walkthrough of how to calculate dropshipping profit are the place to start.

For a wider sanity check on the multiplier: LexisNexis Risk Solutions, surveying 569 fraud and risk executives for its 2025 True Cost of Fraud study, found US ecommerce and retail businesses lose $4.61 for every $1 of fraud. Different metric, same lesson: the visible loss is never the loss.

Fighting it: what a 44.6% win rate is actually worth

The benchmark most merchants quote comes from Chargebacks911's 2026 Chargeback Field Report, which puts the average representment win rate at 44.6% against a net recovery rate of 10.7%. The gap between those two numbers is the story. Winning the argument is not the same as getting the money back and keeping it.

Run the decision on the fees alone, using the Stripe US column and the same illustrative $60 order.

Fighting beats accepting by $18.45 before labour, and by about $10 after you price 20 minutes of someone's attention. Useful, and worth doing. But look at where the line sits.

Break-even is a 20% win rate on fees alone, and roughly 31% once you cost 20 minutes at $24 an hour. Below that, submitting evidence is a hobby. Above it, you are still choosing between two losses.

That is the honest frame. Representment is damage control, not a revenue channel. The only outcome that actually pays is the parcel arriving before the customer thinks about their bank. Everything in the rest of this post is about that.

Why item-not-received is decided on tracking quality, not on being right

Visa's instruction to merchants on Condition 13.1 is one line long and it is not about your intentions. Provide documentation proving that the cardholder, or an authorised person, received the goods as agreed.

The prevention section of the same document is more specific about what counts. Visa names certified mail, and a carrier's certification that the item was delivered to the correct address or was signed for by the cardholder, as the proof that lets you return the dispute.

Stripe operationalises that into two evidence fields: shipping_documentation and shipping_tracking_number. What Stripe does with the second one is the part most people miss. When Stripe compiles your evidence, the tracking numbers "are expanded to include detailed delivery information from the carrier." Shopify does something similar, auto-populating the response with product details, "shipping company used and tracking information," the date and time the order was fulfilled, both addresses, and the customer's IP.

So your defence is not what you say. Your defence is what the carrier's system says when a third party looks up your number.

Three ways that goes wrong on a China-direct order, all of them operational rather than legal:

The number resolves to nothing. If the tracking reference sits on a network the issuer's staff cannot query, the expansion returns an empty record. You submitted a shipping label and a hope.

The number resolves but the story stops. Events read "departed origin country" and then nothing for eleven days. That is not proof of delivery. It is documentary evidence that the customer's complaint was reasonable.

The number resolves to something the customer cannot read. This one costs you the dispute before the dispute exists. A customer in Manchester pastes a code into Royal Mail, gets "not found," emails support once, waits two days, and calls their bank instead. A tracking number the cardholder cannot understand is functionally identical to no tracking. It generates the ticket, then it generates the chargeback.

Support load and dispute load are the same load measured at two different moments. That is exactly the mechanism we picked apart in how we cut returns from 8% to 2%: the fixes that reduce "where is my order" tickets are the same fixes that reduce disputes, because they are the same customer at an earlier stage of losing patience.

The ratio problem: the number that can end card acceptance

Here is the part that turns a cost line into an existential one.

Stripe states it flatly: "All disputes, whether they're won or lost, count towards your dispute rate." And on the network side: "Monitoring programs don't consider dispute outcomes." You can win every single case and your ratio will not move by one basis point.

The thresholds are published.

Two details that matter more than the headline percentages.

Visa tightened the merchant threshold this year. Visa's VAMP fact sheet puts the Excessive Merchant ratio at 220bps for the US, EU, Canada and Asia-Pacific, and footnotes that it drops to 150bps on 1 April 2026. That change is already live. The bar you were measured against last year is not the bar now.

Mastercard adds a per-chargeback surcharge on top of the fine. Once the issuer recovery assessment applies, it is an extra $5 for every chargeback above 300. Stripe's worked example: a merchant in month 4 of ECM with 400 disputes is assessed $5,500, not $5,000. Exiting requires three consecutive months back under the threshold.

Now scale it to a real store. At 500 orders a month, eight disputes is a 1.6% rate. Eight. That is inside Mastercard's ECM band on ratio, and you got there from roughly one and a half unhappy customers a week. Winning all eight changes nothing about the number.

How dispatch speed rewrites the whole calculation

Everything above compresses into one operating variable: the gap between the date the customer expects the parcel and the date it actually lands.

The networks and the processors agree on how to close it. Stripe's prevention guidance for unreceived products is short: ship quickly, hand the cardholder a tracking number at ship time, state shipping times before checkout, and communicate any delay fast with an option to refund. Visa says the same thing to merchants: if goods will be late, tell the customer in writing, give them the revised expected date, and let them cancel.

There is one more lever, and it is the cheapest one on the board. Stripe: "Disputes and chargebacks aren't possible on credit card charges that are fully refunded." A $60 refund issued on day 20, on a parcel that has not scanned in ten days, costs $60 and your processing fee. The same order lost as a dispute costs $134.54 in the illustrative model above, and adds a tick to a ratio that cannot be un-ticked. Refund the stuck parcel. It is a bargain.

Now the part where dispatch speed does the structural work. Four days sitting in a supplier's queue before the first scan is four days of a delivery promise running with nothing to show the customer. It is also four days you cannot get back later in the transit leg, and four days closer to the date the customer decides you are not real.

This is the specific problem we built Peregrine's fulfillment stack around. Orders dispatch from our own Shenzhen warehouse in under 24 hours, so the first tracking event exists on day one rather than day six. Delivery runs 3 to 10 days to 65+ countries at 99.8% delivery accuracy. Every parcel hands off to the local carrier the customer already trusts, USPS, Royal Mail, La Poste, DHL, Australia Post, under branded tracking. That last detail is the one that shows up in your dispute evidence: the cardholder gets a number that resolves in a system they recognise, and so does the issuer's back office when Stripe or Shopify expands it.

None of that makes disputes disappear. Friendly fraud exists, addresses get mistyped, parcels get stolen off porches. What it removes is the category you caused yourself: the dispute that starts with a customer refreshing a tracking page that has said nothing for a week. Routing and timings are on the shipping page, or connect a store and run a real order through it.