Guide

Shopify chargeback win rate: what to expect

Published figures checked at source, August 2026. Where no reliable number exists, this guide says so.

"What percentage of chargebacks can I actually win?" is the first question most merchants ask, and it is the one the category answers worst. There is no published win rate for Shopify Payments. The industry averages that do exist disagree with each other by roughly four times, and every single one of them was published by a company that sells chargeback services. Here is what those numbers actually measure, which ones are worth anything, and what genuinely moves yours.

Nobody publishes a Shopify win rate

Start with the absence, because it matters. Shopify reports your chargeback rate in the admin, and it documents the dispute process in detail, but it publishes no figure for how often merchants win. That is consistent with its stated position on outcomes:

"Shopify isn't liable for chargebacks that occur when using the Shopify platform. Shopify isn't involved in the decision making of chargeback outcomes."

The decision sits with the issuer. In Shopify's words, "the company that issued the cardholder's credit card reviews any evidence, and then resolves the chargeback in either your favor or the cardholder's favor." No platform-level win rate exists because Shopify is the pipe the evidence travels through, not the party deciding.

So every number you will find comes from somewhere else — and knowing where is most of the work.

The published numbers, and how far apart they are

These are the figures currently in circulation, each with the company that published it. Read the third column first.

Three different numbers — 12%, 30%, 45% — all describe themselves as the industry average for the same thing. That is not a rounding disagreement. It means at least two of them are measuring something other than what the label says, and there is no neutral referee: no card network, regulator, or platform publishes an audited figure to check them against.

Win rate of what? The denominator problem

The single largest source of disagreement is not sampling. It is that "win rate" names at least three different fractions, and vendors quote whichever one flatters them:

1. Won ÷ contested. Only disputes you actually responded to. The most common headline, and the most generous — every dispute you ignored drops out of the denominator.
2. Won ÷ received. Every dispute that landed, answered or not. Always lower, and closer to what your bank balance experiences.
3. Net recovery — dollars back ÷ dollars disputed. After later dispute cycles, fees, and the labour. Lowest of the three.

The gap between them is not small. The same Chargebacks911 report that yields a 45% headline win rate puts net recovery at 18%. Chargeflow's own figures show the same shape: a 43.82% win rate on one category of dispute alongside an 8.1% net win rate overall. Whatever else is uncertain, the direction is consistent — the number that reaches your bank account is a fraction of the number on the marketing page.

A worked example makes it concrete. Take 100 chargebacks at $80 each. You respond to 50, and win 20 of those. Your win rate is 40% by the first definition, 20% by the second. In dollars, you have recovered $1,600 of $8,000 disputed, and the chargeback fees on all 100 were charged regardless. Three honest numbers, one outcome.

What moves it most: the question the reason code asks

If you only take one thing from published data, take this. The widest measured spread in the numbers is not by store size, industry, or software — it is by what the dispute is actually about.

Two caveats before you use that table. First, "friendly fraud" and "third-party fraud" are not Shopify reason codes — they are the vendor's own classification of what happened, applied after the fact. Second, all eight numbers come from one company's book of business.

But the 4.7× gap between the top two rows describes something real and structural. When a cardholder disputes a purchase they made themselves, you can show they made it — the order, the login, the delivery, the reply to your support email. When a stolen card was genuinely used on your store, you are being asked to prove an authorization that never happened. Delivery proof does not answer that question; a thief's package arrives too.

The card networks have partly turned that question into a rule rather than an argument. Visa's Compelling Evidence 3.0, which applies to card-absent fraud disputes under network reason code 10.4, lets a merchant qualify by showing two prior undisputed transactions on the same credential, dated 120 to 364 days before the dispute, matching on two evidence elements such as purchase IP and device fingerprint — or one of those plus a shipping address, email, or account id. Read that as a shape rather than a checklist: it rewards stores whose customers come back, and a store whose customers buy once has none of that history to submit. (We have not verified how, or whether, the Shopify Payments evidence form exposes these fields — that is unmeasured on our side.)

Which is the practical version of the rule underneath every one of these numbers: evidence that answers the bank's actual question wins; evidence that answers a different question does not. Which question each of Shopify's eight reasons is asking, and the evidence set for each, is in Shopify chargeback reason codes: what evidence each one actually needs.

What moves it second: whether your evidence is in the file at all

The generic advice in this category is "not responding is an automatic loss." On Shopify Payments that sentence is wrong in a way worth being precise about, because the truth is more dangerous than the myth.

So an ignored dispute is not a forfeit — it is a submission built only from what Shopify can look up: products, addresses, order and fulfillment dates, carrier and tracking, customer IP. The merchant-side slots go out reading "Not provided": the reason the chargeback is invalid, customer communication, shipping documentation, proof of service. Exactly the documents that answer the reason code.

That is why "respond at all" outranks every technique. It is not that silence scores zero. It is that silence sends the shipping label to answer a question about an email. What the automatic response contains, field by field, is in what Shopify submits for you when a chargeback deadline passes.

What moves it third: whether the file can be read

Less discussed and entirely mechanical. Evidence uploads accept PDF, JPEG or PNG only, PDFs must be PDF/A and under 50 pages, each file 2 MB or less, all files together 4 MB or less. Audio, video and external links are rejected outright, so a Drive link to "the full evidence pack" is functionally a blank submission. Shopify also recommends high-contrast images, because a fair number of issuers still receive chargeback evidence by fax — light-grey text on white can arrive as an empty page.

None of that is strategy. It is the difference between an analyst reading your case and an analyst reading nothing, and it costs you cases that were winnable on the merits. The full list, and a method for fitting a case inside 4 MB, is in the file limits that quietly reject chargeback evidence.

Disputes that should never be in your denominator

Before you compute anything, take out the cases that were never yours to win or lose. Otherwise you will be measuring your paperwork against a population that includes disputes decided by other rules entirely.

• Orders covered by Shopify Protect. Coverage is narrower than most merchants assume — Shop Pay, US Shopify Payments, physical goods only, fulfilled within 7 days, in transit within 10, address unchanged — and it only ever applies to fraudulent and unrecognized chargebacks. The complete checklist is in why Shopify Protect didn't cover your order.
• Inquiries. An inquiry takes no money and no fee up front. It is a question you can answer before it escalates, and mixing inquiries into a chargeback win rate inflates it.
• Orders you already refunded. These are not delivery disputes at all — they are about the customer being made whole twice, and the refund record is close to the entire case. Covered in you refunded the order, the chargeback came anyway.

The number Shopify actually enforces

Here is a fact that reorders most merchants' priorities. Per Shopify's documentation, "your chargeback rate is based on all disputes that customers open, not just for disputes that you lose," and — stated flatly — "winning a chargeback doesn't remove it from your chargeback rate."

That rate is the number with consequences attached. Shopify documents the card network monitoring programs it enforces: Visa's acquirer monitoring program uses a 0.5% ratio threshold, and Mastercard's Excessive Chargeback Program starts at 100 disputes and a 1.5% dispute rate, with escalating monthly fines and, at the far end, restriction of Shopify Payments itself.

So a perfect win rate does not protect your ability to take payments. Prevention does — clear billing descriptors, responsive support, accurate delivery estimates, fast refunds. Winning disputes recovers money. It does not recover standing.

How to measure your own — and when it means anything

Measure it once, honestly, and you will stop needing anyone else's average:

1. Count only closed disputes. Anything still under review is not data — Shopify's notification put the issuer's decision window at up to 75 days, so a quarter's cases mostly resolve in the next quarter.
2. Use one denominator and name it. "Disputes where I submitted merchant evidence" is the most useful, because it measures the thing you control.
3. Split by reason code the moment you have enough cases. A blended number across fraud and non-delivery disputes averages two populations that behave nothing alike, and tells you what to do about neither.
4. Track the dollars separately from the cases. Winning six $20 disputes and losing one $400 dispute is an 86% win rate and a net loss.

What no tool can promise

Every win-rate figure in this article was published by a company that sells chargeback services, and most of that category is paid a percentage of what it recovers. That pricing model makes the win rate a sales number as well as a measurement, which is a reasonable thing to be aware of when the same publisher's "industry average" and headline result sit four to six times apart on one page.

The honest version is the one Shopify's own documentation implies: an analyst at an issuing bank decides, using card network rules, the cardholder's account history, and a PDF. Nobody selling you software is in that room. What software can do is make sure the PDF answers the question, contains your evidence rather than blank slots, fits inside the limits, and arrives before the deadline — which is the entire lever merchants actually hold, and the reason a disciplined store with a plain submission routinely beats a disorganised one with a 40-page export.

So the answer to "what win rate should I expect" is: expect the reason code to decide most of it, expect a genuine-fraud dispute to be close to unwinnable, expect your net recovery to be well under your case win rate, and expect any specific percentage you are quoted — including the ones above — to be a number someone had an interest in publishing.

Related guides

• How to fight a Shopify chargeback — the full step-by-step
• Shopify chargeback reason codes: what evidence each one actually needs
• What Shopify submits for you when a chargeback deadline passes
• The file limits that quietly reject Shopify chargeback evidence
• Refunded the order, then got the chargeback anyway
• Why Shopify Protect didn't cover your order
• Success-fee calculator — what a 25% success fee costs at your dispute volume

Disclosure: Evidkit builds merchant-side chargeback tooling, so dispute forms are what we stare at all day. Every figure above is attributed to the company that published it and was checked at source in August 2026; where no reliable number exists, this guide says so rather than estimating, and we publish no win rate of our own — no product is required to act on any of it. Not affiliated with or endorsed by Shopify.