When a cardholder taps "dispute this transaction" in their banking app, the number on the screen — say £65, roughly the UK average disputed value[1] — is the smallest part of what that tap sets in motion.
By the time the case closes, weeks later, that single dispute will have touched the cardholder's issuer, the card network, the acquirer, and the merchant — each adding cost at every hop. The visible loss is the transaction value. The true cost is everything stacked underneath it. As someone who works these cases daily, this is the economics I wish more people in payments talked about.
The cost stack nobody itemises
Start with what the merchant loses. The transaction amount is reversed provisionally on day one. On top of that sits a non-refundable chargeback fee — in the UK typically £15–£25, sometimes far higher[4]. Then the invisible layers: the goods or service already delivered, the shipping and fulfilment spend, one to two hours of staff time assembling a representment pack, and third-party tooling. Mastercard and Javelin put the merchant's average at $82 in internal costs plus $46 in third-party fees per chargeback — before counting the disputed amount itself[1]. Industry analyses consistently land on the same conclusion: the all-in cost of a dispute runs at two to three times the original sale value[4].
The issuer's side is quieter but just as real. Every dispute costs a financial institution roughly $9–$10 in pure processing, and US institutions need about one full-time back-office employee for every $13,000–$14,000 in annual dispute value[1]. That staffing maths is why triage quality is a P&L question, not just a compliance one.
$128
Average merchant cost per chargeback in third-party fees + internal costs — excluding the disputed amount
[1]
261M → 324M
Projected global chargeback volume growth, 2025 to 2028 (Mastercard / Datos Insights)
[2]
~12%
Average merchant net recovery rate — what is actually won back after fighting
[3]
Model it yourself
Rather than take my word for it, here's the arithmetic as an interactive model. Defaults reflect published UK/US averages — drag the sliders to your own reality. The most interesting slider is the last one: how many disputes you resolve before they become chargebacks — by refunding or answering the customer the moment they raise an issue, instead of letting their bank file a formal dispute. The industry calls this "deflection".
Interactive · True Cost Model
Merchant view, per dispute and per year. Figures update live.
—
True cost per dispute — fee + staff time + expected lost value
—
Cost multiple of the transaction's face value
—
Annual exposure at this dispute volume
—
Annual saving from resolving disputes early, before a chargeback is filed
MODEL: fee fixed at £20 (UK range £15–£25). Expected lost value = amount × (1 − win rate); goods & fulfilment are written off on the same share of cases. Cases resolved early are assumed to end in a refund, so they save the fee and staff time but not the transaction value. Illustrative simplification — real unit economics vary by vertical and processor.
Even on conservative defaults, the model lands where the industry data does: a dispute you lose costs close to double its face value once goods and handling are written off — and even blended across your wins, every case runs well above face. The fastest way to shrink the number isn't winning more representments; it's stopping disputes from becoming chargebacks at all.
The friendly-fraud measurement problem
Ask five sources what share of chargebacks is friendly fraud — a cardholder disputing a transaction they actually made — and you'll get answers from 21% to around 70%[3]. That absurd range is itself the finding: intent doesn't appear in dispute data. Networks label the same behaviour differently (Visa leans toward "first-party misuse"), merchants hold different levels of evidence, and issuers see only what the cardholder tells them.
"You cannot manage what you cannot classify — and friendly fraud is, by design, misclassified."
What's not in doubt is the direction of travel. Global chargeback volume is projected to grow roughly 24% between 2025 and 2028[2], and Sift's network data shows first- and third-party fraud together making up about 45% of merchant dispute volume[5]. Generational data adds a behavioural wrinkle: younger cohorts increasingly treat the dispute button as a refund channel[5].
What "good" actually looks like
The single most underused lever is embarrassingly low-tech: talking to each other earlier. Javelin finds customers go straight to their bank in about 75% of dispute cases — yet when they contact the merchant first, 44% of issues are resolved by simply answering a question, and another 31% by a refund[3]. Only a quarter ever needed to become a chargeback.
Layer on the tooling built for exactly this gap — real-time dispute alerts, Visa's Rapid Dispute Resolution, Ethoca/Verifi collaboration rails — and the strategy writes itself: resolve disputes before they are filed, fight the ones worth fighting, and treat win rate as a secondary metric to prevention rate.
The Analyst's Takeaway
A dispute team that measures itself on win rate is optimising the expensive end of the pipeline. The economics point upstream: every dispute resolved before filing saves the full cost stack, every well triaged case avoids a hopeless representment, and every correctly classified friendly-fraud pattern compounds into better policy.
That's the lens I built into both tools on this page — the Mastercard Triage Advisor and the Crypto Card Simulator — and it's the lens I'd bring to a fraud strategy or payments analytics role.
Sources
Luca Caranzano — Chargeback Analyst, Monzo Bank. Written for this portfolio to demonstrate dispute economics literacy beyond day-to-day casework. Figures are third-party published estimates (US-centric where noted), used for illustration; this article is not financial, legal, or operational advice.