“Fraud rate” in ecommerce is not one universal number. Depending on the dataset, it can mean: (1) fraud attack rate (share of orders targeted),
(2) confirmed fraud rate (share of orders confirmed as fraud), (3) fraud loss rate (share of revenue lost), or (4) fraud-related disputes (chargebacks with fraud reason codes).
This page standardizes the terms and shows how to report fraud benchmarks in a way researchers can actually compare.
Back to the hub:
E-commerce Statistics.
Use this together with
payment failure rate,
chargeback rate,
and payment methods share.
Key benchmarks (cite-ready)
Fraud benchmarks are most useful when you publish the metric type (attack vs confirmed vs losses) and the segment (region/category/payment method).
Below are the most commonly cited “anchor points” that appear in reputable industry reporting.
The Europe order-level and revenue-loss benchmarks are useful “anchors” when you want a citeable reference.
For global context, fraud loss forecasts are easier to cite than global fraud rates because fraud definitions vary by processor and fraud stack.
- Europe order/revenue reference points (industry summary referencing MRC): ~3% of orders affected; ~2.8% of revenue lost. 0
- Global online payment fraud loss forecast: losses expected to exceed $362B between 2023–2028 (projection). 1
Definitions (pick the one you mean)
Most “fraud rate” confusion comes from mixing these definitions. Choose one primary definition, then publish supporting ones where possible.
| Metric | What it measures | Best use | Common pitfalls |
|---|---|---|---|
| Fraud attack rate | Share of orders targeted by fraud attempts | Threat level tracking and capacity planning | Can rise even when confirmed fraud stays flat (better defenses) |
| Confirmed fraud rate | Share of orders confirmed as fraudulent | Merchant-to-merchant comparability (if confirmation rules are stable) | Depends on confirmation method (chargebacks, manual review, insurer decisions) |
| Fraud loss rate | Fraud losses as % of revenue | Executive-level impact benchmarking | Often excludes hidden costs (ops, delivery, customer support) |
| Fraud dispute rate | Chargebacks with fraud reason codes (e.g., CNP fraud) | Card-network aligned risk reporting | Understates fraud when “friendly fraud” is coded as non-fraud disputes |
| First-party misuse (friendly fraud) | Disputes by legitimate cardholders (false claims, policy abuse) | Post-purchase risk + disputes strategy | Often reported as “disputes” not “fraud,” but can dominate outcomes |
Visa’s VAMP program is an example of why definitions matter: it uses a single ratio combining fraud and disputes over settled transactions for card-not-present transactions. 2
Segments (one number never describes the market)
If you want your fraud numbers to be linkable and comparable, publish them with at least these segments.
This is the “minimum viable segmentation” researchers expect.
| Segment | What to publish | Why it changes fraud | Pair with |
|---|---|---|---|
| Category / vertical | Attack rate + confirmed fraud by category | Digital goods, subscription-like categories, and high-ticket categories often see different dispute and fraud patterns. | return rate |
| Domestic vs cross-border | Fraud rate for cross-border separately | Cross-border transactions typically face higher risk and different issuer behavior. | cross-border purchase share |
| Payment method | Fraud and disputes by method (cards vs wallets vs transfers) | Authentication and liability models differ by method. | payment methods share |
| Device | Fraud and disputes by device | Mobile UX + authentication flows can affect both fraud and false declines. | mobile revenue share |
| New vs returning | Fraud rate by customer type | First-time buyers are harder to score; returning buyers can be trusted via history. | repeat purchase rate |
| Post-purchase abuse | Refund abuse + “friendly fraud” indicators | Industry reporting shows first-party misuse disputes are rising for many merchants. | chargeback rate |
Why “friendly fraud” must be separated
A large portion of disputes can come from legitimate customers (first-party misuse). If you blend this into “fraud rate,” your numbers become incomparable across merchants.
Reporting template (copy/paste for your dataset)
If you publish fraud benchmarks, include these fields next to the number so researchers can cite it safely.
- Metric type: attack rate / confirmed fraud / fraud losses / fraud disputes
- Time window: month/quarter + year
- Market: country/region; domestic vs cross-border
- Category mix: list top categories included
- Payment mix: include link to payment methods share
- Customer split: new vs returning
- Disputes layer: include link to chargeback rate
- Operational notes: whether 3DS/SCA is used broadly (impacts both fraud and friction)
In regulated contexts, strong customer authentication is widely cited as effective, even while fraud attempts evolve (useful for Europe-focused writing). 4
Sources
Primary and high-signal sources used for benchmarks, definitions, and citation anchors.
- Juniper Research — forecast of online payment fraud losses (2023–2028). Source
- McKinsey — references Juniper fraud loss forecast in payments context. Source
- Visa — VAMP ratio definition combining fraud + disputes over settled transactions (CNP). Source
- Sift (EU summary referencing MRC) — Europe anchor points: fraud share of orders and revenue-loss rate. Source
- Reuters / ECB — payment fraud totals in the EEA and notes on SCA effectiveness. Source
- Riskified — chargeback fraud / first-party misuse evidence points from merchant survey. Source
- Signifyd — trend framing: “fraud pressure” and consumer abuse growth signals (directional trend data). Source
Cite this page
Copy and paste.
Best for Ecommerce. (2026). Fraud rate benchmarks (ecommerce). Retrieved from
/ecommerce-statistics/payments-risk/fraud-rate-benchmarks/
