Key Takeaways
- 22% of merchants reported a rise in account takeover (ATO) fraud in 2023
- Friendly fraud accounted for 44% of all chargebacks in 2022
- Triangulation fraud made up 8% of schemes in 2022
- In 2023, global ecommerce fraud losses reached $48 billion
- Global digital fraud losses projected to hit $343 billion by 2027
- Average cost per fraudulent order was $112 in 2023
- US ecommerce fraud losses totaled $12.5 billion in 2023
- Payment fraud in Europe grew by 12% in 2023
- Brazil saw 1.2 billion fraud attempts in 2023
- Ecommerce fraud attempts increased by 18% year-over-year in 2022
- 15% of online transactions were flagged as suspicious in Q4 2023
- Card-not-present (CNP) fraud rose 20% in 2023 globally
- Fraud detection tools reduced losses by 40% for 65% of retailers using AI in 2023
- Biometric authentication prevented 72% of ATO attempts in 2023 trials
- Machine learning models improved fraud detection accuracy to 98% for top merchants
In 2023, ecommerce fraud surged with account takeovers, promo abuse, and chargebacks climbing sharply worldwide.
Related reading
01 · Category
Common Fraud Types26 stats
Common Fraud Types Interpretation
02 · Category
Financial Losses26 stats
Financial Losses Interpretation
03 · Category
Geographic Distribution27 stats
Geographic Distribution Interpretation
More related reading
04 · Category
Global Prevalence24 stats
Global Prevalence Interpretation
05 · Category
Mitigation Strategies27 stats
Mitigation Strategies Interpretation
Cite This Report
This report is designed to be cited. We maintain stable URLs and versioned verification dates. Copy the format appropriate for your publication below.
Samuel Norberg. (2026, February 13). Ecommerce Fraud Statistics. Gitnux. https://gitnux.org/ecommerce-fraud-statistics
Samuel Norberg. "Ecommerce Fraud Statistics." Gitnux, 13 Feb 2026, https://gitnux.org/ecommerce-fraud-statistics.
Samuel Norberg. 2026. "Ecommerce Fraud Statistics." Gitnux. https://gitnux.org/ecommerce-fraud-statistics.
Sources & references
100 datasets cited across this report · attribution is report-level

