This page collects retail and e-commerce security figures from the primary reports: Verizon’s Data Breach Investigations Reports, IBM’s Cost of a Data Breach study, Sophos’s State of Ransomware in Retail 2026, the Thales (Imperva) Bad Bot Reports, the Nilson Report, UK Finance, the FBI, Sansec, the PCI Security Standards Council, the UK NCSC, and company filings from Marks & Spencer, Victoria’s Secret and UNFI. It covers breach frequency and cost, ransomware, card-not-present fraud, web skimming, bots, holiday trends, the major 2025 incidents and PCI DSS. Each figure carries the year it describes.
How to cite: link to this page or to the primary source beside each figure. Survey figures carry their sample size.
Key numbers at a glance
- 806 confirmed retail data breaches in Verizon’s 2026 dataset, up from 419 the year before. (Verizon 2026 DBIR Retail snapshot; Verizon 2025 DBIR Retail snapshot)
- 42% of retail breaches started with the exploitation of a vulnerability, against 31% across all industries. (Verizon 2026 DBIR Retail snapshot; Verizon 2026 DBIR Executive Summary)
- $3.54 million: the average cost of a retail data breach in IBM’s 2025 study. (IBM Cost of a Data Breach Report 2025)
- 32% of retailers whose data was encrypted paid the ransom in 2026, down from 58% a year earlier and the lowest rate of any sector. (Sophos State of Ransomware in Retail 2026, 217 retailers)
- 41% of web traffic to retail sites came from bad bots in 2025. (Thales 2026 Bad Bot Report)
- £423.5 million lost to remote purchase (card-not-present) fraud in the UK in 2025, up 3%. (UK Finance Annual Fraud Report 2026)
1. How often retailers are breached, and how
Verizon’s DBIR reports retail (NAICS 44-45) as its own sector. The 2026 edition analyzed more than 22,000 confirmed breaches in total. (Verizon 2026 DBIR Executive Summary)
- 997 retail security incidents and 806 confirmed data breaches in the 2026 dataset. Verizon says the number of incidents “rose slightly, but the number of breaches nearly doubled.” (Verizon 2026 DBIR Retail snapshot)
- The year before, the 2025 DBIR counted 837 retail incidents and 419 confirmed breaches. (Verizon 2025 DBIR Retail snapshot)
- 95% of retail breaches fell into three patterns: System Intrusion, Basic Web Application Attacks and Social Engineering. (Verizon 2026 DBIR Retail snapshot)
- 99% of retail breaches were caused by external actors; 1% by insiders. (Verizon 2026 DBIR Retail snapshot)
- 85% of retail breaches were financially motivated. Espionage rose to 19% of breaches with a known motive, from 9% the year before. (Verizon 2026 DBIR Retail snapshot)
- Initial access in retail breaches: exploitation of vulnerabilities 42%, credential abuse 14%, phishing 9%. Across all industries, exploitation was 31% and credential abuse 13%. (Verizon 2026 DBIR Retail snapshot; Verizon 2026 DBIR Executive Summary)
- A third party was involved in 68% of retail breaches, against 48% across all industries. (Verizon 2026 DBIR Retail snapshot; Verizon 2026 DBIR Executive Summary)
- The human element was present in 58% of retail breaches, against 62% overall. (Verizon 2026 DBIR Retail snapshot; Verizon 2026 DBIR Executive Summary)
- Ransomware was the top malware action in retail breaches, and phishing was almost twice as common as pretexting among social attacks. (Verizon 2026 DBIR Retail snapshot)
What data attackers take. Payment cards are no longer the main prize in retail breaches.
- Internal data was compromised in 84% of retail breaches, up from 65% the year before. Credentials were taken in 26% and secrets in 20%. (Verizon 2026 DBIR Retail snapshot)
- In the 2025 DBIR, payment data appeared in just 12% of retail breaches. (Verizon 2025 DBIR Retail snapshot)
What this means: the most common way into a retailer is now an unpatched, internet-facing system, and two-thirds of breaches touch a supplier, platform or service provider. A retail test plan should start with the external perimeter, the storefront and its APIs, and the third-party integrations around checkout. Service providers include outsourced support: an ecommerce customer service outsourcing team that can look up orders and issue refunds holds access an attacker can abuse, so its accounts and permissions belong in the review too. Our guide to e-commerce and retail penetration testing covers how that scope is built.
2. What a retail breach costs
- The average retail data breach cost $3.54 million in IBM’s 2025 study, up from $3.48 million in the 2024 study. IBM defines retail as “brick and mortar and e-commerce.” (IBM Cost of a Data Breach Report 2025, 600 organizations, retail 5% of the sample)
- The 2025 global average across all industries was $4.44 million, and the US average was $10.22 million, up 9%. (IBM 2025)
- IBM’s 2026 report puts the global average at $4.99 million, a 12% increase and a record high. IBM’s public summary of the 2026 report does not break out retail, so the 2025 figure above is the latest retail number cited here. (IBM Cost of a Data Breach Report 2026, 602 organizations, breaches from March 2025 to February 2026, per IBM’s press release)
- Recovering from a ransomware attack cost retailers an average of $1.76 million, excluding any ransom, up from $1.65 million in 2025 and above the $1.70 million cross-sector average. The median was $375,000. (Sophos State of Ransomware in Retail 2026, 217 retailers)
- 6% of retail ransomware recoveries cost $10 million or more, against 2% across all sectors. (Sophos Retail 2026)
Section 8 shows what disruptive attacks cost large retailers in 2025, from their own filings.
3. Ransomware in retail
Figures in this section come from Sophos’s survey of 217 IT and security leaders at retailers in 17 countries that were hit by ransomware, fielded from January to March 2026 as part of a 2,158-organization study. (Sophos State of Ransomware in Retail 2026; summary)
- 26% of attacks on retailers began with an exploited vulnerability, the highest share of any sector; the cross-sector rate was 18%. (Sophos Retail 2026)
- Phishing was the top technical root cause at 29%, up from 23%. Malicious email caused 22%. (Sophos Retail 2026)
- Identity-based vectors (email, phishing, stolen credentials, brute force) started 69% of retail attacks, the lowest of any sector against 79% overall. (Sophos Retail 2026)
- 62% of attacks on retailers encrypted data, up from 48% in 2025 and above the 56% cross-sector rate. (Sophos Retail 2026)
- 19% of attacks both encrypted and stole data, up from 14%. (Sophos Retail 2026)
- When a retailer’s firewall identified the attack before ransomware was deployed, data was still encrypted 60% of the time, against 50% across all sectors. (Sophos Retail 2026)
- 32% of retailers with encrypted data paid the ransom and got data back, down from 58% in 2025. That was the lowest rate of any sector (the cross-sector rate was 48%) and the lowest in five years of retail data. (Sophos Retail 2026)
- 71% restored encrypted data from backups, up from 62%. (Sophos Retail 2026)
- The median ransom demand was $225,000, but the mean was $4.32 million, the highest of any sector, because of a small group of very large demands. (Sophos Retail 2026)
- The median payment was $57,500, the lowest of any sector, down from $1 million in 2025 and $1.4 million in 2024. 63% of retailers that paid handed over less than the initial demand. (Sophos Retail 2026)
- Mean recovery time fell to 2.54 weeks from 3.16 weeks, and 56% recovered within a week, the best result in five years of retail data. (Sophos Retail 2026)
In the cross-sector study, state and local government paid most often, at 72%. (Sophos State of Ransomware 2026, 2,158 organizations)
What this means: retailers recover without paying more often, but more attacks reach encryption. The gap is early in the attack: exposed systems with known vulnerabilities, and alerts that fire without stopping the attacker. Our ransomware statistics page has the cross-industry numbers.
See what a real pentest finds. Read our sample report, then get a fixed quote in writing.
4. Card-not-present and e-commerce fraud losses
- Card fraud losses worldwide were $33.41 billion in 2024, down 1.2% from $33.83 billion in 2023, on total card volume of $51.920 trillion. (Nilson Report, Issue 1298, December 2025; Nilson Report article)
- UK remote purchase fraud, where stolen card details are used to buy online, rose 3% to £423.5 million in 2025. Cases rose 13% to 3.2 million. (UK Finance Annual Fraud Report 2026)
- UK purchase scams, where a buyer pays for goods that never arrive, rose 20% to £118.1 million and made up 71% of all authorized push payment scam cases. (UK Finance 2026)
- Total UK payment fraud losses were £1.28 billion in 2025, up 4%. (UK Finance 2026)
- The FBI received 56,478 non-payment and non-delivery complaints in 2025 (goods paid for and never received, or shipped and never paid for), up from 49,572 in 2024. Reported losses were $503.4 million, down from $785.4 million. (FBI IC3 2025 Internet Crime Report)
- Credit card and check fraud complaints to the FBI reached 18,774 in 2025, with losses of $282.7 million, up from $199.9 million in 2024. (FBI IC3 2025)
Nilson’s detailed breakdowns are subscriber-only, so only its worldwide total is quoted here. FBI figures count only complaints filed with the IC3.
5. Web skimming and Magecart
Web skimming (often called Magecart) injects code into a store’s checkout to copy card details as shoppers type them.
- Magecart infections made up 80% of breaches involving payment cards in Verizon’s 2025 dataset, but only 1% of System Intrusion breaches. (Verizon 2025 DBIR)
- In a multi-year contributor dataset of 43,324 Magecart-compromised websites, the median site had around 7,000 monthly visitors and the median infection lasted less than 30 days. Verizon concluded that attackers pick victims by opportunity, not size. (Verizon 2025 DBIR)
- Six weeks after Adobe’s emergency patch for the SessionReaper flaw (CVE-2025-54236), 62% of Magento stores were still vulnerable, as attacks began in October 2025. (Sansec, October 2025)
- After Sansec disclosed the PolyShell file upload flaw in Magento and Adobe Commerce in March 2026, it observed attack attempts against 79.5% of the stores it protects. (Sansec, March 2026)
- One PolyShell attack wave compromised 471 stores in a single hour. (Sansec, March 2026)
- One loader group injected code into 953 Magento stores between July 2025 and October 2026. (Sansec, October 2026)
- A fake storefront network skimming cards under hundreds of real brand names reached 5,714 sites by June 2026, growing by about 70 domains a day. (Sansec, June 2026)
- Just before PCI DSS’s script controls took effect, 38% of Magento stores used a Content Security Policy and 13% had CSP reporting enabled, across the 400,000 largest stores Sansec tracks. (Sansec, March 2025)
What this means: skimming is a patching and change-control problem as much as a payment problem. Exploitation of e-commerce platforms now arrives in mass waves, hundreds of stores at a time, and many stores are still unpatched weeks after a fix ships. A web application penetration test of the checkout, admin panel and extensions, and a review of every third-party script on payment pages, address both halves.
6. Bots and account takeover
Figures from the Thales (Imperva) Bad Bot Reports, built on traffic across Thales’s network, which blocked 17.2 trillion bot requests in 2025. (Thales 2026 Bad Bot Report)
- Bots made up more than 53% of all web traffic in 2025, up from 51%; bad bots alone were 40%. (Imperva, 2026 Bad Bot Report summary; 2026 Bad Bot Report)
- Retail web traffic in 2025 was 41% bad bots, 14% good bots and 45% human. (Thales 2026 Bad Bot Report)
- The previous edition put bad bots at 33% of retail traffic in 2024, up from 26% in 2023, and ranked retail the second most attacked industry with 15% of all bot attacks. (Imperva 2025 Bad Bot Report)
- 55% of bot attacks on retail in 2025 were classed as advanced, the second highest share of any industry after law and government. (Thales 2026 Bad Bot Report)
- Retail was the top target for AI bots, receiving 20% of AI bot attacks. (Thales 2026 Bad Bot Report)
- Retail was also the most targeted industry for business logic abuse, with 24% of those attacks (travel was second at 17%). Examples include loyalty abuse, transaction manipulation and bots that fill carts to block inventory. (Thales 2026 Bad Bot Report)
- Account takeover attacks rose 70% from July 2024 to July 2025 across all industries. Financial services drew 46% of them. (Thales 2026 Bad Bot Report)
The report gives no retail-only account takeover figure; the retail risk shows up as business logic abuse against login, loyalty, cart and checkout flows. Those flows run on APIs, which drew 27% of all bot attacks in 2025 (Imperva); that is why an API penetration test is part of a retail scope.
7. Holiday-season attack trends
- Retail traffic rose 37% above the November average over Black Friday 2025, and bot attacks on retail sites rose 50%. (Imperva, Black Friday 2025 in Review)
- Malicious Black Friday traffic was concentrated in the US (46%), Australia (12%) and the UK (11%). (Imperva, Black Friday 2025)
- During major sales events, bot traffic to retailers rises 2 to 3 times compared with normal days. (Imperva 2025 Bad Bot Report)
- In 2024, bot traffic against retail spiked around April and September as well as from October onward, rather than climbing steadily toward November as it did in 2022 and 2023. (Imperva 2025 Bad Bot Report)
What this means: test well before peak season, and treat every major launch or sale as a bot event, not only Q4.
8. Major retail incidents in 2025, from official statements
Marks & Spencer (UK). On 22 April 2025 M&S announced it was managing a cyber incident and paused orders on its UK and Ireland websites and apps.
- M&S estimated the impact at around £300 million on 2025/26 operating profit before cost management, insurance and other mitigation. (M&S full year results, May 2025)
- The website was available for browsing only from late April to early June. UK online sales fell 42.9% to £382.3 million in the first half, and Fashion, Home & Beauty sales fell 16.4%. (M&S interim results, November 2025)
- For the full 2025/26 year, M&S booked £131.3 million of direct incident costs and £100.0 million of insurance proceeds. Adjusted profit before tax was £671.4 million, down from £881.1 million. (M&S full year results, May 2026)
- The NCSC estimates the cost to M&S and its insurers at more than £300 million. (NCSC Annual Review 2025)
Co-op Group (UK). The Co-op was hit on 25 April 2025, according to an open letter from its chief executive published in the NCSC’s annual review.
- The NCSC’s review says the attacks by the DragonForce ransomware group left customers unable to make payments and saw the data of all 6.5 million Co-op members stolen. (NCSC Annual Review 2025)
The UK picture. The NCSC handled 429 incidents in its 2025 reporting year. 204 were nationally significant, up from 89, and 18 were highly significant, a 50% increase and the third annual rise in a row. After the retail attacks, the NCSC mobilized its Retail Sector Trust Group to brief the wider sector. (NCSC Annual Review 2025)
United States.
- Victoria’s Secret detected a security incident on 24 May 2025 and closed its US e-commerce site for four days. It estimated the incident cut net sales by about $20 million and operating income by about $14 million. (Victoria’s Secret & Co. Form 10-Q, quarter ended 2 August 2025)
- UNFI, a grocery wholesaler that supplies retailers, detected unauthorized activity on 5 June 2025, which temporarily disrupted its ability to fulfill and distribute orders. (UNFI Form 8-K) It later reported $26 million of incident costs, mainly shrink and remediation, in its fiscal fourth quarter. (UNFI Q4 fiscal 2025 results)
- Verizon’s 2026 retail analysis cites the Hot Topic breach, which affected 57 million customers. (Verizon 2026 DBIR Retail snapshot)
The common thread in these filings: most of the reported cost came from stopped operations (paused websites, disconnected warehouse systems, lost sales), not from the stolen data itself. Our e-commerce industry page covers how online retailers are tested for that kind of disruption.
9. PCI DSS context
Every retailer that takes cards falls under PCI DSS. The changes most relevant to e-commerce took effect in 2025.
- PCI DSS v4.0.1 requirements 6.4.3 (managing payment page scripts), 11.6.1 (detecting changes to payment pages) and 12.3.1 (targeted risk analysis) became effective on 31 March 2025. (PCI Security Standards Council, January 2025)
- For merchants that validate with SAQ A, the PCI SSC removed 6.4.3 and 11.6.1 from the questionnaire and replaced them with an eligibility criterion: the merchant must confirm its site “is not susceptible to attacks from scripts” that could affect its e-commerce systems. The underlying PCI DSS requirements were not removed. (PCI SSC, January 2025)
- A February 2025 FAQ explained that merchants can meet that criterion with techniques like those in 6.4.3 and 11.6.1, or with written confirmation from a PCI DSS compliant payment provider whose embedded payment form protects the page. (PCI SSC, 28 February 2025)
PCI DSS also requires penetration testing at least annually and after significant changes (Requirement 11.4), separate from quarterly ASV scans. Our PCI DSS compliance checklist covers all twelve requirements, PCI DSS penetration testing requirements explains 11.4 line by line, and ASV scan vs penetration test explains what each one proves.
What to expect in 2027
These are sourced forecasts, or trends labelled as such, drawn from the data above.
- More exploitation of retail systems (Sophos forecast). Sophos expects exploited vulnerabilities to climb again as organizations struggle to keep up with patches for flaws found by frontier AI models. Retail is already the sector most often breached this way. (Sophos Retail 2026)
- Slower patching meets mass exploitation (trend). Only 26% of critical vulnerabilities on the CISA Known Exploited Vulnerabilities list were fully fixed in 2025, down from 38%, while e-commerce platform flaws were mass-exploited in 2025 and 2026, in one case compromising 471 stores in an hour. (Verizon 2026; Sansec)
- More bot traffic, more of it AI-driven (trend). Bad bot share of retail traffic rose from 26% (2023) to 33% (2024) to 41% (2025), and retail is the top target of AI bots. (Imperva 2025; Thales 2026)
- Fewer ransom payments, more encryption (trend). Retail payment rates fell from 58% to 32% in one year while encryption rose from 48% to 62%. (Sophos Retail 2026)
- Supplier and platform risk keeps growing (trend). Third parties were involved in 68% of retail breaches, and breaches with third-party involvement grew 60% across all industries. (Verizon 2026 Retail snapshot; Verizon 2026 DBIR Executive Summary)
Related statistics
Part of our cybersecurity statistics hub. See also small business cyber attack statistics and data breach statistics.
Sources
- Verizon, 2026 Data Breach Investigations Report: Retail snapshot, 2026
- Verizon, 2026 Data Breach Investigations Report: Executive Summary, 2026
- Verizon, 2025 Data Breach Investigations Report and 2025 Retail snapshot, 2025
- IBM and Ponemon Institute, Cost of a Data Breach Report 2025, 2025
- IBM, Cost of a Data Breach Report 2026 and press release, 2026
- Sophos, The State of Ransomware in Retail 2026 (summary), 2026
- Sophos, The State of Ransomware 2026, 2026
- Thales (Imperva), 2026 Bad Bot Report and summary, 2026
- Imperva, 2025 Bad Bot Report, 2025
- Imperva, Black Friday 2025 in Review, December 2025
- The Nilson Report, Issue 1298: Card Fraud Losses Worldwide, 2024, December 2025
- UK Finance, Annual Fraud Report 2026, June 2026
- FBI Internet Crime Complaint Center, 2025 Internet Crime Report, 2026
- Sansec, research on SessionReaper (2025), PolyShell and the PolyShell attack wave (2026), the CMS block loader campaign (2026), GorgonAgora (2026) and CSP adoption (2025)
- PCI Security Standards Council, Important updates for merchants validating to SAQ A and FAQ on SAQ A eligibility, 2025
- National Cyber Security Centre, Annual Review 2025, 2025
- Marks and Spencer Group plc, full year results May 2025, interim results November 2025 and full year results May 2026
- Victoria’s Secret & Co., Form 10-Q for the quarter ended 2 August 2025, 2025
- United Natural Foods, Inc., Form 8-K, June 2025 and Q4 fiscal 2025 earnings release, 2025
Frequently asked questions
How many retailers are breached each year? Verizon’s 2026 DBIR counted 997 retail security incidents and 806 confirmed data breaches, up from 419 confirmed breaches in the 2025 edition. These are the cases in Verizon’s dataset, not a count of every breach. (Verizon 2026; Verizon 2025)
What does a data breach cost a retailer? $3.54 million on average in IBM’s 2025 study, below the $4.44 million all-industry average. Recovering from ransomware cost retailers $1.76 million on average in 2026, excluding any ransom, according to Sophos’s survey of 217 retailers. (IBM; Sophos)
How do attackers get into retailers? Most often through unpatched vulnerabilities. Exploitation was the initial access vector in 42% of retail breaches in Verizon’s 2026 data, against 14% for credential abuse and 9% for phishing. In Sophos’s ransomware survey, retail had the highest share of attacks starting with an exploited vulnerability (26%). (Verizon; Sophos)
Do retailers pay ransoms? Less than any other sector. In 2026, 32% of retailers whose data was encrypted paid and recovered it, down from 58% in 2025, and the median retail payment fell to $57,500. (Sophos)
How much card-not-present fraud is there? In the UK, remote purchase fraud reached £423.5 million across 3.2 million cases in 2025. Worldwide, all card fraud losses were $33.41 billion in 2024, according to the Nilson Report. (UK Finance; Nilson)
How much retail traffic comes from bots? 41% of retail web traffic was bad bots in 2025, and bot attacks on retail sites rose 50% above the November average over Black Friday 2025. (Thales; Imperva)
Written by
Wahid Iqbal, Head of Compliance
20+ years in the field. Focus: PCI DSS, ISO 27001 & ISO 22301. Part of the team that scopes, tests, and reports for our clients. Full profile →