2014 · Germany · K - Financial and insurance activities · Data theft and extortion via compromised public-website database
Analysed 2026-11-06 by Hermes
Open this case in the tool →Data theft and extortion via compromised public-website database
Expertise to Execution
Evidence coverage: 3/6 stages, weight 16/60
Decisive stage: egress (clearest evidence) · Alternative: Strategy to Execution
(1) As the eurozone's central bank, the ECB was expected to protect even the modest personal data it held - the contact details of people registering for its conferences and visits - and it publicly presents data security as taken extremely seriously. (2) In practice, attackers exploited an undisclosed vulnerability in the database serving the public website and took around 20,000 contact records; the encryption the ECB applied to 95 percent of that database stopped short of the email-address, street-address and phone-number fields, so what left the network was readable plaintext; the ECB detected nothing itself - the theft came to light only when the attacker sent an anonymous ransom demand - and at disclosure it could not say whether the whole database or only parts had been copied. Containment, recovery and the response were sound by contrast: the database was physically separate from internal systems, the vulnerability was fixed, passwords reset, affected people contacted, the police informed and the ransom publicly refused. (3) The classification rests on the egress stage: the remedy that was missing (field-level encryption) was not an unknown - it was in live use in the same database - so the chain broke between available technical expertise and operational architecture, inside the material column, with no source showing any leadership decision that deferred or under-funded it; that is Expertise to Execution. The entry vulnerability was never technically specified in public reporting and is not load-bearing for this classification. (4) RM26 (Expertise to Execution lens): encryption expertise existed and was applied to most fields, yet the specialists' own protection did not reach the contact fields, leaving them in plaintext so the stolen data arrived readable and the amplifying effect fell exactly on the fields that gave the extortionist leverage (causal role: Amplifying). RM34 (Expertise to Execution lens): monitoring and alerting on the internet-facing web database were not configured or tuned to catch the intrusion and the bulk read of ~20,000 records, which obscures the incident until the attacker himself announces it (causal role: Obscuring). RM33 (Expertise to Execution lens): the ECB could not reconstruct the scope of the loss - it stated at disclosure that it did not know whether the entire database or only parts had been copied - so the logging and forensic evidence needed to scope the egress was not available to its own specialists (causal role: Obscuring).
RM26 — Data-protection / cryptographic weaknessRM33 — Logging / forensic-readiness gapRM34 — Monitoring / detection / alert-handling gapSource type informationisbeautiful: the row's one-sentence description matches the July 2014 incident and the stated victim, so no data-quality mismatch. Caveat: the dataset's source text says an anonymous call; the ECB's own press release and all other reporting opened say an anonymous email - the substance (ransom demand for the stolen data, refusal to pay) is the same. The dataset's source_url (cityam.com/1406190300/...) no longer resolves in that form; cited instead is the current City AM article URL opened in this task. The exploited vulnerability was never specified publicly (ECB gave no technical details), attribution remained unknown, and no regulator investigation, fine or prosecution of the 2014 incident is publicly documented (EDPS searches surfaced nothing). A separate, later ECB incident exists (August 2019 closure of the BIRD portal website), but it is not linked as a related row and its comparability was not assessed here. No consent-gated steps were skipped in this unattended run.