← Back to news 22.11.2026 Series: Basic terms, part 13

Restricted is not erased, pseudonymised is not anonymous

Basic terms, part 13: the deletion concept says "erase", the database has a tick at "inactive", and the name has been replaced by a number. Four words are often used as if they meant the same. Under the GDPR, only two of them lead out of data protection law.

The ticket says "Erase customer data after the contract ends" and has the status "done". In the database, the customer now has a tick at "inactive", in the report the name has been replaced by a customer number, and the mapping table sits in the same system. Nothing has been erased.

Erasure, restriction (in German often still called "Sperren", blocking), anonymisation and pseudonymisation are often used interchangeably in daily work, as if they all meant "gone". Under the GDPR, only two of these four routes lead out of data protection law. With the other two, the data remain personal data, with all obligations.

Erasure: the last processing step

The GDPR does not define "erasure" separately. But Article 4(2) lists it among the forms of processing, at the end of the list: "restriction, erasure or destruction". Erasure is itself still an operation on personal data, the last one. Under Article 17(1), the controller has the obligation to erase personal data "without undue delay" where one of the grounds listed there applies, for example where the data are "no longer necessary" for their purpose.

What counts as erased technically is not regulated by the GDPR. In practice, the question is whether the data can still be restored with reasonable effort. A record flagged "deleted" that lives on in every backup is not erased.

Blocking is now called restriction of processing

"Sperren" is the word from the old German Federal Data Protection Act. The GDPR speaks of restriction and defines it in Article 4(3) as "the marking of stored personal data with the aim of limiting their processing in the future". The data remain stored, they are only marked.

Article 18(1) lists the cases in which data subjects can request it: where they contest the accuracy, where the processing is unlawful and they want restriction instead of erasure, where they need the data for legal claims, or while an objection is being examined. Under paragraph 2, the data may then, apart from storage, only be processed with consent or for the reasons listed there, such as the defence of legal claims.

German law adds another case. Where erasure in non-automated processing is impossible or only possible with disproportionate effort and the data subject's interest is low, Section 35(1) of the German Federal Data Protection Act (BDSG) replaces erasure with restriction of processing. The GDPR still applies, because the data are still there.

Pseudonymisation: a safeguard, not an exit

Article 4(5) defines pseudonymisation as processing in such a manner that the data "can no longer be attributed to a specific data subject without the use of additional information", provided that this additional information is kept separately and protected by technical and organisational measures. The name gives way to a number, and the mapping is kept elsewhere.

Recital 26 draws the consequence: pseudonymised data which could be attributed to a person by the use of additional information "should be considered to be information on an identifiable natural person". Recital 28 describes the benefit: pseudonymisation "can reduce the risks to the data subjects concerned".

The CJEU clarified this on 4 September 2025 (C‑413/23 P, on the parallel regulation for EU bodies). For the body that pseudonymised the data and holds the mapping, the data remain personal data despite the pseudonymisation (para. 76). For a recipient who neither has the mapping nor can establish it by other means, they need not be personal data in every case and for every person (para. 86). Pseudonymisation within your own organisation therefore changes nothing about your own obligations.

Anonymisation: only when nobody can attribute any more

Anonymous data fall outside the GDPR. According to Recital 26, the principles of data protection should not apply to data rendered anonymous "in such a manner that the data subject is not or no longer identifiable". The standard is strict: account must be taken of all the means "reasonably likely to be used" by the controller "or by another person", considering costs, the amount of time required and the available technology.

A table without names but with date of birth, postcode and department is therefore often not anonymous. And what cannot be attributed today may well be tomorrow, with a new data set.

The four compared

TermData still present?Still personal data?
Erasurenono
Restriction (blocking)yes, markedyes
Pseudonymisationyes, without direct nameyes, for your own organisation
Anonymisationyes, without personal referenceno, if nobody can attribute any more

The test question

For every line in the deletion concept that says "erase": what happens technically, and is there afterwards still a mapping anywhere, in a table, a backup or an export? If so, it is not erasure but one of the other two routes, and the obligations continue to apply.

In the deletion concept of isidaten, each deletion procedure records whether data are erased, overwritten, anonymised or pseudonymised, and the retention coverage shows where only soft-deleted data still remain.

Sources

Regulation (EU) 2016/679 (GDPR), Article 4(2), (3) and (5), Article 17(1), Article 18(1) and (2), Recitals 26 and 28. Section 35 BDSG (German). CJEU, judgment of 4 September 2025, EDPS v SRB, C‑413/23 P, paras. 76 and 86.

In the next part: TOM, measure, safeguard, and what Article 32 requires compared with Article 46. The previous part: Receiving data does not make you a processor.

All parts of the series

Matching solution isidaten for GDPR

Questions about this update?

Talk to us – we are happy to show you this feature in a demo.