What is profiling?
When does personalisation become profiling?
The GDPR defines profiling as automated processing that uses personal data to evaluate personal aspects of a natural person – in particular to analyse or predict work performance, economic situation, health, personal preferences, interests, reliability, behaviour, location or movements. In marketing, examples include affinity and lead scores or predictions of who is likely to cancel.
For direct marketing, the balancing of interests under Art. 6(1)(f) GDPR is often the legal basis. According to the German supervisory authorities, however, intrusive measures – automated selection procedures for creating detailed profiles, behavioural predictions or analyses that lead to additional insights – tend to indicate that the interests of the data subjects prevail; in that case consent is generally required (DSK guidance on direct marketing, February 2022).
Data subjects can object to processing for direct marketing at any time; this also applies to related profiling (Art. 21(2) GDPR). Art. 22 GDPR protects against decisions based solely on automated processing that have legal or similarly significant effects. Where personal aspects are evaluated systematically and extensively, a data protection impact assessment must be considered (Art. 35 GDPR).
The selection “Tariff ends in 60 days” uses a contract date and is comparatively unintrusive. A model that predicts each customer’s likelihood of switching from usage and response data, on the other hand, is profiling – with higher requirements for legal basis and transparency.
Distinction
| Term | Difference |
|---|---|
| Segmentation | Division by existing attributes, without a new evaluation |
| Automated individual decision-making (Art. 22 GDPR) | Decision with legal or similarly significant effect; profiling can be its basis |
| Hyper-personalisation | Marketing approach; may include profiling, but does not have to |
How the PBM Campaign Platform supports it
Personalisation rules are readable expressions; lead scoring is an explainable points model; for AI functions with predictive control, transparency, human oversight and the ability to switch them off are provided for; every AI action is logged.
Related pages
-
Platform
Consent & compliance -
Trust Centre
Trust Centre -
Blog In preparation
[BLOG: Hyper-personalisation and the GDPR: legal bases, purpose limitation, profiling]