
Data on health, origin, religion, sexual orientation and other sensitive categories has reinforced protection. Platforms block part of the targeting based on it, but there are grey areas. Advertisers should know the limits, avoid delicate inferences and design campaigns that do not expose people.
The essentials in a definition
Targeting with sensitive data (health, beliefs, orientation, origin, among others) is restricted by the GDPR and largely blocked by platform policies, such as Google Ads policies and Meta’s advertising standards. The risk exists even unintentionally: ads that suggest you know something sensitive about the person may be disapproved and generate complaints.
What sensitive categories are
GDPR identifies special categories of data, such as health, racial or ethnic origin, religious and political beliefs, sexual orientation and biometric data. Platforms have their own policies, often going beyond the law, on targeting and related messaging.
There are also categories sensitive for advertising purposes, such as finance, employment, housing and gambling, with specific rules.
What platforms already do
They have removed many options for targeting by sensitive interests, restrict customer lists in certain sectors and review ads that appear to infer personal characteristics. However, some indirect combinations may still suggest sensitive categories.
The rules change. What is allowed today may be restricted tomorrow.

Advertising without exposing sensitive data
- Identify pages and events that reveal sensitive categories
- Do not send that data to platforms
- Write about the service, not the person
- Review policies and take legal advice
The risk of inference
Even without using sensitive data directly, ads and pages that suggest the person has a condition or belongs to a group (“For those who suffer from…”) can expose and upset. This affects trust and may breach policies.
Prefer messages centred on the product or service, not on the recipient’s personal condition.

An ad should not tell people what the company thinks it knows about them.
Good practice
Avoid collecting sensitive data without necessity and a legal basis. Do not send it to ad platforms. Review landing pages and events that may reveal sensitive categories (for example pages about medical conditions).
In sectors such as health and finance, work with legal support and read each platform’s policies carefully.
Decision table
| Practice | Risk | Alternative |
|---|---|---|
| Targeting by health condition | Very high | Target by context and general interest |
| Messages revealing inferences | High | Neutral, useful messages |
| Lists with sensitive data | Very high | Do not upload; minimise |
| General educational content | Low | Prefer and measure |
Common mistakes to avoid
- Text suggesting knowledge of the person. “For you, who have diabetes” may breach policies and create discomfort.
- Using waiting lists or sensitive customer records. Do not share data of that nature with platforms.
- Ignoring restricted categories. Review sector policies before launching.
How to measure the result
Track disapprovals by policy, complaints and the performance of neutral versus direct messages, and keep a list of expressions to avoid.
Limits and exceptions
Platform policies and the law change, and sensitive categories are interpreted differently. A message accepted on one platform may be rejected on another. When in doubt, simplify the text, avoid references to personal conditions and consult the policies and, if necessary, a lawyer. This content is informational and does not replace legal advice.
Illustrative example
Illustrative example: a nutrition clinic reformulates ads that said “If you have weight problems” into neutral messages about healthy habits and booking a consultation. Instead of targeting by condition, it uses search keywords and educational content. Disapprovals disappear and the booking rate holds, without exposing anyone to messages that reveal inferences.
Frequently asked questions
Can I target health interests?
Platforms restrict it. Check policies and validate with legal support.
Are medical-condition pages a problem?
They can be, if events reveal the condition. Review what is sent.
How do I write ads in sensitive sectors?
By centring the message on the service and avoiding statements about the person.
Can I talk about health conditions in ads?
With great care, without inferring the person has the condition and complying with sector policies and law; when in doubt, check the rules.
How do I avoid risks?
Neutral messages, contextual targeting, data minimisation and policy review before launching.
Avoid targeting or suggesting inferences about sensitive data; use neutral messages, context and educational content, and check platform policies and your sector’s law.
Sources and recommended reading
Last editorial review: October 2026. Informational content; always confirm current conditions in the official sources before deciding.