Direct Answer: Five Attributes, US Scope
Google's publisher policy text on privacy-related standards states that in the US and Canada, personalized ads for credit-related products and services may not be selected or targeted by an audience's gender, age, parental status, marital status, or postal code. The policy names credit cards as an example alongside credit and loans, bank and checking accounts, and debt-management products. This article addresses the US scope only; Canadian requirements need separate review.
That is a targeting restriction, not a forecast that a specific campaign will be rejected or an account suspended. The retrieved text notes possible account-level consequences for policy violations but does not establish campaign-level enforcement outcomes.
Scope Test: Is Your Offer Inside the Covered Category?
Do not reduce the rule to “credit card.” The covered category includes credit-related products and services, bank products and services, and some financial planning and management services. The examples include credit cards and loans, bank and checking accounts, and debt-management products. A co-branded card, secured card, or card-plus-loan campaign still sits close to those named examples.
Teams often inherit plans labeled “card comparison,” “balance transfer,” or “bank account” and treat each label as a separate compliance universe. The retrieved text does not draw those fine distinctions. For an ambiguous offer, verify the live English policy and ask the platform whether the product is covered. Record the reasoning for the next refresh.
Attribute-by-Attribute Check
The clause works best as a line-item audit, not a vague instruction to avoid demographics. Map each audience layer against the five named attributes. The table separates that clause from the sensitive-information rule.
| Audience attribute or signal | Status in personalized US credit card ads (per retrieved policy text) | Rule basis |
|---|
| Gender | Not permitted | US/Canada credit clause — named attribute |
| Age | Not permitted | US/Canada credit clause — named attribute |
| Parental status | Not permitted | US/Canada credit clause — named attribute |
| Marital status | Not permitted | US/Canada credit clause — named attribute |
| Postal code (ZIP code in the US) | Not permitted | US/Canada credit clause — named attribute |
| Low credit rating or high debt burden signals | Not permitted | Sensitive-information rule — separate from the US/Canada credit clause |
All five attributes have the same status. No credit-specific exception appears in the retrieved text. The only carve-out visible in the section is the employment bullet, for certain pre-identified US government advertisers relying on bona fide occupational qualifications. The credit bullet states no comparable exception. A business justification, small audience, or geographic relevance argument does not create one.
Postal code deserves special care because “ZIP code” is the US label teams use. The clause names postal code directly, so ZIP is not a harmless localization layer. Remove or rebuild a legacy audience containing ZIP layers rather than relabeling it.
The Second Rule People Miss: Sensitive Financial-Status Signals
Removing age or ZIP targeting does not make a credit-card audience compliant by itself. Google's personalized-advertising rules separately prohibit selecting or targeting personalized ads with inferred or actual sensitive information, including negative financial-status information such as signals that a user has a low credit rating or a high debt burden. That rule is distinct from the five-attribute clause.
Ask two review questions. First, does the audience definition use gender, age, parental status, marital status, or postal code? Second, does it derive from sensitive financial-status signals, including credit-seeking or hardship behavior implying a low credit rating or high debt burden? A list built from “credit repair seekers,” “debt relief visitors,” or similar content can fail the second test after every demographic and ZIP layer is removed. The retrieved text gives no complete list of acceptable sources, so edge cases need platform guidance or qualified review.
A neutral segment name is not a cure. “Prime prospects” may still encode financial status in the underlying data. Review the list source, inference logic, and cookie or partner-data permissions.
Documentation: Data Rights, Ad Notice, and Self-Regulatory Alignment
For personalized advertising, the policy requires all necessary rights to use audience data such as cookie lists. It requires a disclosure in the ad noting that the ad is interest-based, for example the AdChoices icon. Parties must also follow applicable industry self-regulatory principles, such as the Digital Advertising Alliance's principles for online behavioral advertising.
These duties sit alongside the attribute checks. A team can remove age and ZIP layers yet lack rights to a purchased list, omit the ad notice, or fail to document self-regulatory alignment. Keep list provenance, permission records, ad-notice placement, and the framework relied on in the pre-launch file. Check the current live version of the linked disclosure resource.
Pre-Launch Audit and Verification Limits
Make the check a recurring gate. Confirm the offer's category; inventory every audience layer, including inherited ones; mark each against the five attributes; review list sources for negative financial-status signals; then check data rights, the interest-based ad notice, and self-regulatory alignment. Keep the evidence with the launch record.
The hard parts are usually scope and hidden data. A co-branded, secured, or card-plus-loan offer may need platform confirmation. A legacy ZIP layer may have been added for localization, but the clause states no localization exception. The retrieved text is Chinese-language and consolidated, so verify the current English page and interface labels. It does not resolve contextual or placement targeting, Canadian requirements, or US fair-lending, ECOA/Reg B, CFPB, and state-law duties; those need separate review. AdSense program policies say policies may be revised, publishers must keep current, and violations can lead to ad serving being stopped and/or account deactivation. This is policy research, not legal advice.