Personalized ads are no longer just a performance lever. In Google Ads, they are also a policy risk.
Google’s restricted targeting rules for personalized advertising draw a harder line between audiences advertisers build themselves and audiences Google defines inside the platform. That distinction now matters most for brands operating in sensitive categories, as well as advertisers promoting housing, employment, and consumer finance offers in the United States and Canada.
The Audience Source Now Matters As Much As The Campaign Goal
Google’s policy separates audience tools into two broad buckets.
Advertiser-curated audiences include lists and segments built from a brand’s own data or audience logic. Customer Match, your data segments, audience expansion, and lookalike segments fall into the restricted side when campaigns promote products or services in sensitive interest categories.
Predefined Google audiences are treated differently. In-market segments, affinity audiences, life events, and some demographic targeting can still be used in sensitive categories because Google says those audiences are configured without sensitive user signals.
That distinction is not cosmetic.
For advertisers, it changes how campaign teams need to evaluate targeting. A campaign may have compliant ad copy, a compliant landing page, and a legitimate business goal, yet still run into restrictions because the audience layer contains a list or segment Google does not allow for that category.
Health is the obvious example. So are alcohol, gambling, birth control, clinical trial recruitment, political affiliation, religious belief, sexual orientation, race and ethnicity, relationship hardship, negative financial status, and other sensitive interest categories listed in Google’s policy.
The policy also states that users under 18 are not eligible for personalized advertising of any kind. Advertisers using advertiser-curated audiences are also prohibited from uploading customer information collected from viewers of child-directed content.
That narrows the margin for sloppy list hygiene.
Customer Match Is Powerful, But Not Universal
Customer Match has become a central tool in paid media because it lets advertisers use first-party customer data across Search, Shopping, Display, YouTube, and Gmail campaigns. Google’s separate Customer Match policy confirms that access depends on account standing, payment history, and, for some features, account history and lifetime spend.
The same policy also applies sensitive category restrictions to Customer Match.
Advertisers cannot use Customer Match to target ads to users or promote products and services tied to sensitive interest categories. Google also says customer lists cannot be combined with other targeting criteria in a way that produces an overly narrow or specific audience.
That creates a practical problem for performance teams.
Many modern Google Ads accounts are built around layers: first-party lists, geographic limits, conversion signals, exclusions, automated bidding, and campaign-type automation. In ordinary retail or B2B campaigns, those layers can sharpen efficiency. In restricted categories, the same setup can trip a policy review if it creates a sensitive or overly narrow targeting pattern.
The issue is not whether first-party data is useful. It is whether the campaign context allows it.
That distinction connects with Google’s broader push toward automated advertising systems. Recent changes, such as Google Ads’ new prospects mode show how much audience definition now depends on exclusions, signals, and platform-side interpretation rather than a simple targeting list.
Housing, Employment, And Finance Face A Separate Line
Google treats housing, employment, and consumer finance differently from sensitive interest categories.
These “Access to opportunities” rules apply in the United States and Canada. Their purpose is narrower: limiting personalization in categories where targeting could affect access to important life opportunities.
Under the policy, advertisers in these categories cannot target audiences using gender, age, parental status, marital status, or ZIP codes in the U.S. and Canada. Predefined Google audiences are allowed, but not when they rely on the restricted demographics.
Location targeting is not banned outright. Radius, city, and country-based targeting remain allowed. Radius targeting must be set at least one kilometre around a given location. In Canada, Google says the first three letters of a postal code, known as the forward sortation area, can be used for ad targeting.
That matters for local advertisers.
A mortgage broker, recruitment firm, apartment developer, or credit provider may still run geographically relevant campaigns. What they cannot do is use restricted demographic controls or ZIP code targeting to shape who sees opportunity-related ads.
The policy also instructs advertisers flagged under these categories to set gender, age, parental status, and marital status to “Enable” and remove ZIP code targeting.
For agencies managing cross-border accounts, Canada and the U.S. cannot be treated as generic North American targeting regions. The restriction is specific, operational, and easy to miss during campaign duplication.
Demand Gen And Discovery Can Run Into The Rule Faster
Google notes that Discovery and Demand Gen campaigns use advertiser-curated audiences by default and may be restricted from serving when the promoted products or services fall within sensitive interest categories.
That single detail could affect how advertisers troubleshoot disapprovals or limited serving.
A Search campaign using predefined in-market segments may behave differently from a Demand Gen campaign built around first-party lists, custom segments, or engagement audiences. The landing page may be the same. The offer may be the same. The policy outcome can still differ because the campaign type leans on a restricted audience setup.
Demand Gen has become more important as Google consolidates visual and discovery-style inventory. TechWyse recently reported on Google Display Ads moving into Demand Gen, a shift that gives advertisers more reach inside a campaign type where audience configuration already carries policy implications.
Custom segments also require closer review. Google’s policy says custom segment audiences using sensitive creative assets or pointing to sensitive landing pages will serve only with Display campaigns to non-sensitive audiences or contextually. Other campaigns using custom segments with sensitive creative assets or sensitive landing pages will not be eligible to serve.
That is a technical sentence with a blunt result.
The creative, the landing page, and the audience source are all part of the policy review surface.
Data Collection Is Part Of The Same Compliance Chain
Google’s restricted targeting policy does not stop at campaign settings.
The data collection section prohibits ads from collecting or containing personally identifiable information unless a Google-provided ad format is designed for that purpose. It also bars advertisers from using PII with anonymous or pseudonymous information, sharing PII with Google through data segment tags or product feeds, or sending precise location information without first obtaining user consent.
Advertisers are also told not to use data segments in ways that target an overly narrow audience.
That last point is easy to overlook because it sits between privacy and targeting. It matters because many accounts use audience layering to compensate for rising media costs, weaker cookie signals, or lower-quality lead flow. Narrowing an audience too aggressively may improve short-term efficiency in some categories, but it can also create a policy problem.
Practical implications for marketers and SEOs: campaign teams need to review audience sources before treating performance issues as bidding or creative problems. For restricted categories, Customer Match, remarketing lists, lookalike-style segments, and audience expansion should be checked at the campaign, ad group, and asset group level. Landing pages also need review, because sensitive content can affect whether a custom segment is eligible to serve. In regulated or opportunity-related categories, privacy documentation, consent handling, and geographic targeting settings should be audited alongside ad copy.
The compliance layer is also becoming harder to separate from measurement. Google’s move to unify consent controls across Analytics and Ads, covered by TechWyse in Google Analytics and Ads consent controls, puts more pressure on advertisers to understand when data can be used for personalization and when it cannot.
Policy Fixes Are Campaign Work, Not Paperwork
Google gives advertisers several routes when ads are flagged.
For sensitive interest categories, advertisers can remove advertiser-curated audiences from ad, ad group, asset group, and campaign targeting. They can also edit or remove policy-triggering content from the ad, site, or app. If they believe the decision is wrong, they can appeal from the Google Ads account.
For access-to-opportunity categories in the U.S. and Canada, the targeting fix is more specific: enable all restricted demographics and remove ZIP code targeting.
Those fixes sound simple on paper. In larger accounts, they are rarely one-click changes.
Audience lists may be shared across campaigns. Asset groups may inherit audience logic from old builds. Campaigns migrated into newer formats may carry targeting assumptions from earlier structures. As Google Ads continues adding automation, including features covered in TechWyse’s reporting on AI Max controls and campaign migration, policy troubleshooting increasingly depends on knowing where the targeting signal actually enters the account.
Google’s documentation places responsibility on advertisers to ensure compliance with laws and regulations wherever ads are served. The platform may review, restrict, or require corrections, but the setup begins inside the account.
For restricted categories, the cleanest campaign is not always the most aggressively targeted one.
