BRAND SAFETY PMAX: EXCLUDING UNWANTED PLACEMENTS

Brand safety in Performance Max is one of the most underestimated aspects when setting up PMax campaigns, and one of those with the most visible consequences in terms of reputation. A PMax campaign not optimized for brand safety can show a company's ads next to controversial content, on YouTube channels misaligned with the brand's values, or on websites that no marketing manager would approve. Keeping your ads brand safe is not a secondary concern: in digital advertising, a single high-profile placement in a damaging context can undo months of brand-building.

The paradox is that Performance Max, by definition, maximizes coverage by automating placement selection across all Google channels. The broader the coverage, the harder it is to guarantee that every individual appearance context is consistent with the brand's identity. This is a structural tension between automation and control that has no complete solution, but it has concrete mitigation tools worth knowing and configuring carefully in all active PMax campaigns on the account.

In this article I analyze the tools available to manage brand safety in Google Ads within PMax: starting from official documentation and flagging where declared features have significant operational limitations that only emerge from direct use on real accounts.

Google's Official Tools for Brand Safety in PMax

According to official Google Ads documentation, brand safety in Google campaigns, including PMax, is managed through three main control levels. Google increasingly frames this under brand suitability, a broader concept than pure safety, covering not just avoiding harmful content but ensuring safety and suitability across the full range of placements where ads appear:

• Content exclusions, filters for sensitive content categories applicable at campaign level
• Inventory type, selection of the advertising inventory type based on the content sensitivity level
• Placement exclusions, exclusion of website URLs, YouTube channels and specific apps at account level via the Brand Safety Center

These three tools also exist for standard Display and YouTube campaigns, but their behavior in PMax has specific characteristics that the documentation describes only partially. The actual level of control in PMax is lower than in traditional Display campaigns: you cannot select specific placements to include, you don't see in real time where ads are being served, and some exclusions have different scopes than you might expect.

What experience shows: many advertisers set up PMax without touching any brand safety settings, leaving Google's default settings active. The default settings, Standard Inventory and no content exclusions beyond those mandatory by policy, are designed to maximize reach, not to protect the brand. For any brand with a corporate, family or ethically specific audience, the defaults are not sufficient.

Content Exclusions in PMax: How to Configure Them and What They Actually Exclude

According to official Google documentation, content exclusions in PMax let you exclude potentially sensitive content categories from your ad placements. The categories available in PMax are:

• Sexual content, adult and sexually suggestive content
• Shocking content, graphic images or descriptions of violence, blood or disgust
• Profanity and crude content, swearing, insults and explicit language
• Gambling content, sites and channels dedicated to gambling
• Sensitive news, coverage of tragic events, conflicts, crimes or disasters
• Political content, content related to political parties, politicians or election campaigns
• Socially sensitive content, divisive topics of social relevance or inappropriate content that falls outside accepted norms for mainstream advertising

Content exclusions are set in the "Settings" section of each PMax campaign, under "Content your ads appear with." They operate at campaign level, unlike the placement exclusions in Brand Safety Center, which act at account level. They can be configured individually, it is not necessary to exclude all categories at once.

A critical point that the documentation doesn't highlight with sufficient clarity: content exclusions in PMax apply to Display and YouTube placements, but not to the search network. Text ads generated by PMax on Search queries are not subject to content exclusions because the search network has no "page content" context in the traditional sense, the context is the user's query, not the site on which the ad appears.

Inference on the effectiveness of content exclusions: on accounts where content exclusions were configured to exclude "Sensitive news" and "Shocking content," a visible reduction in impressions on crime news sites and true crime YouTube channels was observed. The reduction is real but not total: Google classifies content automatically, and classification has errors. Content that objectively falls within an excluded category may not be classified as such by Google, and therefore won't be filtered. Content exclusions reduce exposure to problematic content, they don't eliminate it.

Inventory Type in PMax: Standard, Expanded or Limited

Inventory type in PMax is a setting that defines the range of advertising inventory the campaign can access, based on the content sensitivity level. According to official Google documentation, the three available levels are:

• Expanded Inventory, the broadest level: includes mature, sensitive and some explicit adult content that complies with Google's policies. Maximizes reach but exposes the brand to contexts potentially inappropriate for most businesses.
• Standard Inventory, the default level: excludes explicitly adult content but includes mature and sensitive content considered acceptable by Google. Appropriate for most generalist brands.
• Limited Inventory, the most restrictive level: excludes mature, violent, divisive or sensitive content even when it doesn't violate Google's policies. Significantly reduces available reach but guarantees the highest level of brand protection.

The setting is found in PMax campaign "Settings," under "Inventory type." The default setting is Standard Inventory, but Google doesn't highlight this with a warning during campaign setup. Many advertisers are unaware of this setting and leave the default value active.

When to choose which level: from experience across accounts in different sectors, a more granular practical rule emerges than what the documentation suggests:

• Limited Inventory is the right choice for: brands with family audiences (children, parents), B2B companies with solid corporate identity, brands in the healthcare, education sectors, or with relevant ethical certifications. The reach reduction it brings, averaging 15–25% versus Standard, is an acceptable cost for brand protection.
• Standard Inventory is appropriate for: generalist brands, e-commerce of products for adults without sensitive connotations, campaigns with maximum reach objectives and non-critical brand equity.
• Expanded Inventory is rarely advisable for established brands: the marginal reach increase it offers doesn't justify exposure to contexts that almost no brand would welcome. It is used primarily by advertisers in adult entertainment or gaming.

Inference on inventory type in PMax: on PMax campaigns where inventory type was lowered from Standard to Limited, a reduction in Display and YouTube impression volume of 18–22% on average was observed, but without significant impact on conversion CPA. Nearly all lost impressions were on low-quality inventory, with CTR and conversion rates below campaign average. In terms of value generated per impression, Limited Inventory often produces a better ROAS despite the reduced volume, because the remaining impressions are genuinely safe and suitable contexts where users are more receptive.

Excluding Websites and YouTube Channels from Performance Max

The third control level for brand safety in PMax campaigns is the exclusion of specific URLs and YouTube channels. According to official Google Ads documentation (support.google.com/google-ads/answer/14187974), placement exclusions for PMax are managed through the Brand Safety Center, accessible from Tools and settings → Brand Safety Center → Placement exclusions.

A fundamental aspect that documentation mentions but doesn't sufficiently emphasize: placement exclusions configured in the Brand Safety Center operate at account level, not per individual campaign. This means a domain or YouTube channel that is excluded is excluded from all campaigns in the account, including standard Display, YouTube and Discovery, not just PMax. This should be considered before adding large-scale exclusions.

How to exclude websites from PMax

To exclude sites from Performance Max, the path in the interface is: Tools and settings → Brand Safety Center → Placement exclusions → + Add exclusions. Individual URLs (e.g., "example.com") can be entered or a list of URLs uploaded via file. Google applies the exclusion to the entire domain and its subdomains.

Access to the placement report, which shows where PMax ads have actually been served, is limited compared to standard Display campaigns. Google has progressively introduced greater visibility into PMax placements, but the available aggregated report doesn't have the granularity of traditional Display campaign reports. To identify problematic sites to exclude, the most effective path is: PMax Campaigns → Insights → "Where your ads appeared" section, which shows the main domains by impression volume.

How to exclude YouTube channels from PMax

Excluding YouTube channels from PMax follows the same Brand Safety Center path: in the "Placement exclusions" section you can add specific YouTube channel URLs (e.g., "youtube.com/channel/UCXXXXXXXX") as well as individual video URLs.

A significant limitation: the YouTube placement report in PMax only shows channels with the highest impression volume, not all channels where ads were served. Channels with few impressions don't appear in the report, making a complete audit of YouTube placements impossible without third-party tools.

Inference on placement exclusions in PMax: the most effective practice, from experience on accounts with high brand safety requirements, is to build a proactive exclusion list based on known site categories, not just a reactive one after identifying problematic placements in the report. Lists of low-quality domains (MFA, Made for Advertising sites), known controversial YouTube channels, and mobile apps with poor-quality inventory can be uploaded in advance. Some industry-curated exclusion lists (e.g., MFA lists) are freely available and significantly reduce exposure to problematic inventory before even seeing the first report.

The Real Limits of Brand Safety in Performance Max

Having a realistic view of the limits of the brand safety tools available in PMax is just as important as knowing how to configure them correctly. Google documentation describes available features without quantifying their effectiveness limits, limits that emerge clearly from use on real accounts. Understanding these gaps is essential to ensure brand protection that meets real brand safety standards, not just default settings.

1. You cannot select placements to include

The most significant gap compared to standard Display campaigns: in PMax it is not possible to create placement whitelists, i.e., select the sites and YouTube channels where you want to appear. You can only exclude. The algorithm autonomously chooses where to show ads in the universe of available placements after exclusions, and there is no way to limit distribution to a predefined list of approved placements.

For brands with very strict brand safety requirements, where prior approval of placements is a company policy, this limitation makes PMax incompatible with that level of control. In these cases, standard Display and YouTube campaigns with specific placement targeting remain the only alternative that guarantees pre-approved placements.

2. Google's content classification is not perfect

Content exclusions and inventory type are based on Google's automatic classification of content. This classification occurs via machine learning systems that may incorrectly classify edge-case content. Content that objectively falls within an excluded category (e.g., "Sensitive news") may not be classified as such if Google's system hasn't yet indexed it correctly or if the content was updated after the last classification.

Inference on automatic classification: on accounts with active content exclusions for "Shocking content" and "Sensitive news," impressions were still observed on crime news sites and violent YouTube content in an estimated 5–10% of Display and YouTube impressions. This is not a figure published by Google, but it is an observable pattern by regularly monitoring the placement report and comparing it with active exclusions.

3. Placement visibility remains partial

PMax's placement report only shows the main domains and channels by impression volume, not the complete universe of placements. For campaigns with high budgets and high frequency, the long tail of placements (sites and channels with few impressions each but many in number) remains opaque. These tail placements are often the lowest quality and least relevant for the brand, but identifying them requires external monitoring tools or access to Google Ads APIs for more granular reports.

The practical consequence is that brand safety in Performance Max requires active and recurring monitoring, it is not a "set and forget" configuration. Monthly review of the placement report, combined with periodic updating of exclusion lists, is the operational practice that produces the best results in terms of brand protection over time.