Google will enable landing page image extraction by default for Performance Max campaigns created from December 4, 2026. The change, announced on October 21, applies in the Google Ads interface, Google Ads Editor and the API. Existing Performance Max campaigns will retain their current configurations, according to the company.
Until now, a campaign without an explicit choice for this feature inherited the account-level setting. The new behavior enables extraction at campaign level even when dynamic images are disabled in the account. Explicitly disabling the feature for each new campaign remains possible.
Website photographs enter the ad creation process
The feature uses Google AI to identify high-resolution images on the supplied landing pages. These materials can supplement the campaign's visual assets. The operation starts with images already on the website and should be distinguished from generating a new image from a description. In the API documentation, extraction is identified separately through the GENERATE_IMAGE_EXTRACTION automation type.
Google defines Performance Max as a goal-based campaign type that provides access to Search, YouTube, Display, Discover, Gmail and Maps inventory from a single campaign. Images are one creative component, alongside text and video. Ad delivery uses the materials, settings and objectives provided in the account; the change announced for December concerns the origin of visual assets, without replacing these other campaign components.
Assets are organized into groups built around a theme or audience. The system selects and combines images, headlines, descriptions and videos according to the channel on which the ad appears. A group can have one or more final pages relevant to its objective. Google Ads API documentation allows up to 100 groups in a campaign, with each group associated with only one campaign.
Materials within a group can be combined with one another for compatible advertising formats. Google provides previews of image and text combinations, and its advertiser documentation also describes adding assets manually. For retailers using a Merchant Center feed, there is a separate option to build ads from the connected account's materials. Landing page image extraction joins this ecosystem of assets and automation.
The campaign setting is explicit in the API
In an API integration, automation options are represented by pairs: an automation type and its status. The AssetAutomationSetting resource includes the asset_automation_type and asset_automation_status fields. This structure allows a choice for image extraction to be recorded separately from other creative features. Applications managing campaigns therefore use an identifiable configuration at campaign level, rather than an assumption based solely on the account's general preferences.
The OPTED_IN status enables generation and serving for the corresponding automation type. OPTED_OUT stops these operations for that type. The technical reference also lists UNSPECIFIED, for an unspecified status, and UNKNOWN, used as a returned value when the API version does not recognize the status. An explicit decision to disable the feature consequently has a different meaning from the absence of a value in the submitted configuration.
Google warns that updating the asset_automation_settings list replaces the entire list, without automatically merging new entries with existing ones. A change for images must retain the other desired options in the request. This behavior concerns applications that create or update campaigns on advertisers' behalf and explains why the visual setting cannot be treated in isolation from the rest of the automation list.
Updates through the Google Ads API also use a field mask that specifies which properties are being changed. Fields submitted but absent from that mask are ignored. Documentation for the Python library describes comparing the initial and new configurations to construct the mask. The submitted automation list and the update mask have distinct roles: the former contains the choices, while the latter identifies the property to which the change applies.
Extraction, image enhancement and text have different controls
The automation catalog separates image extraction from GENERATE_IMAGE_ENHANCEMENT, which covers adjustments such as automatic cropping and image enhancement. There are also automation types for video variants and text assets. The names and statuses of these features are distinct in the API; a choice for extracting photographs does not, by itself, describe the entire policy for creating and adapting advertising materials.
Text customization can create headlines and descriptions using ads, page content and generative artificial intelligence. Final URL expansion has a separate function: it can select a more relevant landing page and associate text suited to that page. Google documentation treats these mechanisms as campaign settings, alongside options for videos and the company's brand identity.
At asset level, an image uploaded to Google Ads is an object that can be linked to one or more groups. The API manages the file separately from its association with an ad or campaign. ImageAsset resources cannot be changed in place after upload; associations are managed through the corresponding operations. For campaigns with brand guidelines enabled, the company's name and logo are associated at campaign level.
In its announcement about the new setting, Google asks advertisers to check rights and permissions for images published on landing pages, including their commercial and advertising use. The communicated timetable is specific: the new default behavior begins on December 4 for new campaigns. Existing campaign configurations remain as previously established, and explicit control over image extraction continues to be available.
