Google's PMax Image Default Flips December 4: What Your Build Process Needs Now

On October 21, the Google Ads Developer Blog announced that starting December 4, 2026, newly created Performance Max campaigns will have image extraction enabled by default across the UI, Editor, and API. That includes accounts where the account-level "Dynamic images" setting is turned off. If your team builds PMax campaigns using templates, Editor uploads, or API automation, the governance model you relied on just changed.

The practical consequence: your landing pages are now part of your ad-creative supply chain unless someone explicitly says otherwise at campaign creation.

What Image Extraction Actually Does

Google describes image extraction as AI that scans a campaign's landing pages for relevant images and adds them to the campaign's creative assets. Those images can then serve across Search, YouTube, Display, Discover, Gmail, and Maps. Until now, if your account-level Dynamic images toggle was off, new PMax campaigns inherited that setting.

That inheritance breaks on December 4. The campaign-level default overrides the account setting for every newly created campaign. Existing campaigns keep their current configuration (Google was explicit about this), so the risk is forward-looking: it's about what happens the next time someone on your team hits "create."

Why This Matters for B2B Ops Teams

If you manage a single PMax campaign, this is a checkbox. If you manage dozens across products, regions, or business units, it's a process gap waiting to bite you.

Brand suitability. Landing pages often carry partner logos, customer screenshots, or team headshots. None of those were designed to be cropped into ad formats. As PPC Land noted, image rights for website display may not cover use in paid ads across YouTube, Gmail, and Display. Google's own developer post asks advertisers to confirm they have the rights or permissions for images used in advertising. That responsibility sits with you.

Configuration drift. The control point moves from one account-level toggle to a per-campaign decision. For teams using API-based campaign creation or Editor templates, the opt-out step has to be baked into the build workflow. Miss it once across a multi-market launch and you've got inconsistent settings that are tedious to audit after the fact.

Scope creep via Final URL expansion. If Final URL expansion is enabled, PMax can route users to pages on your domain other than the one you supplied, widening the set of pages whose images might be extracted. Google's own help documentation describes how Final URL expansion can send users to a different relevant page on the advertiser's domain. URL exclusions become part of creative governance, not just traffic governance.

How to Opt Out (Three Paths)

Google Ads UI: Open the campaign, go to Settings, then Asset optimization, then the Images tile. Uncheck "Landing page images" and save. For new campaigns, do this during setup before you launch.

Google Ads API: Set GENERATE_IMAGE_EXTRACTION to OPTED_OUT in the campaign's asset_automation_settings. The catch: asset_automation_settings is a repeated field. An update replaces the whole array. If you send only the image extraction opt-out, you wipe out your other settings (like TEXT_ASSET_AUTOMATION). Send every setting you want to keep in the same update.

Google Ads Editor: Google's developer post confirms the default-on behavior applies in Editor but doesn't document a dedicated Editor opt-out control. Confirm settings in the UI after any Editor upload.

To audit what you've got, run this query and look for campaigns missing an image extraction entry (those are inheriting the account setting today):

SELECT campaign.id, campaign.name, campaign.asset_automation_settings FROM campaign WHERE campaign.advertising_channel_type = 'PERFORMANCE_MAX'

When to Leave It On

Not every team should opt out. If your landing pages carry current, high-quality product imagery that's fully licensed for ad use, extraction adds creative variety without production cost. The conditional framing from Single Grain gets this right: image extraction works when landing pages are well curated, but opting out is safer when pages include licensed stock, partner imagery, or outdated campaign assets.

The Checklist Before December 4

Treat this as a build-process update, not a one-time fix. Four things to lock in:

• Add a landing-page images check to your PMax campaign creation template or QA doc. Every new campaign, every time.
• If you create campaigns via API, update your code to explicitly set GENERATE_IMAGE_EXTRACTION to OPTED_OUT (and include all other asset_automation_settings entries).
• Audit landing-page imagery as if it were part of your ad-creative library: brand fit, cropping resilience, licensing for paid distribution.
• Review your Final URL expansion settings and URL exclusions. If expansion is on, know which pages (and which images) are in scope.

The trade-off is clear: default-on extraction trades creative control for velocity. For teams with strict brand systems or complex licensing, the velocity isn't worth the cleanup. And the real risk isn't that the control doesn't exist. It does, in every surface Google offers. The risk is operational drift, one missed checkbox compounding across campaigns until someone spots an off-brand image serving on YouTube and starts asking questions nobody wants to answer.