Google Ads is preparing to change the default behavior of Lookalike segments in Demand Gen campaigns. Search Engine Land reported on October 12 that Lookalikes will move toward a suggestion mode by default, while advertisers who need the original behavior can opt out and keep them as targeting constraints. Google’s help page now frames the same choice clearly: suggestion-led expansion for flexibility, or a targeting constraint when the selected Lookalike audience must remain the delivery boundary.
This is not a small settings note. Lookalikes are built from first-party seed data: customer lists, website visitors, app users and YouTube audiences. If that seed list represents high-value buyers, strict control can protect acquisition quality. If the campaign goal is efficient reach, suggestion mode may give Google’s systems more room to find incremental converters. The wrong default can turn a useful expansion tool into either a volume ceiling or a quality leak.
Start with the business risk, not the toggle
The first question is whether the audience is a boundary or a signal. A boundary is appropriate when the seed represents a narrow commercial truth: recent purchasers, high-LTV customers, qualified leads or users from a regulated market. A signal is appropriate when the brand needs scale and can tolerate exploration outside the seed profile. The same Lookalike may be a boundary in one campaign and a signal in another.
Next, check whether the seed is strong enough. Google says the submitted seed lists need more than 100 active matched people, and Lookalikes can use narrow, balanced or broad reach. A weak or mixed seed list makes the constraint less meaningful. Before preserving strict behavior, remove stale buyers, low-quality leads and users who do not match the campaign’s commercial goal.
A practical control checklist
Use four tests. First, value: does the seed list reflect the customers the business actually wants more of? Second, margin: can the campaign afford broader reach if conversion quality drops? Third, measurement: will reports show conversion value, customer quality or only cheap leads? Fourth, governance: who approves the opt-out choice, and when will it be reviewed?
If the answers are unclear, split the decision. Keep targeting constraints for the highest-value or most sensitive ad groups. Test suggestion mode on lower-risk campaigns with enough budget, a clean holdout period and clear success metrics. Monitor CPA, conversion value, new-customer quality and audience insights, not just impressions or clicks.
The takeaway
Demand Gen automation is useful when teams know what they are handing over. The Lookalike update should push advertisers to document where audience control is still strategic and where algorithmic exploration is acceptable. Do not preserve constraints because control feels safer, and do not accept suggestion mode because scale sounds modern. Make the setting follow the business risk.
Source references
- Search Engine Land: Google Ads gives Demand Gen advertisers a way to keep Lookalike targeting restrictions
- Google Ads Help: Use Lookalike segments in Demand Gen campaigns as a Targeting Constraint
