The Shift to AI Control: Navigating Changes to Demand Gen Lookalike Audiences
by Joe Smith, Kinase Video Growth Manager
In a move that mirrors the broader industry evolution across Google and Meta, we are seeing another shift away from advertiser targeting control toward (we hope) smarter AI-driven systems.
This is the latest update to Demand Gen Lookalike (LaL) audiences. From 15th March these will move away from the existing three similarity thresholds, to using the original audience as a seed list signal for Google AI to reach an audience it identifies as in line with your campaign goals.
At Kinase, our focus is on managing and getting the most from these changes. Here is how we are advising brands to navigate this transition in both the short and long term.
Short Term: The "Test, Test, Test" Phase
Because what works for one brand rarely works for all, our immediate priority is testing the impact of adopting the new Lookalike AI targeting against retaining the existing one.
While some may see an increase in volume, the quality of that traffic is the real question. This change is particularly sensitive for specific sectors where precise targeting is a pillar of the strategy:
B2B: Since audiences are already niche, even a standard 2.5% lookalike can feel broad; further AI expansion may dilute this further.
Lead Gen (without lead scoring): You might see a surge in lead volume, but a potential degradation in quality as AI may not yet distinguish between a "lead" and a "qualified lead".
Luxury Brands: These brands rely on specific income signals and exclusivity; AI might find users with similar browsing habits who lack the disposable income to convert.
Strict CPA Targets: Brands operating with tight performance targets may benefit from opting out initially to see how the broader market reacts.
Note: If you choose to opt out of these changes (and thus enable testing), you currently need to fill out a Google form for each specific ad group.
Long Term: Content is King (Again)
This shift serves as a wake-up call: creativity is the cornerstone of performance. As targeting becomes more automated, the "who" is handled by the machine, making the "what" (your creative) the primary lever for success.
Across Social, YouTube, and Demand Gen we are already finding that a diverse, creatively strong mix of content leads to drastic performance improvements. To bridge this gap, Kinase is focusing on:
AI-assisted creative
Influencer partnerships and UGC
Robust creative testing frameworks
The Bottom Line
These changes are here to stay. The most successful brands will be those that stop fighting for manual control and start mastering the art of testing and adapting to an AI-led landscape.
