The carefully constructed walls that once defined digital advertising audiences are systematically being dismantled by the very platforms that helped build them, pushing marketers into a new era of automated trust. This year, Google has implemented a fundamental change to how Lookalike audiences function within its Demand Gen campaigns, a move that is less of a minor tweak and more of a philosophical pivot. This update reengineers the relationship between advertiser input and algorithmic output, forcing a profound reevaluation of targeting strategies and the very definition of control in performance marketing. For advertisers accustomed to precise audience curation, this shift from direct instruction to AI-led guidance represents a critical juncture.
Are You Still in Control of Your Ad Targeting?
The central question now facing every digital marketer is how this evolution challenges their traditional role as the primary architect of audience selection. With the new system fully active in Demand Gen campaigns, the days of setting firm targeting boundaries and expecting the platform to adhere to them strictly are over. This change marks a pivotal moment, transforming what was once a clear directive from the advertiser into a collaborative, and at times opaque, process with Google’s artificial intelligence.
This transition forces advertisers to move from a mindset of micromanagement to one of strategic oversight. The core responsibility shifts from defining who to see an ad to defining what a successful outcome looks like, and then trusting the system’s expansive reach to achieve that goal. Consequently, marketers must adapt their skills to focus more on providing high-quality input signals—such as first-party data and conversion tracking—rather than hand-picking the audience themselves.
Understanding the Core Shift from Fence to Compass
Previously, the Lookalike model operated like a “fence.” Advertisers selected a similarity tier—narrow, balanced, or broad—and this choice created a rigid boundary. Ad delivery was strictly confined to the users identified within that specific cohort, providing a clear and predictable targeting framework. This method offered a high degree of perceived control, as marketers knew their budget was being spent on a precisely defined group of users who mirrored their best customers.
In contrast, the new AI-driven model transforms these tiers into a “compass.” The selected Lookalike segment no longer acts as a containment field but as an initial optimization signal. It gives Google’s AI a starting point, a direction in which to begin its search for users likely to convert. The algorithm is now empowered to venture far beyond that initial list, making the Lookalike segment a suggestion for optimization rather than a strict targeting constraint.
Dissecting the New Mechanics of Audience Expansion
This updated functionality works in tandem with Optimized Targeting, an existing feature designed to find new conversion opportunities by reaching beyond an advertiser’s manually selected audiences. When the new signal-based Lookalike is used alongside Optimized Targeting, the system effectively stacks multiple automation signals. This combination grants the algorithm maximum freedom to expand its reach in pursuit of a better cost-per-acquisition (CPA) or higher conversion volume, exploring a much wider audience landscape than ever before.
However, Google has provided a temporary recourse for advertisers who are hesitant to cede this level of control. Through a dedicated request form, marketers can opt out of the new system and retain the legacy behavior of strict Lookalike targeting. This option allows for a transitional period, but it also underscores the platform’s definitive move toward an automation-first future, positioning the traditional method as a legacy feature rather than a permanent alternative.
The Broader Context Why Google is Trading Control for Automation
This update is not an isolated event but part of a larger industry trend toward AI-first audience expansion, drawing clear parallels to similar strategies implemented by platforms like Meta. According to digital marketer Dario Zannoni, there are two primary drivers behind this strategic shift. First, he notes that overly strict Lookalike targeting can artificially limit a campaign’s scale and overall performance by excluding potential converters who fall just outside the defined similarity parameters.
Second, Zannoni highlights the increasing complexity of maintaining high-quality similarity models in a privacy-conscious digital ecosystem. As data signals become more fragmented, relying on broad automation becomes a more efficient and scalable path for platforms to deliver results. By shifting the focus from precise audience matching to predictive conversion modeling, Google can leverage its vast data sets more effectively while navigating the challenges of a changing ad landscape.
A Practical Playbook for Performance Marketers
For many performance marketers, the most immediate concern is the perceived loss of granular control over audience definitions, which has long been a cornerstone of campaign management. The inability to ensure ads are shown exclusively to a hand-picked similar audience introduces a new layer of uncertainty. This anxiety is valid, as campaigns may now reach users who, on paper, seem far removed from the core customer profile.
To navigate this new reality, the focus must shift from resistance to adaptation and rigorous testing. Marketers should initiate A/B tests, comparing the performance of legacy Lookalikes (accessed via the opt-out) against the new signal-based approach. By closely monitoring key metrics—specifically CPA, reach, and incremental conversions—it becomes possible to measure the real-world impact. This data-driven approach allows advertisers to re-evaluate their strategies, learning how to leverage this increased automation to their advantage rather than fighting against it.
This period of adjustment required marketers to build new frameworks for measuring success. The ultimate goal was to determine if the algorithm’s freedom truly uncovered more efficient pockets of conversions that manual targeting had missed. For many, the evidence suggested that embracing this AI-driven expansion, while initially unsettling, led to improved performance and scale, cementing automation’s role at the heart of modern digital advertising.
