What Are Google Ads Lookalike Audiences and How to Use Them

Google Ads Lookalike Audiences help you reach people who share traits with your best customers. Instead of guessing who might be interested in your product, Google finds users with similar behaviors, interests, and demographics. This expands your reach while keeping your targeting precise. Lookalike audiences are built from your existing data like website visitors, past purchasers, or engaged users. With machine learning, Google identifies patterns and finds new potential customers more likely to convert.

Inside Google Ads, these audiences are called Lookalike segments, and they run only in Demand Gen campaigns across YouTube, Discover, and Gmail. Since March 2026 they also behave differently by default, which this guide covers alongside the setup steps.

Lookalike Audience vs. Similar Audiences: What’s the Difference?

Similar Audiences were Google’s older, auto-generated way to reach people who resembled your remarketing lists, and Google retired them in 2023. Lookalike segments replace that idea inside Demand Gen campaigns. You choose the first-party seed lists, set how closely the audience should match them, and create each segment yourself during campaign setup.

Before Lookalike segments, Google offered Similar Audiences to help marketers expand their reach to consumers who behaved like those on existing remarketing lists. Google generated these lists automatically, and while marketers enjoyed broader exposure, they also had limited ways to customize them. As privacy expectations and third-party cookie limitations reshaped ad targeting, Google announced the sunset of Similar Audiences in November 2022. It stopped creating new similar segments in May 2023, and by August 2023 it had removed them from all campaigns. Nowadays, Google’s Lookalike Audience focuses on first-party data from channels you control, like your CRM systems and website traffic. Google Ads pinpoints individuals who share traits and interests with your best customers. And since the data is collected with consent, your customer targeting aligns with privacy standards. The Lookalike Audience also lets you customize how closely you want to match your seed users. You're no longer stuck relying on cookie-based assumptions, so your ad spend targets people more likely to convert.

What Changed for Lookalike Audiences in 2026

Starting March 2026, Google treats Lookalike segments in Demand Gen as signals for its AI by default. Your seed lists and reach setting guide delivery, and the system can reach likely converters outside the original similarity thresholds. Advertisers who prefer strict targeting can opt out through a Google form and keep the original behavior.

Under the new default, Google’s help documentation explains that your seed list acts as a signal for Google’s models, so the system may reach highly qualified customers outside the Narrow, Balanced, or Broad thresholds. The reach setting still matters, but it guides the AI toward conversions without capping the audience. In reporting, a signal tag appears next to the Lookalike segment row, and performance gains from the change are reported in that row.

The update applies only to Lookalike segments attached to Demand Gen campaigns in Google Ads. As Search Engine Land reported in February 2026, the change trades some targeting control for scale, so it is worth testing its effect on CPA, reach, and conversion volume. If you rely on tight audience boundaries, Google offers an opt-out form, with removal expected in about three days. Opted-out segments keep working as a strict targeting constraint.

How to Create a Google Ads Lookalike Audience

Lookalike segments are created inside a Demand Gen campaign, from the audience section of an ad group. You pick up to 10 seed lists, choose the countries to target, set a reach level, and save. Google recommends building the campaign and segment two to three days before you want ads to start serving.

Creating a Lookalike Audience in Google Ads is a straightforward process that allows you to reach new customers similar to your existing ones. Follow these steps to set up and optimize your campaigns for better performance.

1. Start a Demand Gen Campaign

First, sign in to your Google Ads account and check you have the right permissions to manage audiences and campaigns. Make sure your account is properly set up and linked to any relevant Google services, such as Google Analytics, to fully utilize audience data. Then create a Demand Gen campaign following the usual steps. Lookalike segments can only be built during Demand Gen setup. You can view them in Audience Manager afterward, but you can’t create, edit, or delete them there.

Give yourself some lead time. Google recommends creating your campaign and Lookalike segment two to three days before you want ads to serve, so the segment has time to populate.

2. Choose Your Data Segment

Select the audience lists you plan to use as your seed. Draw from your CRM or remarketing lists of high-value users like purchasers, frequent site visitors, or active email subscribers. Google accepts website visitors, app users, customer lists, Google Analytics audiences, and YouTube audiences as seed lists, and you can combine up to 10 of them in one segment. Consider segmenting your seed audience based on specific behaviors or purchase history to create more targeted Lookalike Audiences. The more defined your seed data, the more accurately Google can find users who resemble your desired customer profile.

3. Create the Lookalike Audience

At the ad group level, add an audience and select “New Audience.” Open the “Add or create Lookalike segment” search box, select “New Segment,” and add your seed lists. Enter the countries or regions you want the segment to target, then name your new audience for easy tracking (e.g., “Balanced Lookalike - 30-Day Purchasers”). Be thoughtful with the naming conventions of your audiences to keep your campaigns organized, especially if you plan to create multiple Lookalike Audiences.

4. Define Your Audience Expansion Level

Google Ads offers Narrow, Balanced, or Broad expansion. Narrow closely mirrors your seed audience for precise targeting, aiming for about 2.5% of users in your target location. Balanced aims for about 5%, and Broad aims for about 10%, sacrificing some accuracy for wider reach. Experiment with different expansion levels so that you can find the optimal balance between reach and relevance for your specific marketing goals. Keep in mind that a wider audience might lower your cost per impression but could also reduce conversion rates if the audience is less targeted. Balanced sits in between and is a practical place to begin. You can adjust later based on campaign data.

Under the 2026 default, these levels act as guidance for Google’s AI instead of hard limits. If you opted out, they keep working as strict reach boundaries.

5. Save and Run Your Demand Gen Campaign

Select Save, and the segment attaches to that ad group. Verify that your ad creatives and messaging align with the interests and behaviors of your Lookalike Audience to maximize engagement. Regularly monitor the performance of your campaigns and make necessary adjustments to bids, budgets, or ad content to improve results.

Keep the segment inside Demand Gen. If you add a segment containing a Lookalike audience to another campaign type, Google shows it as Eligible - Limited, because Lookalike segments can’t serve outside Demand Gen. Once live, Google refreshes the segment automatically every one to two days based on your available customer data.

Best Practices for Using Google Ads Lookalike Audience

Lookalike results depend on the seed lists you feed Google and on how cleanly you structure the campaign. Start from your most valuable customers, segment them by behavior, test expansion levels, and pair Lookalikes with retargeting or custom segments. Under the 2026 default, seed quality matters even more because it steers Google’s AI.

To maximize the benefits of Google Ads Lookalike Audiences, it’s important to implement specific strategies that can help you improve your targeting.

Use High-Quality Seed Data

Prioritize your most engaged and profitable users as your seed. Those who have made purchases or shown strong brand interest produce more accurate lookalike groups. If you opt out of the 2026 signal model, your seed lists together need more than 100 active matched people, and Google says that minimum no longer applies under the new default. Size alone won’t carry a segment, though. Remember that the quality of your seed data directly influences the effectiveness of your Lookalike Audiences. By carefully selecting high-value users, you increase the likelihood of reaching new prospects who are genuinely interested in your products or services, and prevent overspending.

Segment Your Data

When you understand consumer behavior, you can focus on specific segments and customize your messaging to address the unique needs and preferences of each group. Divide your audience based on demographics, buying habits, or site behavior to create highly relevant ads. Also, subdivide seed users into smaller groups and tailor campaigns to each one. This improves engagement, boosts conversion rates, and refines return on ad spend. Also, consider using performance creative strategies to make your ads even more relevant.

Test Different Expansion Levels

A Narrow level suits niche offerings where you need precision, while a Broad level goes further if you’re promoting a more general product and need scale. Run tests to see which settings deliver the best results, and keep refining over time. Adjusting expansion levels allows you to find the sweet spot between audience size and relevance. Regular experimentation helps you understand how different settings impact your campaign metrics, and enables you to optimize for both performance and cost-effectiveness. Because the levels now guide Google’s AI by default, compare results against your pre-March 2026 benchmarks before drawing conclusions.

Combine Lookalike Audience with Other Targeting Methods

You can amplify your campaign's effectiveness by combining the Lookalike Audience with retargeting or custom audience tactics. Retargeting focuses on people who previously showed interest, while the Lookalike Audience finds new, promising leads. When these methods are combined, they create smarter and more cohesive campaigns. Incorporate multiple targeting strategies to ensure that you cover all stages of the customer journey. By reaching out to both new and existing customers, you create a comprehensive marketing approach that maximizes your reach and reinforces your brand message across different channels. If you also run Meta, our guide to Meta Lookalike Audiences covers how the same idea works there.

Keep different audience types in separate ad groups so reporting stays clean. If you want to measure Lookalike performance on its own, remove other audiences from the ad group and turn off Optimized Targeting. If you leave Optimized Targeting on, watch how it interacts with the new Lookalike signal in your reports.

Common Mistakes to Avoid

Weak seed data causes most Lookalike problems, followed by messy structure and too little testing. Keep seed lists current and clean, segment users by behavior, avoid overlapping audiences, and check performance regularly. Remember that Lookalike segments only serve in Demand Gen, so plan them into that campaign type from the start.

Although Google Ads Lookalike Audiences can significantly improve your campaigns, it’s important to be aware of what could go wrong so you can avoid it:

• Seed audience too small: Relying on a weak or outdated seed audience hampers the accuracy of your lookalike audiences. To achieve better targeting, gather data from multiple channels such as email lists and recent conversions.
• Low-quality data: If your data contains errors or inaccuracies, it leads to flawed lookalike results. Regularly clean your lists to remove duplicates or incorrect information.
• Poor segmentation: Combining all your users into one big pool dilutes your message. Instead, segment your audience based on user traits such as browsing patterns or past purchases.
• Overlapping audiences: Audience overlap can waste advertising spend and muddle performance insights. Use exclusions and combinations to keep your targeting sharp and avoid serving ads to the same users multiple times.
• Lack of testing: Without A/B testing, you may miss opportunities to optimize your campaigns. Try different audience settings, ad creatives, and offers to find what resonates best with your audience.
• No ongoing monitoring: Skipping regular performance checks makes it difficult to pinpoint what’s effective. Track key metrics and make adjustments to address engagement or performance issues in your Google Ads campaigns.
• Using Lookalikes outside Demand Gen: Adding a Lookalike segment to other campaign types leaves it Eligible - Limited, so it won’t reach anyone there.

How to Measure Lookalike Audience Performance

Judge Lookalike segments on conversion outcomes such as CPA, ROAS, and conversion value, and read them in their own ad groups so results stay separate. Under the 2026 default, look for the signal tag on the Lookalike row in reporting, since Google credits gains from the new model to that row.

Compare each Lookalike ad group against your other prospecting audiences on the same conversion goal, and review audience insights to see who Google is actually reaching. When a segment underperforms, try a different seed list or reach level before retiring it. For the wider picture of competing efficiently on Google, see our guide to beating competitors on Google Ads without overspending.

Key Takeaways

• Google Ads Lookalike segments run only in Demand Gen campaigns and are created during campaign setup, from an ad group’s audience section.
• Seed lists can combine up to 10 first-party sources, including customer lists, website visitors, app users, Google Analytics audiences, and YouTube audiences.
• Narrow, Balanced, and Broad aim for about 2.5%, 5%, and 10% of users in your target location.
• Since March 2026, Lookalikes act as AI signals by default, and an opt-out keeps the original strict targeting.
• Seed quality is the biggest lever you control, so start with recent, high-value customers.