Lookalike Audience
An audience segment created by advertising platforms that mirrors the characteristics of your existing customers using machine learning algorithms that analyze hundreds of behavioral, demographic, and interest signals. The source audience (called a seed) can be based on customer lists, website visitors, or app users. Platforms like Meta offer percentage-based sizing from 1% (most similar) to 10% (broader reach), where smaller percentages yield higher match quality but smaller audience pools. Google offers Similar Audiences and Optimized Targeting, while LinkedIn provides Predictive Audiences with analogous functionality. The minimum seed size for effective lookalikes is typically 1,000-5,000 users.
Why It Matters
Lookalike audiences let you scale prospecting campaigns efficiently by finding people who resemble your best customers. They typically outperform broad targeting by 2-3x on conversion rate because the platform uses machine learning to identify high-value behavioral and demographic patterns. The quality of your seed audience directly determines lookalike performance: a seed built from high-LTV repeat purchasers will generate better results than one built from all customers. Refreshing seed audiences quarterly ensures the algorithm adapts to evolving customer profiles. For best results, test multiple lookalike sizes (1%, 3%, 5%) simultaneously and compare CPA and ROAS across each to find the optimal balance between audience quality and scale.
Example
A DTC brand uploads a list of 5,000 high-value customers to Meta Ads and creates a 1% lookalike audience of 2.3 million users. This lookalike audience converts at 2.8% compared to 0.9% for interest-based targeting, tripling campaign efficiency. The team then tests a 3% lookalike (5.8 million users) and a 5% lookalike (9.5 million users). The 3% lookalike achieves a 2.1% conversion rate at a CPA only 15% higher than the 1% audience, making it the best balance of scale and efficiency. By allocating 60% of prospecting budget to the 1% and 3% lookalikes, the brand reduces overall prospecting CPA by 40% compared to the previous interest-targeting approach.
Related Terms
Retargeting
A digital advertising strategy that serves ads to users who have previously visited your website or interacted with your content, using browser cookies, pixels, or platform-specific identifiers to build audience lists. Retargeting can be site-based (targeting all visitors), page-based (targeting visitors of specific product pages), or action-based (targeting users who added to cart but did not purchase). Common retargeting platforms include Google Display Network, Meta, and programmatic DSPs. The typical retargeting window ranges from 7 to 90 days depending on the purchase cycle, with most conversions occurring within the first 7 days of the initial visit.
DSP
A Demand-Side Platform is software that allows advertisers to buy digital ad inventory programmatically across multiple ad exchanges in real time. It provides centralized bidding, targeting, and campaign management across display, video, native, audio, connected TV, and digital out-of-home formats. DSPs use real-time bidding (RTB) to evaluate and bid on individual impressions in under 100 milliseconds. Major DSPs include The Trade Desk, Google DV360, Amazon DSP, and MediaMath, each offering different inventory access, audience data integrations, and optimization algorithms.