Meta 广告
Lookalike Audiences
基于现有客户数据生成相似受众，找到与高价值用户行为相近的潜在新客群。
主要用途
Lookalike Audiences extend prospecting by letting Meta find people statistically similar to a source audience you provide. You pick the source, such as a customer list, purchase Pixel event, app activity, or page engagers, choose a region, and set a size from 1 percent to 10 percent of that region's population, where smaller percentages are more similar and larger ones trade precision for reach. Meta models shared characteristics from its behavioral signals rather than copying list attributes. Source quality drives everything: five hundred recent purchasers usually outperform fifty thousand page viewers. Since Advantage+ audience options arrived, lookalikes increasingly act as guiding signals that Meta may expand beyond, rather than hard boundaries.
所在链路
Lookalikes operate at the launch stage as a prospecting input in Ads Manager, built on the event and customer data collected through the Pixel and Conversions API.
核心功能
- Lookalike modeling from customer lists, Pixel events, or engagement
- Adjustable similarity size from 1 to 10 percent per region
- Multiple stacked ranges from one source for tiered testing
- Automatic refresh as the source audience updates
- Compatible with exclusion lists for clean prospecting
适合谁用
- Advertisers with a meaningful base of customers or high-intent events
- Ecommerce brands scaling beyond retargeting pools
- Teams testing structured prospecting tiers against broad targeting
新手提示
- Build from your highest-intent source available, ideally purchasers from the last 90 to 180 days, not from page likes or video views.
- Test lookalikes against broad Advantage+ targeting before assuming they win; on large accounts broad now often performs as well or better.
- Exclude existing customers and the source audience itself, or your prospecting spend quietly turns into retargeting.