Lookalike Audience can help you find audience groups that share commonalities with your existing customers. It could help you find audiences that are interested in your product and expand your audience range.
How Lookalike Audience Works
The model will use algorithms to study the attributes of the custom audience you create. For example, we will analyze attributes of the users from the Custom Audience type you selected, such as demographics, location, operating system, interests, etc. Then, we will look for other users and groups who share similar attributes.
The accuracy of the user list produced by the model is based on the audience you provide us. To ensure the model runs efficiently, we recommend a size of 10,000 people for your source audience. It is very important you ensure that the list and attributes you provide are correct.
Another factor the model considers is which option you select for Audience Size, either Specific, Balanced or Broad. The more specific the audience size, the more similar of an audience the model will look for.
Lookalike Audiences Application
You can use Lookalike Audience throughout the different phases of the Ad Group delivery.
Cold start: During this phase, Lookalike Audience can shorten the exploration time and help the algorithms learn more quickly which group of users will be interested in your ads, and the product you are promoting, helping you obtain conversions faster.
Mature delivery phase: When your delivery curve is at its peak and has reached a satisfactory number of conversions, you can use Lookalike Audience to expand your audience further and maximize your ROI.
Decline phase: When your Ad Group delivery is in decline, you can use Lookalike Audience to extend the life cycle of your delivery.