Leverage Lookalike Models and Predictive Audiences

When you need to reach new users that look like your best customers, create Predictive Audiences powered by Lytics Lookalike Models. Campaigns run with Predictive Audiences improve customer engagement, drive higher conversion rates, and reduce ad spend.

Using Predictive Audiences can benefit marketers in a number of ways:

• Helps you define better targeting criteria based on behavioral data rather than demographics or third-party data.
• Enables you to optimize each audience for reach or accuracy based on campaign goals.
• Deliver use cases that drive conversions across the customer lifecycle.

Better targeting criteria

Predictive Audiences simplify your targeting strategies by eliminating the need to manually define parameters to sort your customers. Lytics gives you direct insight to help answer the question, "how do I know what to target users on?" Each Lookalike Model in Lytics has a Feature Importance chart that shows the most important attributes that predict user conversions for your source and target audiences. These attributes include Lytics behavioral scores, content affinities, and other user fields.

In the example above, Lytics created a Lookalike Model to understand which of our documentation site users are mostly likely to move from "Casual Visitors" to "Deeply Engaged Users". As you can see in the chart, our model determined that "Total Pageview Count", "Quantity", and an affinity for "marketing" are the top three attributes of users mostly likely to become deeply engaged. Your Lookalike Models may simply confirm assumptions, but in others cases, they can offer a new understanding of your customers' behaviors and which factors really influence desired actions.

Optimize based on goals

Once a Lookalike Model is built, you can easily create different Predictive Audiences to optimize the balance between reaching more people and targeting your best customers for each campaign. When reach is a priority, you can build Predictive Audiences with higher reach but lower accuracy to target more users. When Return on Ad Spend (ROAS) is a priority, you can build Predictive Audiences with higher accuracy but lower reach to target only the users mostly likely to convert.

For example, if you're targeting single purchasers to encourage them to become repeat purchasers, accuracy may be more important than reach. Learn more about how to optimize the trade-off between accuracy vs. reach.

Use cases across the customer lifecycle

Predictive Audiences help you engage and convert customers across their lifecycle, from anonymous users visiting your website for the first time, to high lifetime value (LTV) customers that are brand advocates. Below are common examples of how to build Lookalike Models that power Predictive Audiences through different stages of your marketing funnel.

Earlier stages:

• Which of my anonymous users are most likely to convert to known users?
• Which of my users with free trials are most likely to convert to subscribers?

Mid stages:

• Which of my single purchasers are most likely to become multi-purchasers?
• Which of my highly engaged users are most likely to become subscribers?

Later stages:

• Which of my subscribers are most likely to churn?
• Which of my engaged users are likely to make a purchase soon and do not require additional marketing?