Skip to main content

Create a Lookalike Segment from a Seed

Reach more users like your top customers by creating lookalike segments from your seed audience.

J
Written by Juan Sebastian Franco
Updated over 3 months ago

Lookalike segments help you find new users who share traits with your highest‑value customers. By analyzing the demographic and behavioral patterns of your “seed,” Loop identifies a broader audience that closely matches those characteristics.

What Is a Seed?

A “seed” is the set of users whose attributes you want to replicate through enrichment. It typically contains your best or most engaged customers. The lookalike algorithm then finds new individuals who share interests or demographics with that seed.

Example Use Case

Imagine you’re promoting an online fashion store. You have a list of repeat purchasers (your seed) who consistently buy the latest trends. By creating a lookalike segment, you can target more people who mirror those customers’ style preferences—boosting the likelihood of conversions and sales.

Steps to Build a Lookalike Segment

  1. Click Segments, then Create Segment.

  2. Select Lookalike as the segment type.

  3. Choose the Country you want to target.

  4. Pick your Seed Data (e.g., loyal buyers, frequent visitors).

  5. (Optional) Include or exclude certain Device Types (cookies, Android, iOS, or any combination).

  6. Define the Range Factor (1–10%).

    • Lower percentages produce a smaller segment that is more similar to your seed.

    • Higher percentages produce a larger, less similar segment.

  7. Add a Description, then click Create to finalize your lookalike segment.

How to build a Lookalike Segment in Loop

demo LaL.mp4 [video-to-gif output image]

Lookalike segments help you scale your campaigns by targeting people with traits that match your top audiences. By balancing similarity (range factor) and reach (number of potential new users), you can extend your campaign to fresh prospects who are likely to engage and convert.

Did this answer your question?