Lookalike Audience

An audience of new users Snap identifies as similar to a seed list of your existing customers.

You give Snap a seed audience, usually your customers or high-value pixel events, and its models find users across Snapchat who resemble that group. You pick how tightly to match on a scale from Similarity, which is the closest and smallest, through Balance, to Reach, which is the broadest and loosest. Tighter means more relevant and fewer people; broader means more scale and more slack.

You seed a lookalike with your last ninety days of purchasers and choose Similarity to start. It is a small, precise pool, so once it performs you build a second lookalike on Balance from the same seed to add scale, keeping the tight one running rather than replacing it.

A lookalike inherits the quality of its seed. Feed it a list of newsletter signups when you care about purchases and it will faithfully find you more signups, not more buyers. Seed on the outcome you actually want, and give it enough records to be worth modeling, not a few hundred rows.

Garbage seed, garbage twins. Seed on the outcome.