TikTok Ad Targeting Options Explained (Core to Custom)

Most new advertisers assume TikTok targeting works like a search engine: pick the right interests and the right people appear. But the reality of how TikTok ad targeting options actually function is closer to teaching an algorithm than filling out a filter.

TikTok Ad Targeting Options Explained: Core vs Custom

Core (Demographic) Targeting

The most basic layer: age, gender, location and language. Useful for excluding clearly irrelevant groups, but rarely precise enough on its own to drive strong results.

Interest and Behavior Targeting

TikTok categorizes users by content interests, such as fashion, gaming or fitness, and by in-app behaviors, such as video interactions, ad interactions and creator interactions. This layer narrows an audience further, but still describes a broad group of people rather than known customers.

Custom Audiences

Built from a business's own data: website visitors tracked by the Pixel, a customer list uploaded directly, engagers with the brand's TikTok videos or account, or app activity. This is the first targeting layer based on real interaction with the business rather than general interest categories.

Lookalike Audiences

Built from a source Custom Audience, ideally past purchasers, Lookalike Audiences ask TikTok to find new users who share similar characteristics and behavior patterns, at an adjustable similarity level from narrower-but-closer to broader-but-looser.

Automatic Targeting vs Manual Targeting

TikTok increasingly pushes advertisers toward Automatic Targeting, where the algorithm decides who sees an ad based purely on the optimization event, without any interest or demographic filters set manually. This tends to outperform manual targeting once the Pixel has enough purchase data logged, but can perform worse early on, before there's a real signal to learn from.

Minimum Audience Size in Practice

TikTok generally recommends keeping an ad group's eligible audience in the hundreds of thousands at minimum, since a heavily filtered audience of only a few thousand people gets exhausted quickly and forces the algorithm to keep showing the same ad to the same small pool. When a combination of interest, behavior and demographic filters shrinks reach too far, removing one filter is usually more effective than simply raising the budget.

How Broad Should Targeting Really Be?

• In the first week or two of a new campaign, broader targeting generally helps the algorithm find its footing faster than a narrow, over-filtered audience.
• Once a Custom Audience of purchasers exists, a Lookalike built from it usually outperforms interest-based targeting for a Conversions objective.
• Excluding existing customers from acquisition campaigns keeps budget focused on genuinely new buyers rather than people who already convert organically.
• Overlapping audiences across multiple ad groups can cause a business's own campaigns to compete against each other in the auction, worth checking periodically.

Refreshing Audiences as the Business Grows

A Custom Audience or Lookalike built once during setup quietly goes stale as a customer base grows and shifts over time. Rebuilding source audiences from the last few months of purchasers, rather than relying on a static list from launch day, keeps Lookalikes matched to who the business actually sells to now, not who it sold to a year ago.

Common Mistakes With TikTok Ad Targeting

• Stacking so many interest and behavior filters that the eligible audience becomes too small for the algorithm to deliver efficiently.
• Building a Lookalike Audience from a source list that's too small or too generic, such as all website visitors rather than actual buyers.
• Never revisiting Custom Audiences as a business grows, running ads against a customer list that's a year out of date.
• Switching between automatic and manual targeting too frequently, which resets some of the algorithm's learning each time.
• Setting a filtered audience so narrow it falls below TikTok's recommended minimum size, then blaming targeting for what is really a reach problem.

Getting targeting right is only half the job, the other half is watching which audience segment is actually converting profitably today and shifting budget toward it before spend drifts toward a segment that's stopped working. That daily comparison is exactly what AGUDOT automates, reading real campaign metrics from a business's TikTok, Facebook and Google accounts every day and adjusting spend automatically against the budget the owner has set.