Detailed Targeting

Detailed targeting is Meta's audience-definition layer: the builder where advertisers combine interests (inferred from engagement), behaviors (purchase and device patterns), and demographics into an audience, with include, narrow ('must also match'), and exclude logic. It is the system most people mean by 'Facebook targeting' — and a decade of privacy change and delivery-AI evolution has quietly demoted it from precision instrument to starting suggestion.

What the categories actually are matters: interests are inferences from likes, engagement, and connected activity — directionally useful, individually noisy (an 'interested in running' member may have liked one video once), which is why interest stacks behave like soft probability filters rather than census facts. The privacy decade pruned the catalog: thousands of sensitive and granular categories (health, ethnicity, beliefs, the DSA-and-DMA-era removals this glossary's law entries explain) are gone, and DETERMINISTIC first-party tools — CUSTOMER-LIST audiences, LOOKALIKES from real buyers — took the precision crown detailed targeting once claimed. Meanwhile the delivery system itself changed the contract: Advantage+ products expand beyond stated targeting by default, and Meta's own guidance increasingly favors broad targeting — letting conversion signal steer the auction — over hand-stacked interests, because the ranking models read behavior richer than any advertiser's five-interest guess. The craft that survives: detailed targeting as exploration (interest splits as creative-research probes), as guardrails where law or brand requires them, in cold-start accounts whose pixels haven't earned signal yet, and as exclusions. The standing failure is the narrowness fetish — stacking interests until the audience is tiny, expensive, and no better than broad, while the CPM premium of micro-audiences quietly taxes the whole plan.

Detailed targeting matters in cold-start accounts, regulated categories needing explicit guardrails, genuinely niche products where interest signal beats nothing, and as the research lens it quietly excels at. It matters less than habit insists everywhere else: accounts with conversion signal and first-party seeds usually buy better with broad-plus-signal than with interest archaeology. The discipline is testing the demotion honestly — broad versus stacked against lift-grade measurement — and spending the saved complexity on creative and first-party data, the inputs the auction actually rewards now.

Detailed targeting was the precision pitch of the Facebook ads decade — thousands of interest and behavior categories built from engagement data — before privacy regulation pruned the catalog and delivery AI (Advantage+ era) demoted hand-built audiences to suggestions the system may override; the feature survives, its mythology mostly retired.

Etymology: source.

Search interest for this term over the last five years:

View interest-over-time on Google Trends →

Areas of marketing where detailed targeting is a core concern: