Audience overlap is the intersection of two or more audiences a brand is paying to reach concurrently — distinct from reach (users actually exposed) and audience size (the targetable universe). It comes in two flavors measurable to very different degrees: within a single ad platform, and across platforms.
Within a platform, overlap is in principle directly measurable from the platform’s own user graph, and most major ad platforms have at some point surfaced an overlap-reporting diagnostic — though the specific feature labels and entry points shift between releases. Across platforms there is no shared user identifier, so cross-platform overlap has to be inferred: through first-party identity matching in clean rooms, deterministic match via uploaded customer lists, panel data, or media-mix modeling. None are exact. This asymmetry is part of why blended measurement like MER and incrementality testing exist.
Two failure modes dominate account audits. Heavy overlap between prospecting and retargeting means the prospecting budget is partially paying to re-reach already-engaged users, inflating prospecting’s ROAS while deflating retargeting’s. Two lookalike audiences with high overlap usually mean the seeds were too similar — the model has converged on a narrow pool instead of expanding reach. Overlap also explains frequency creep: a user sitting in three separately-capped audiences can receive 3× the intended exposure, because per-audience caps don’t compose.
Overlap is the structural reason channel-level attribution and incrementality bias point the same way — the same user is counted as reachable, served, and sometimes converted across multiple audiences whose budgets each claim credit. It is a planning and diagnostic signal, not a delivered-impression metric: it shapes audience construction and budget allocation, not the daily performance dashboard.