Skip to Content
InsightsAudience Overlap

Audience Overlap

The Audience Overlap dashboard compares two audiences to see how much membership they share. Use it to plan deduplication, design exclusion targeting, and understand how distinct two audiences really are.

What Audience Overlap Shows

Pick two audiences and Zeotap evaluates both against your warehouse and computes their intersection. The result answers:

  • How many members are in both — the shared population
  • How many are exclusive to each — the incremental reach each audience brings
  • How similar they are overall — as a Jaccard index, which normalises for size

Overlap Metrics

MetricFieldFormulaDescription
Overlapoverlap_size|A ∩ B|Members in both audiences
Only Aonly_a|A| − |A ∩ B|Members in A alone
Only Bonly_b|B| − |A ∩ B|Members in B alone
Jaccard indexjaccard_index|A ∩ B| / |A ∪ B|Similarity from 0 (disjoint) to 1 (identical), independent of how large either audience is

The Jaccard index is the one to read when the two audiences are very different sizes. A 5,000-member audience fully contained in a 500,000-member one overlaps 100% of itself and 1% of the other; the Jaccard index says 0.01, which is the honest answer to “are these the same audience?”

The three counts add up to the union, so the total unique population across both is overlap_size + only_a + only_b.

How to Use Audience Overlap

Deduplication Planning

If you’re syncing multiple audiences to the same destination, overlap analysis helps you understand the actual incremental reach of each audience:

  1. Pick two audiences that sync to the same destination
  2. Review the overlap — a high one means you’re sending largely the same people twice
  3. Consider consolidating overlapping audiences into a single audience to reduce sync volume and destination API costs

Example: “High-Value Customers” and “Repeat Buyers” may have 80% overlap. Instead of syncing both to your CRM, create a single “High-Value Repeat Buyers” audience or sync only the one that better matches your use case.

Exclusion Targeting

Use overlap analysis to build exclusion audiences for advertising:

  1. Compare your “Current Customers” audience with your “Prospecting” audience
  2. If there’s significant overlap, members who are already customers are seeing prospecting ads
  3. Create a sync for “Current Customers” as a suppression list in your ad platform
  4. The ad platform will exclude these members from prospecting campaigns, reducing wasted spend

Understanding Audience Composition

When designing a new audience, compare it against existing audiences to understand who you’re targeting:

  1. Create your new audience (e.g., “Holiday Campaign 2024”)
  2. Compare it, one pair at a time, against your key reference audiences:
    • “All Active Customers” — What fraction of your active base is being targeted?
    • “Previous Holiday Campaign” — How different is this year’s targeting from last year?
    • “High-Value Customers” — Are you including your best customers?
  3. Adjust the new audience’s filter criteria based on the overlap insights

Funnel Analysis

Compare audiences that represent different stages of the customer lifecycle to understand conversion flow:

Signed Up (10,000)
  → First Purchase (3,500) — 35% conversion from signup
    → Repeat Buyer (1,200) — 34% conversion from first purchase
      → VIP (300) — 25% conversion from repeat buyer

Compare each adjacent pair. A later stage should be a proper subset of the one before it — only_b should be zero when B is the later audience — so a non-zero count there means the filters are not nested the way you think.

Running an Overlap Analysis

  1. Navigate to Insights > Audience Overlap
  2. Choose two audiences to compare
  3. Zeotap evaluates both against your warehouse and returns the counts

Two things govern how often that query actually runs:

  • Results are cached for an hour, keyed on the pair. Asking the same question twice inside that window returns the first answer rather than re-querying.
  • A workspace may compute at most ten overlaps per hour. Beyond that the request is refused until the window rolls forward. Overlap is the most expensive query the Insights module issues — two full audience evaluations and a join — so it is deliberately rationed rather than left to be run in a loop.

API Endpoint

GET /api/v1/workspaces/{id}/insights/audience-overlap?audience_ids=aud_a,aud_b

Query parameters:

ParameterTypeDescription
audience_idsstringTwo audience IDs, comma-separated. Required, and exactly two — a different count is refused with 400.

Response:

{
  "overlap_size": 12405,
  "only_a": 3210,
  "only_b": 88771,
  "jaccard_index": 0.1189
}

A request that exceeds the ten-per-hour limit is refused rather than queued.

Last updated on
Activation CoverageOverview