Skip to content

Segment Overlap

Segment overlap exploration shows how many users belong to two or three segments simultaneously. It renders as a Venn diagram with intersection counts and a table of metrics for each segment combination. The primary use is understanding relationships between different user behaviors and characteristics.

Audience research: What percentage of your mobile users are also purchasers? What percentage of organic search users also came from email at some point?

Campaign analysis: How many users were exposed to both email and paid search before converting? Are these users more likely to purchase than users who saw only one channel?

Product analysis: What proportion of your trial users are also active blog readers? Do trial users who engage with the product tour have higher conversion to paid?

Attribution investigation: How much do your acquisition channels overlap? If a large percentage of paid search users also have an organic search visit, last-click attribution may be significantly understating organic’s contribution.

  1. Go to Explore → Create new exploration → Segment overlap.

  2. In the Variables panel, add the segments you want to compare. You can compare 2 or 3 segments.

  3. Drag the segments from Variables to Segments in the Tab Settings panel.

  4. Select the metrics to display for each intersection (users, conversions, revenue, etc.).

  5. The Venn diagram and intersection table populate automatically.

The Venn diagram shows:

  • Each circle = one segment
  • The number inside each circle (outside overlaps) = users in that segment only
  • Overlap areas = users in multiple segments simultaneously

Below the diagram, a table shows each segment combination and the selected metrics.

For two segments A and B:

  • A only: users in A but not B
  • B only: users in B but not A
  • A ∩ B: users in both A and B

For three segments, you get seven combinations: each segment alone, each pair, and all three together.

Segments:

  • Segment A: Device category = Mobile
  • Segment B: Event = purchase

Question: What percentage of purchasers are mobile? What percentage of mobile users purchase?

If Segment A has 60,000 mobile users and Segment B has 8,000 purchasers, and the intersection is 3,200 users — 40% of purchasers are mobile, but only 5.3% of mobile users purchase. If desktop purchasers represent 4,800 users out of 40,000 desktop users, the purchase rate is 12% for desktop. This tells you mobile has significant room to improve or may need a different checkout experience.

Segments:

  • Segment A: Sessions from Paid Search
  • Segment B: Sessions from Email
  • Segment C: Sessions from Organic Social

Question: How many users have been touched by multiple channels?

A large intersection of A ∩ B ∩ C suggests your retargeting and email sequences are reaching users who initially arrived via paid search — this is the multi-touch journey you want to understand for attribution purposes.

Segments:

  • Segment A: Users who viewed 3+ blog posts
  • Segment B: Users who completed a trial signup

Question: Is there a correlation between content engagement and trial signup?

If the A ∩ B intersection is proportionally large relative to Segment B total (blog readers make up 40% of trial signups but only 15% of total users), content engagement is a positive signal for conversion. This supports investment in content marketing.

Choose metrics in the Tab Settings that are most relevant to your analysis question:

  • Users: the baseline — how many users in each segment
  • Conversions: did overlap affect conversion behavior?
  • Purchase revenue: are multi-segment users higher-value?
  • Engagement time: do overlapping users engage more deeply?
  • Sessions per user: do overlapping users visit more frequently?

You cannot use calculated metrics in segment overlap by default, but you can compute ratios manually from the table output.

Maximum 3 segments: Segment overlap supports at most 3 simultaneous segments. For more complex analyses, consider comparing multiple 2-way overlap explorations.

User scope only: Segment overlap operates at the user level. You cannot do session-level overlap analysis.

Sampling: Large properties with many users will experience sampling. The Venn diagram counts are estimates. Check the sampling indicator.

No dimensional breakdowns: Unlike free-form explorations, you cannot add secondary dimensions to segment overlap. The output is the segment intersections and your selected metrics only.

Segment overlap vs. comparisons in standard reports

Section titled “Segment overlap vs. comparisons in standard reports”

Standard reports let you add comparisons (up to 4) that filter the same report by different conditions. This is similar to segment overlap but:

  • Comparisons in standard reports show the same metrics side-by-side for each filter
  • Comparisons do not show intersections (how many users are in multiple groups)
  • Segment overlap specifically calculates and shows the intersection populations

Use standard report comparisons when you want to compare segment performance on standard metrics. Use segment overlap when the intersection size itself is the question.

Comparing segments with very different scales

Section titled “Comparing segments with very different scales”

If Segment A has 500,000 users and Segment B has 200 users, the Venn diagram is nearly useless visually — the circles are disproportionate and the intersection numbers are all tiny. Segment overlap works best when segments are on a comparable scale.

Segment overlap shows correlation, not causation. The fact that blog readers are more likely to convert does not prove that reading blogs causes conversion — it may be that users who are more interested in your product both read your blog and convert at higher rates for independent reasons. Use experimentation to establish causation.

Segment overlap is calculated over the exploration’s date range. Users who were purchasers last year but not in the current date range will not appear in the “purchaser” segment for a current-period exploration. Define your date range carefully to ensure the segments reflect the population you intend to compare.