Audience mistakes are expensive because they propagate. A condition that is too broad can send an irrelevant message to thousands of people; one that is too narrow can make a campaign appear broken. Waiting until an audience is saved, synchronized and used in a campaign turns a simple rule problem into an operational investigation.
Preview the definition, not yesterday's saved version
SegmentHub evaluates the rules currently in the editor. That distinction matters: a marketer can change a condition, add an exclusion or select a rolling date and immediately inspect the proposed result without replacing the live audience.
The preview returns the number of matching profiles, the total profile population and the calculation time. Masked sample profiles provide recognizable context without exposing raw identifiers in the preview. A zero-match warning catches impossible combinations; a universal or extremely broad warning catches definitions that may be technically valid but operationally risky.
See what every rule contributes
A total count alone cannot explain why the result is surprising. Rule contribution applies the definition cumulatively and shows the count after each condition. Consider a retention audience:
- Started with members who viewed a product.
- Restricted them to customers seen within the last two weeks.
- Excluded anyone who already purchased.
If the population collapses at step two, the recency window—not the purchase exclusion—is the first place to investigate. If a rule does not change the count at all, it may be redundant or based on data that is not populated as expected.
Use real customer context without leaking identifiers
Masked sample profiles help answer questions a count cannot: Do the matches look like the intended lifecycle stage? Are expected attributes present? Does a known edge case appear? From the profile view, “why included?” explanations connect the person back to the qualifying audience conditions.
Rolling dates keep audiences from going stale
Fixed dates are useful for one-off analysis. Recurring customer journeys need moving windows such as today, yesterday, one week ago or three months ago. SegmentHub resolves these dynamic values when the rule runs, so “seen within the last two weeks” remains meaningful next month without an editor changing the date.
A practical pre-publish checklist
- Confirm the matched count is plausible relative to the total population.
- Review broad, empty and zero-match warnings.
- Inspect where the cumulative count changes most sharply.
- Check masked samples for the intended customer context.
- Compare against an audience already used in market when overlap matters.
- Only then save, publish and activate.
Audience preview turns segmentation from a configuration exercise into an observable decision. Explore the full workflow on the Audience Intelligence page.