> Conditions and groups — the audience builder

TL;DR: A condition is one field + operator + value. A group bundles conditions with AND or OR. Groups can nest. The right grouping changes which records match.

Expand

The atomic unit — a condition

A condition is:

{field} {operator} {value}

Examples:

• icp_code = "Mid-market SaaS"
• persona_score >= 70
• engagement_stage in ["Warm", "Hot", "In-market"]
• last_signal_at >= 14 days ago

Conditions are evaluated independently. Each returns true or false for a given record.

Grouping with AND / OR

A group bundles N conditions with a logical operator:

• AND — all conditions in the group must be true
• OR — at least one condition in the group must be true

The toggle at the top of each group flips between AND and OR. The default is AND.

Single-level example

[AND]
icp_code = "Mid-market SaaS"
persona_score >= 60
engagement_stage in ["Warm", "Hot"]
last_signal_at >= 14 days ago

Matches records that are in the Mid-market SaaS ICP AND have a Persona score >= 60 AND are Warm or Hot AND fired a signal in the last 14 days. All four conditions must hold.

Two-level example with OR

You want records that are Hot OR very recently active:

[AND]
icp_code = "Mid-market SaaS"
persona_score >= 60
[OR]
engagement_stage = "Hot"
last_signal_at >= 3 days ago

The outer AND requires the ICP match AND the Persona score floor AND the inner OR. The inner OR matches if EITHER Hot stage OR a signal in the last 3 days. Records that aren't Hot but fired yesterday still qualify; records that are Hot but haven't signaled in months still qualify.

Three-level — when it's useful

Most audiences don't need 3 levels. When they do, it's usually multi-source intent capture:

[AND]
icp_code = "Mid-market SaaS"
[OR]
[AND]
signal_source = "Replies"
reply_recency_days <= 7
[AND]
signal_source = "Demo Viewed"
demo_view_recency_days <= 14
[AND]
signal_source = "Website Intent"
pricing_page_visit_recency_days <= 3

This audience matches Mid-market SaaS records who EITHER replied in the last week OR viewed a demo in the last 2 weeks OR visited the pricing page in the last 3 days. Different recency windows per source.

Adding, reordering, deleting

In the builder:

• + Add condition — adds a condition to the current group
• + Add group — adds a sub-group to the current group
• Drag the handle on the left of any condition or group to reorder
• The × on the right deletes the condition or group

Empty groups are auto-removed on save.

Common mistakes

• Using AND when you want OR. "Show me records that replied OR visited the pricing page" is OR, not AND. With AND, you'd need them to do both, which is a much smaller set.
• Putting recency at the wrong level. If last_signal_at <= 14 days is at the outer AND, every match must have a signal in the last 14 days. If it's inside an OR with other conditions, only some matches need it.
• Nesting too deeply. Beyond 3 levels, audiences become hard to reason about and hard to debug. If you find yourself going 4+ levels deep, consider splitting into two audiences.

Live count preview

The top of the builder shows the count of records currently matching the conditions. It updates as you add / remove / edit conditions. Two reads:

• Count is 0 — your conditions are mutually exclusive. Usually an AND between two fields that can't both be true.
• Count is the whole workspace — your conditions are too loose. Usually an OR at the outer level that always trivially matches.

A healthy audience is usually 0.1%–5% of your workspace. Wider than 5% and you're probably not segmenting; tighter than 0.1% and you might be over-fitting.

Related

• Building an audience — Copilot first, then manual
• What audiences are
• Audience composition and Reachable %