Combine Audience Filters With AND and OR

An audience filter is a list of rules and one choice: every rule must hold, or any one is enough. That choice can multiply or divide your audience. Watch the count while you build the filter, and you'll know who a send or a test will reach before anything goes out.

Illustrative scenario: the brand, the people and every figure on this page are fictional.

Subject

For the grunge fans: the 90s Edit

An audience filter is a list of rules and one choice: every rule must hold, or any one is enough. That choice can multiply or divide your audience. Watch the count as you build the filter, and you'll know who a send or a test will reach before anything goes out.

This series has three parts. This one builds a filter. Part 2 looks at who the filter targets before you sample, and part 3 saves audiences for reuse.

The examples follow Retro Revival, a fictional vintage clothing shop, with made-up numbers. Its contact list has a genre and a decade column, as set up in Import subscribers with the attributes you'll test on.

Rules and one match mode

You'll find the same filter builder on an email's Send page (Audience filter), in a recurring send, in a journey's branch and in an experiment's Audience & sample step. Each rule names a field, an operator and a value, such as "genre equals Rock–Alternative". Above the rules, pick the match mode:

  • Match ALL rules (and): a contact must meet every rule. Each rule you add can only shrink the audience.
  • Match ANY rule (or): one rule is enough. Each rule you add can only grow it.

The same two rules, two audiences

Retro Revival has 46,320 eligible subscribers. 11,240 of them have "Rock–Alternative" as their genre, and 14,870 have "1990s" as their decade. The two rules give very different audiences depending on the match mode:

FilterContacts
genre equals Rock–Alternative11,240
decade equals 1990s14,870
Both rules, Match ALL rules (and)6,420
Both rules, Match ANY rule (or)19,690

AND keeps only the 6,420 who like both: the audience for a grunge collection. OR takes anyone who likes either, which counts the 6,420 once: 11,240 + 14,870 − 6,420 = 19,690. The figure shows how the AND filter narrows the eligible list to the people Retro Revival targeted.

  1. In the list50,874
  2. Eligible46,320

    Left out

    • Unsubscribed or bounced · 2,870
    • No marketing consent · 704
    • Verified invalid address · 612
    • Hasn't confirmed their subscription · 236
    • Opted out of this email's topic · 98
    • Missing or malformed address · 34
  3. Targeted6,420

    Rock–Alternativeand1990s

  4. In the experiment4,000
    • Arm A · 10% off · 2,000
    • Arm B · Free shipping · 2,000
Contacts at each stage of the audience
StageContacts
In the list50,874
Eligible46,320
Targeted6,420
In the experiment4,000
Left out: Unsubscribed or bounced2,870
Left out: No marketing consent704
Left out: Verified invalid address612
Left out: Hasn't confirmed their subscription236
Left out: Opted out of this email's topic98
Left out: Missing or malformed address34
Arm A · 10% off2,000
Arm B · Free shipping2,000
Each stage narrows the one before. Contacts who can't be emailed are left out first. Then the filter keeps contacts with both interests, and the sample is drawn from them and split between the arms.

Watch the count while you build

In an experiment's Audience & sample step, the Audience counts update as you edit the filter: In the list, Eligible, Targeted. Targeted is the number your rules produce, after eligibility. On a Send page, Apply filter to selection ticks the matching contacts in the recipient list below, so you can see and check them before you send.

Add one rule at a time and watch the number. A count that drops to a handful usually means a typo in a value, or a rule that only fits a few people. A count that barely moves means the rule adds little.

Pick the right operator

  • equals compares the whole value, without caring about capitals. Use it for columns holding one value, such as a favourite genre.
  • contains matches part of a value. "contains Rock" also matches "Punk Rock", so prefer equals where you can.
  • does not equal and does not contain also match contacts with no value in that field. Add "is not empty" with Match ALL rules (and) if blanks shouldn't count.
  • is empty and is not empty find contacts with or without a value, which is handy for finding gaps to fill.
  • greater than, at least, less than and at most compare numbers as numbers, and dates written as 2026-10-15 in date order.

Rules don't have to read a column. Depending on where you use it, the builder also offers groups such as Behavior (opened, clicked, purchased, optionally within the last N days), Email verification and Deals. They combine with field rules under the same match mode.

One match mode per filter

A filter has one match mode for all its rules, with no brackets. "Rock–Alternative AND (1990s OR 1980s)" can't be written as a single filter. Two ways around it:

  • Add a column that groups the values. A decade_group column with "80s–90s" for both decades turns the OR part into one rule, so the filter becomes two rules joined by AND.
  • Turn the OR into AND with exclusions. If decade only takes a few values, "decade does not equal 1970s" and "decade does not equal 2000s" together leave the 1980s and 1990s. Remember that does not equal also keeps blanks, so add "decade is not empty".

Getting started

Open a draft experiment's Audience & sample step, pick the contact list and add your first rule. Add the second, switch between Match ALL rules (and) and Match ANY rule (or), and watch Targeted change. When the audience is right, read part 2 to see who's in it before you draw a sample.

A typical build-and-send sequence

  1. 1

    Pick the contact list

    Choose the list in an experiment's Audience & sample step, or link it as the email's audience before using the Send page's filter.

  2. 2

    Add the first rule

    Pick a field, an operator and a value, such as genre equals Rock–Alternative.

  3. 3

    Choose the match mode

    Match ALL rules (and) narrows the audience with each rule. Match ANY rule (or) widens it.

  4. 4

    Watch the count

    The experiment's Audience counts update as you edit. On the Send page, Apply filter to selection ticks the matching contacts.

  5. 5

    Check the operators

    Prefer equals to contains, and remember that does not equal also keeps contacts with no value.

  6. 6

    Work around one match mode

    For a rule like A AND (B OR C), add a column that groups B and C, or use exclusions.

Frequently asked questions

What's the difference between the two match modes?

Match ALL rules (and) keeps contacts who meet every rule, so each rule shrinks the audience. Match ANY rule (or) keeps contacts who meet at least one, so each rule grows it.

Can I nest AND and OR in one filter?

No. A filter has one match mode for all its rules. Add a column that groups the values you'd put in brackets, or turn the OR part into exclusions joined by AND.

Why does does not equal include blank contacts?

A blank value isn't equal to the value you typed, so it matches. Add an is not empty rule with Match ALL rules (and) when blanks shouldn't count.

Are filters case-sensitive?

No. Filters compare text without caring about capitals. Numbers compare as numbers, and dates written as 2026-10-15 compare in date order.

Does the count include unsubscribed contacts?

No. The Targeted count in an experiment comes after eligibility, so unsubscribed, bounced, non-consenting and unconfirmed contacts are already left out.

Build this in the studio

Start on the free plan — every interactive block and the full fallback engine are included on every tier.