Targeting Conditions

Targeting Conditions

A targeting condition is a filter that determines which users a rule applies to. Each rule on a Feature can have zero or more conditions. All conditions on a rule must be true for the rule to match the current user. If any condition is false, the user does not match that rule and GrowthBook moves on to the next one.

Adding a condition in the GrowthBook UI

1. On a Feature’s detail page, click Add Rule (or click the pencil icon on an existing rule).
2. In the rule editor, click Add Condition.
3. Select the Attribute from the dropdown (e.g. country).
4. Select the Operator (equals, not equals, contains, regex, etc.).
5. Enter the Value (e.g. TH).
6. Click Save Rule.

You can add as many conditions as you need. Each additional condition narrows the audience further.

Condition operators

GrowthBook supports a wide range of operators to match different attribute types:

• equals / does not equal — exact match on string, number, or boolean
• contains / does not contain — substring match for strings
• matches regex — test a string against a regular expression
• is greater than / is less than — numeric comparisons
• is in list / is not in list — check whether a value appears in a comma-separated set

Advanced JSON mode

For power users, GrowthBook supports a MongoDB-style condition JSON — the same format used internally by the evaluation engine. Click Advanced in the rule editor to enter raw JSON instead of using the visual builder.

Match users where country is TH and plan is pro:

{
  "country": "TH",
  "plan": "pro"
}

Match users where email ends with @company.com:

{
  "email": { "$regex": "@company\\.com$" }
}

The advanced editor is useful for complex expressions (nested $or, $and, $not) that the visual builder does not expose.

import { GrowthBook } from '@growthbook/growthbook';

const gb = new GrowthBook({
  apiHost: 'https://cdn.growthbook.io',
  clientKey: 'sdk-YOUR_DEV_KEY',
});

await gb.init({ timeout: 2000 });

gb.setAttributes({
  id: 'user-456',
  country: 'TH',
  email: '[email protected]',
});

// Returns the feature value if the user matches all conditions on a rule,
// or the default value ('false') if no rule matches.
const isBetaEnabled = gb.getFeatureValue('beta-feature', false);
console.log('beta-feature:', isBetaEnabled);

Copy conditions example ⧉

🧪 Conditions let you ship to a precise audience

Combine multiple conditions on a single rule to narrow the audience precisely — e.g. employees in Thailand on the pro plan. Users who do not match all conditions skip the rule and fall through to the next one.

Trade-offs

Option	Benefit	Cost
Complex targeting conditions	Much more precise	Harder to debug when results don’t match expectations
Simple conditions	Easier to debug	Coarser, covers fewer use cases

Common mistakes

• Writing self-contradicting conditions (e.g. country = TH AND country = US), matching no users at all
• Nesting so many condition layers that no one on the team can read them
• Not testing the condition against real users before shipping it

💡 Real-world example

Growth teams commonly combine several layers of targeting conditions (country + device + subscription tier) to release a feature to the segment most likely to respond well first.

A rule has two conditions: country equals TH and plan equals pro. A user has country: TH and plan: free. Does the rule match?

Yes — at least one condition matches No — all conditions on a rule must be true for the rule to match Yes — the first condition takes precedence It depends on the rule order in the feature

The GrowthBook Advanced JSON condition editor uses which format?

GraphQL filter syntax SQL WHERE clause syntax MongoDB-style condition JSON JSONPath expression syntax

A user does not match any condition on a rule. What happens next?

GrowthBook throws an evaluation error The feature is disabled for that user The user skips that rule and GrowthBook evaluates the next rule in order GrowthBook returns null

Submit

Mark as complete