Targeting Rules

How Targeting Works#

When you create a flag, you can define targeting rules as conditions on user attributes. The rules are checked against the context of the user: the user id, the anonymous id and the attributes you provide.

Conditions are combined in a group with AND (all must match) or OR (at least one must match). If no targeting rules are defined, the flag applies to all users. A user who does not match gets the flag's off value with reason NOT_TARGETED.

Setting User Attributes#

Pass user attributes to the SDK so targeting rules can evaluate against them:

Client-Side SDKs

// React Native / Next.js (client)
await flags.identify('user-123', {
  plan: 'pro',
  country: 'US',
  age: 28,
  email: 'user@company.com',
});

// Later: replace the whole set of attributes
await flags.setAttributes({ plan: 'business', country: 'US' });

// Flutter
await flags.identify('user-123', attributes: {'plan': 'pro', 'country': 'US'});

// Android
flags.identify("user-123", mapOf("plan" to "pro", "country" to "US"))

// iOS
flags.identify("user-123", attributes: ["plan": "pro", "country": "US"])

Server-Side SDKs

// Node.js / Next.js (server) — pass context per evaluation
const isEnabled = flags.isEnabled('premium_feature', {
  userId: 'user-123',
  attributes: {
    plan: 'pro',
    country: 'US',
  },
});

What an attribute can be

Attributes are sent to Abrova Flags with each fetch. Send what your rules need, and avoid personal data you do not target on.

Comparison Operators#

Targeting rules support these operators:

The rule builder in Abrova Console offers is, is not, is one of, is not one of, contains, greater than, less than, matches regex, exists, doesn't exist, is in segment and is not in segment.

How values are compared

• Text comparison is case-sensitive: Pro does not equal pro.
• A number and a numeric string are equal when the numbers are equal: 28 equals "28". The number operators also accept numeric strings.
• A boolean equals the strings "true" and "false".
• If the attribute is an array, a condition matches when any item matches. For example tags: ['beta', 'staff'] matches tags equals beta.
• If the attribute is missing, equals, in, contains and the other positive operators do not match, and not_equals, not_in, not_contains and not_in_segment do match.
• regex looks for the pattern anywhere in the value. Use ^ and $ to match the whole value.

Rule Examples#

A rule set is a group with an operator and a list of conditions. You build it in the dashboard; this is the shape the server SDKs receive:

{
  "operator": "AND",
  "rules": [
    { "attribute": "country", "operator": "in", "value": ["US", "CA", "UK"] },
    { "attribute": "plan", "operator": "equals", "value": "premium" }
  ]
}

Here are common patterns:

Country Targeting

{ "attribute": "country", "operator": "in", "value": ["US", "CA", "UK"] }

Plan-Based Access

{ "attribute": "plan", "operator": "equals", "value": "premium" }

Email Domain Targeting

{ "attribute": "email", "operator": "regex", "value": "@company\\.com$" }

Combined Rules (AND Logic)

{
  "operator": "AND",
  "rules": [
    { "attribute": "country", "operator": "in", "value": ["US", "CA"] },
    { "attribute": "plan", "operator": "equals", "value": "premium" },
    { "attribute": "age", "operator": "greater_than_or_equal", "value": 18 }
  ]
}

All three conditions must be true for the flag to be enabled for this user.

Either Condition (OR Logic)

{
  "operator": "OR",
  "rules": [
    { "attribute": "plan", "operator": "equals", "value": "enterprise" },
    { "attribute": "segment", "operator": "in_segment", "value": "beta-testers" }
  ]
}

A segment is a saved group of users defined by its own attribute conditions. The flag is enabled for enterprise users and for everyone in the beta-testers segment.

Targeting and Rollouts Together#

Targeting is checked before the rollout percentage. With rules for plan equals premium and a rollout of 20%, the flag is on for 20% of premium users and off for everyone else. See Percentage Rollouts and the Evaluation Flow.