name: Statsig Vocabulary description: >- Domain vocabulary for Statsig's feature management and experimentation platform covering feature gates, A/B testing, dynamic configuration, and analytics concepts. version: '1.0.0' created: '2026-05-02' modified: '2026-05-02' terms: - term: Feature Gate definition: >- A boolean flag that controls whether a feature is enabled for a given user. Evaluated using targeting rules based on user properties such as user ID, email, country, or custom attributes. synonyms: - Feature Flag - Feature Toggle tags: - Feature Management - Core - term: Dynamic Config definition: >- A key-value configuration object that returns different values for different users based on targeting rules. Enables server-driven configuration without app deployments. synonyms: - Remote Config - Dynamic Configuration tags: - Configuration - Core - term: Experiment definition: >- An A/B test or multivariate test that assigns users to different groups and measures the impact of parameter variations on key metrics. Experiments have a hypothesis, groups with parameter values, and an allocation percentage. synonyms: - A/B Test - Split Test - Multivariate Test tags: - Experimentation - Core - term: Layer definition: >- A namespace for parameters shared across multiple mutually exclusive experiments. Users are assigned to only one experiment per layer, preventing interaction effects between tests. tags: - Experimentation - Organization - term: Holdout definition: >- A group of users excluded from receiving specific features to measure the cumulative long-term impact of feature launches. Holdouts help identify interaction effects and sustained improvements. tags: - Experimentation - Measurement - term: Autotune definition: >- A multi-armed bandit optimization system that automatically shifts traffic to the best-performing variant based on a target success metric, without manual intervention. tags: - Optimization - Machine Learning - term: Exposure Event definition: >- An automatically logged event that records when a user evaluates a feature gate, dynamic config, or experiment. Used to accurately count experiment participants and calculate metric impacts. tags: - Events - Analytics - term: Rule definition: >- A targeting condition within a feature gate or dynamic config that defines which users match and what value they receive. Rules include conditions (user ID, email, country, custom fields), a pass percentage, and a return value. tags: - Feature Management - Targeting - term: Condition definition: >- A single predicate within a targeting rule that evaluates a user attribute against a target value using an operator. Condition types include user_id, email, country, custom_field, and segment membership. tags: - Targeting - Rules - term: Server Secret Key definition: >- A confidential API key used by server-side SDKs and server-side HTTP API calls. Must only be used in trusted server environments and never exposed in client-side code. tags: - Authentication - Security - term: Client SDK Key definition: >- A publishable API key used by client-side SDKs in mobile apps and front-end web applications. Safe to embed in public code as it only provides read access to evaluated configurations. tags: - Authentication - SDKs - term: Console API Key definition: >- A management API key for the Statsig Console API, granting full CRUD access to project configuration. Should be kept confidential. tags: - Authentication - Administration - term: Download Config Specs definition: >- The endpoint used by server SDKs to download the complete project configuration for local evaluation. Enables sub-millisecond feature gate checks by eliminating per-request network calls. tags: - Server SDK - Performance - term: Local Evaluation definition: >- The process of evaluating feature gates and experiments entirely within the SDK using downloaded configuration specs, without making network requests for each evaluation. tags: - Server SDK - Performance - term: Segment definition: >- A reusable user group definition used in targeting rules. Segments can be ID-list based (specific user IDs) or rule-based (matching conditions). tags: - Targeting - Organization - term: Target App definition: >- A definition that scopes feature configurations to a specific application (e.g., iOS app, Android app, web app), enabling multi-app management from a single Statsig project. tags: - Organization - Multi-App - term: Audit Log definition: >- A timestamped record of all changes made to project configuration through the Console or API, tracking who changed what and when. tags: - Governance - Compliance - term: Override definition: >- A manual assignment forcing specific users or IDs to always receive a specific gate value or config, bypassing normal targeting rules. Used for testing, debugging, and internal users. tags: - Testing - Feature Management - term: Custom ID definition: >- An alternative identifier for users beyond the standard userID, allowing targeting by company ID, device ID, or other custom identifier types. tags: - Targeting - User Identity - term: Private Attributes definition: >- User attributes used for evaluation but stripped from event logs before being sent to Statsig servers, protecting sensitive user data. tags: - Privacy - User Identity categories: - name: Core Concepts terms: - Feature Gate - Dynamic Config - Experiment - Layer - name: Advanced Features terms: - Holdout - Autotune - Segment - Target App - name: Targeting terms: - Rule - Condition - Override - Custom ID - name: Authentication terms: - Server Secret Key - Client SDK Key - Console API Key - name: Analytics terms: - Exposure Event - name: Performance terms: - Download Config Specs - Local Evaluation - name: Governance terms: - Audit Log - Private Attributes