LaunchDarkly

Overview

LaunchDarkly is a feature flagging and experimentation platform used by engineering teams to manage feature rollouts, A/B tests, and operational kill switches. Its client-side SDK evaluates feature flags based on targeting rules without requiring code deployments.

What This Script Does

LaunchDarkly's JavaScript SDK (ldclient-js) connects to LaunchDarkly's streaming infrastructure and performs the following:

• Flag evaluation: On initialization, the SDK sends a user context object (containing attributes like user key, email, plan, or custom properties) to LaunchDarkly's servers. It receives back the evaluated flag values for that user.
• Streaming updates: Maintains a persistent connection (via Server-Sent Events or polling) to receive real-time flag changes without page reloads.
• Event tracking: Sends analytics events back to LaunchDarkly, including flag evaluation counts, custom conversion events (for experimentation), and optional click/pageview events for A/B test measurement.
• Local storage: Caches flag values in localStorage to reduce latency on subsequent page loads and provide fallback values when offline.
• User identification: The SDK requires a user key to evaluate targeting rules. This can be an anonymous auto-generated key (stored in a cookie or localStorage) or an identified user key provided by your application.

The SDK communicates with app.launchdarkly.com, events.launchdarkly.com, and clientstream.launchdarkly.com.

Consent & Compliance

Consent Category: Functional

LaunchDarkly's primary purpose is feature management, which is an operational concern rather than marketing or analytics. However, the consent classification depends on how it is configured:

• Functional use (feature flags only): If the SDK is used purely for feature rollouts and operational toggles with anonymous user keys, it falls under functional/strictly necessary. The localStorage cache and anonymous key cookie support core application behavior.
• Analytics/experimentation use: If LaunchDarkly is used for A/B testing with conversion tracking, click events, and pageview events, the experimentation data collection moves into analytics territory and may require consent.
• GDPR consideration: If identified user attributes (email, name, plan) are sent as part of the user context, this constitutes personal data processing. LaunchDarkly acts as a data processor in this case.

Should You Block This Without Consent?

Conditional. If LaunchDarkly is used strictly for feature flags with anonymous user keys and no experimentation events, it can be classified as functional and does not require consent. If you are running A/B experiments with conversion tracking or passing identified user data, obtain analytics consent before initializing the SDK. Review your targeting rules and event configuration to determine which category applies to your implementation.