Event Streaming Overview
Event streaming lets you send the data in Amplitude across your stack. With event streaming, you can use the rich behavioral data in Amplitude to enrich customer profiles and stream data to marketing, sales, and infrastructure tools.
Event streaming includes powerful, no-code, configuration-based tools that give fine-grained control over the data you send. Filter by user, group, and event properties to control what you forward to your other tools. You can also see important metrics like event volume, end-to-end latency, and detailed Delivery Issue information to understand the performance and health of your integration.
- No transformations: Amplitude event streaming supports raw (untransformed) events, event properties, and user properties. Transformed events and properties (such as merged properties) aren't supported.
- User Properties Format: All forwarded user properties are currently sent as strings.
- Reserved Keywords: Specific keywords, including "_all" and "_identify," can't be used as event names when streaming events from Amplitude.
- Billing Efficiency: Amplitude tracks event volume based on distinct events streamed out. If you send the same event to multiple event streaming destinations, it's counted once for billing.
- Latency Target: For latency, Amplitude targets an end-to-end p95 latency of 60s. This means 95% of Events streamed must deliver in 60s or less. Amplitude has internal processes, monitors, and alerts in place to monitor and support this target.
- Starting from Event Streaming Setup: Amplitude's streaming integrations focus on data from the setup point forward. Historical data isn't included in this process, which ensures that Amplitude transmits only events captured post-configuration.
- Reliable Retry Mechanism: Amplitude addresses intermittent errors through in-memory retries with exponential backoff for initial sends. A robust retry pipeline attempts up to 10 times within a 4-hour window to handle retry errors. Amplitude applies this mechanism to all Event Streaming destinations.
- Streamlined Monitoring and Management: The Event Streaming Debugger UI in Amplitude Data allows users to monitor pending retries and their progress. Expired payloads that exhaust retry attempts are clearly marked. The UI provides insights into error categories and also offers samples of failed payloads for analysis.
What's the difference between Cohort syncing and Event Streaming?¶
- Cohort Syncing provides a valuable mechanism for syncing cohort membership data from Amplitude to third-party tools such as SFMC or Braze. This empowers you to delve into behavioral targeting and comprehensively analyze the effects of your targeting strategies. However, it's important to note that Cohort Syncing primarily focuses on data synchronization and behavioral analysis, without involving the creation or maintenance of users within the third-party tool.
- Event Streaming, goes beyond syncing cohorts. It facilitates user creation and management through the Amplitude UI. By setting up Event Streaming, you benefit from a single Amplitude instrumentation setup, allowing data and events to flow seamlessly to various supported platforms. This eliminates the need for ongoing engineering adjustments. Additionally, Event Streaming provides precise control, allowing you to cherry-pick specific events, users, or properties for retention within the destination platform. This tailored approach ensures only relevant data reaches its intended target. Finally, Event Streaming unlocks the real-time potential of conversion events. For instance, in a scenario like SFMC, these events can serve as triggers within workflows or journey orchestrations. This strategic integration empowers the optimization of look-alike targeting, thereby augmenting the effectiveness of your initiatives.
What are some examples of how customers are using Event streaming?¶
- Marco Polo: Used Event Streaming to power a real-time 'Welcome' email campaign by streaming sign-up events from Amplitude to Braze.
- RefaceAI: Utilized Amplitude's Kinesis Data Stream integration for real-time event piping from mobile clients to back-end services. This avoided the complexity of building and maintaining a custom pipeline.
- Invoice Simple: Leveraged Event Streaming for a robust engagement campaign, customizing messaging based on a series of events to enhance engagement effectiveness.
Event streaming destinations¶
- Amazon Kinesis Data Stream
- Amazon Kinesis Firehose
- Google Analytics 4 (iOS/Android)
- Google Analytics 4 (Web)
- Google Ads
- Google Pub/Sub
- Kochava (Post Install)
- Kochava (Install)
- Meta Pixel
- :movable-ink: Movable Ink
- :sfmc: Salesforce Marketing Cloud