AWS Pinpoint Sessions per endpoint lower than 1

The metric "Sessions per endpoint" in AWS Pinpoint represents the average number of sessions per unique endpoint (user device) over a specific period. Seeing values lower than 1 for this metric can be puzzling, but there are a few scenarios and reasons why this might occur:

Possible Reasons for "Sessions per Endpoint" Less Than 1

1. Short Time Frame with Low User Activity: If you are looking at a very short time frame (e.g., an hour or a day) and there is a low volume of user activity, it's possible to have some endpoints with no sessions, leading to an average that is less than 1. For instance, if you have 10 endpoints and only 5 sessions, the average would be 0.5 sessions per endpoint.

2. New Endpoints with No Sessions Yet: If new endpoints (devices) are being registered but haven't yet initiated a session, these endpoints will lower the average sessions per endpoint. For example, if you have 100 registered endpoints but only 50 sessions in a given time frame, the average will be 0.5 sessions per endpoint.

3. Incorrect or Inconsistent Session Reporting: There might be issues in how sessions are being reported from the Android app. Ensure that sessions are being started and ended correctly. For example:

   • A session should start when the app is opened or brought to the foreground.
   • A session should end when the app is closed or sent to the background.

4. Endpoint Duplication or Multiple Endpoints per User: If the same user is represented by multiple endpoints due to reinstallations or other reasons, it can affect the session count. Ensure that endpoints are being managed correctly and de-duplicated where necessary.

5. Time Zone Differences: Ensure that you are considering time zones correctly. Sessions might appear to be missing if there is a mismatch between the time zone of the device and the reporting period in Pinpoint.

Troubleshooting Steps

1. Review Session Tracking Implementation: Ensure that your Android app is correctly implementing session tracking. Verify that sessions are started and ended appropriately using the Pinpoint SDK.

   // Example of starting a session
   pinpointManager.getSessionClient().startSession();
   
   // Example of stopping a session
   pinpointManager.getSessionClient().stopSession();
   

2. Check Endpoint Registration: Ensure that endpoints are being registered correctly and not duplicated. Check if there are any patterns or anomalies in endpoint registration.

3. Analyze Time Frame and Data Volume: Review the time frame you are analyzing. If you are looking at very short periods with low user activity, try expanding the time frame to get a more representative average.

4. Monitor and Debug Events: Use logging and monitoring to ensure that events are being tracked correctly. Check for any errors or issues in event reporting.

5. Pinpoint Console and Reports: Use the Pinpoint console to check the metrics and reports for anomalies. The console provides detailed insights that can help identify where the discrepancies are occurring.

Example: Investigating with Pinpoint Console

1. Log into AWS Pinpoint Console.
2. Navigate to your Pinpoint project.
3. Go to the Analytics section.
4. Check the Session Metrics: Look at the session metrics over different time frames and compare them with the number of endpoints.
5. Review Event Streams: Check the event streams for any anomalies in how sessions are being recorded.

Seeing "Sessions per endpoint" values lower than 1 in AWS Pinpoint for your Android app could be due to low user activity, new endpoints without sessions, incorrect session reporting, endpoint duplication, or time zone differences. By reviewing your session tracking implementation, checking endpoint registration, analyzing the time frame and data volume, monitoring events, and using the Pinpoint console for detailed insights, you can identify and resolve the issue.