Experience Analytics Mobile Sessionization FAQ

What is sessionization?

Sessionization is the process during which the data pipeline groups events together according to the session ID set up by the mobile SDK. The sessions are then sent to our database, ClickHouse, where they are available for the platform permanent requests.

Good news! This process was updated and improved, resulting in better data quality.

What changed concretely?

As more sessions and events are taken into account, more data has to be processed. Thus, you may observe a significant increase in the number of sessions, as well as a variation in other metrics:

• Data insertion was reduced from 2 hours to a couple of seconds on Contentsquare
• Session data will be updated up to 7 days after it is inserted (late data will be taken into account)
• Session duration limitation has increased
• Dynamic variables sent before the first screen view of the session will be processed
• Transaction collection reliability was improved

More data gives better insights!

How will this impact my analysis on Contentsquare?

As the number of sessions and events considered increases, the data for some metrics is likely to change. For example, the number of sessions may increase significantly. Other metrics impacted include time spent related metrics, number of screen views, sessions with transactions and in-screen events (gestures, increase in dynamic variables when more screen views are retrieved).

Why didn't we have these sessions before?

Before Flink (our new data processing engine), if additional data was received after the session was inserted, that data was discarded. Now, if we receive data after a session has been inserted, the session will be updated. Thus, the value of a KPI displayed for a specific Analysis Context can vary up to 7 days after the selected period. This is due to the late entry of offline data that is then used to update the results.

We can receive late data in the following situations:

• the entire session happens offline

• the device loses connectivity until the end of the session

• the app crashes or is force killed (and isn't restarted within 30 minutes)

• the device is shut down by the user or turns off (no battery) and the app isn't restarted within 30 minutes

To sum up, if you see data discrepancies when you run an analysis, don't worry, it's just the result of better data collection after our migration to Flink.