In order for us to decommission Session Length v1 dataset (disable data collection and the data pipeline), we must provide a replacement.

We can address the limitations of v1 (lack of granularity particularly with shorter sessions, extremely limited segmentation capability) by

• utilizing session identifiers and contextual attributes (e.g. performer_is_logged_in)
• improving granularity of the ticks (e.g. more ticks in the first minute, rather than evenly spaced 1 minute ticks)

Acceptance criteria

• Requirements gathered from stakeholders
  • Dimensions that would be useful for analysis
  • Granularity of ticks
• Instrument to replace sessionTick.js (v1)
• Data pipeline to replace session length (v1) daily DAG
• Dataset explorable in Superset, replacing session length (v1) dashboard