In this week newsletter the second statistical ghost: the fallacy of session based metrics. Or in other words: why should you optimize for users and not sessions?
One user can have multiple sessions in the same experiment and the amount of sessions per user can influence the conversion rate for example. At Skyscanner the conversion rate is higher for visitors who have more sessions.
In this article there is a thorough explanation of this issue and how you can avoid being fooled by this statistical ghost.
Also in this newsletter: end experiments early with optimal stopping. In this article, Netflix’s experimentation team shows a way you can stop experiments earlier.
Furthermore in this newsletter:
- Succeeding as a data scientist in small companies
- Top 16 Google Tag Manager extensions worth trying out
- Sitespeed. io
- CXL Live 2019 event
Enjoy your week.