Skip to content

Split examples for clustered analysis #464

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 1 commit into
base: main
Choose a base branch
from
Open

Split examples for clustered analysis #464

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
5 changes: 3 additions & 2 deletions docs/experiment-analysis/clustered-analysis.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,7 +4,8 @@ Clustered analysis is available in cases where there is a need to compute metric

Common cases include but are not limited to:
- B2B companies that randomize at the company level, but wants to look at user level metrics
- Social media companies that randomize at the user level (for a consistent user experience), but want to look at session-level metrics
- Marketplaces and Social media companies who want to cluster users more likely to interact with each to share common feature
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Is this suggesting randomizing on those clusters of users? I worry that in many cases there won't be enough clusters to satisfy the CLT. My understanding is that the current approach to clusters on works if you have at least a few hundred cluster samples

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Also, this sentence could be clearer:

.. who want to cluster users more likely to interact with each to share common feature

- Any company that randomize at the user level (for a consistent user experience), but want to look at session-level metrics

Configuring clustered analysis is easy in Eppo. It requires two steps: mapping a new subentity column in the Assignment SQL and choosing clustered analysis as the analysis type when creating an experiment. Metric creation and all other configuration remains the same.

Expand All @@ -26,4 +27,4 @@ Once created and running, there are a few differences on how experiment results

Explore charts behave the same way. Assignment numbers are reported, but metrics and their split bys will use the subentity numbers for the denominator.

![Subentities in Explore charts](/img/experiments/clustered-analysis/explore-with-subentities.png)
![Subentities in Explore charts](/img/experiments/clustered-analysis/explore-with-subentities.png)