Motivation
We would like to understand how useful the "show related temporary accounts" mode is on Special:Contributions for temporary accounts.
Background
- When a user with IP reveal rights visits Special:Contributions for a temporary account that has related temporary accounts (who have shared the same IP addresses), a warning box is shown, with a link to show also the contributions from the related temporary accounts.
- Clicking on the link reloads the page. The warning box is also shown on the reloaded page, with a link to hide the contributions from the extra temporary accounts.
- This is not shown to a user without IP reveal rights
- This is not shown if there are no other temporary accounts have used the same IP address
Instrumentation
Add some instrumentation to capture the following:
| Event | Information |
|---|---|
| The Special:Contributions page is visited for a temporary account, by a user who has IP reveal rights | Used as a denominator to calculate what proportion of the time a page is visited that has related temporary accounts |
| The warning message is shown, when the page is not already in "show more" mode | The numerator to calculate what proportion of the time a page is visited that has related temporary accounts. Also the denominator to calculate what proportion of the time the "show more" link is clicked |
| The "show more" link is clicked | Numerator to calculate the proportion of times the user clicks to show related temporary accounts, when the warning box is shown. |
How we will use the information
The instrumentation will help us measure the following:
- How often is the feature relevant.
- How often it is noticed.
- How often it is used when available.
We would like to understand how often do patrollers land on a TA that has related TAs to show. Are many temporary accounts part of clusters? Do patrollers frequently encounter situations where this broader context is available? This will help us understand whether this is a common part of the workflow or it’s an edge case.
When the feature is shown, we would like to understand whether patrollers actually choose to use it . This will tell us whether the feature becomes part of the natural patrolling workflow or is ignored.
We want to asses the numbers on a weekly basis until the end of the quarter