Page MenuHomePhabricator

Implement bucketing for Usability Improvements A/B test
Closed, ResolvedPublic

Description

This task is about implementing the changes necessary to ensure people are being bucketed in the Usability Improvements A/B test (T302358) as expected.

Bucketing requirements

  • Bucketing should include both logged-in and logged-out users logged-in users
  • Everyone who is accessing a desktop talk page on any of the wikis listed in the Participating wikis, logged in and logged out, should have a 50% chance of being included/bucketed into the A/B test's control or test group
  • Bucketing should be done on a per-Wikipedia basis. 50% of people within a given wiki should be placed within the control group; 50% should be bucketed in the test group
  • People should remain in the same group for the duration of the test (even if they explicitly turn on or off discussion tool preferences).
  • Testing should be done across various desktop skins (not limited to just Vector 2022)

Instrumentation-Related Requirements

  • Within the A/B test, we are able to distinguish all events logged for the control group and the test group.
  • For all wikis in the test, we are able to join events from DesktopWebUIActionsTracking to EditAttemptStep (confirmed in T330938)
  • We are able to track if/when someone in the test group adjusts any of the discussion tool-related preferences.
  • We can track the desktop skin type being viewed by a user in the AB test (confirmed in T335159)

Participating wikis
Defined in T332946 (still being finalized)

Event Timeline

MNeisler triaged this task as Medium priority.Apr 10 2023, 2:57 PM
MNeisler moved this task from Triage to Current Quarter on the Product-Analytics board.
MNeisler renamed this task from Implement bucketing requirements for Usability Improvements A/B test to Implement bucketing for Usability Improvements A/B test.Apr 26 2023, 6:45 PM
MNeisler updated the task description. (Show Details)

I'm not sure we can apply this to logged out users in any reliable way -- it's going to hit the same caching issue as we encountered with the last A/B test, in that it's a change to the page-content that'd be based on a cookie and so it'd just trigger at a page level and flip around every time the page fell out of the cache based on whatever user happened to view it first.

Change 916902 had a related patch set uploaded (by DLynch; author: DLynch):

[mediawiki/extensions/DiscussionTools@master] Update a/b test code for visual enhancements a/b test

https://gerrit.wikimedia.org/r/916902

Change 916902 merged by jenkins-bot:

[mediawiki/extensions/DiscussionTools@master] Update a/b test code for visual enhancements a/b test

https://gerrit.wikimedia.org/r/916902

Change 917160 had a related patch set uploaded (by DLynch; author: DLynch):

[mediawiki/extensions/DiscussionTools@wmf/1.41.0-wmf.7] Update a/b test code for visual enhancements a/b test

https://gerrit.wikimedia.org/r/917160

Change 917160 merged by jenkins-bot:

[mediawiki/extensions/DiscussionTools@wmf/1.41.0-wmf.7] Update a/b test code for visual enhancements a/b test

https://gerrit.wikimedia.org/r/917160

Mentioned in SAL (#wikimedia-operations) [2023-05-08T20:08:19Z] <taavi@deploy1002> Started scap: Backport for [[gerrit:917160|Update a/b test code for visual enhancements a/b test (T333715)]], [[gerrit:916903|Enable DiscussionTools visual enhancements a/b test (T302358)]]

Mentioned in SAL (#wikimedia-operations) [2023-05-08T20:09:49Z] <taavi@deploy1002> kemayo and taavi: Backport for [[gerrit:917160|Update a/b test code for visual enhancements a/b test (T333715)]], [[gerrit:916903|Enable DiscussionTools visual enhancements a/b test (T302358)]] synced to the testservers: mwdebug2002.codfw.wmnet, mwdebug2001.codfw.wmnet, mwdebug1001.eqiad.wmnet, mwdebug1002.eqiad.wmnet

Mentioned in SAL (#wikimedia-operations) [2023-05-08T20:20:13Z] <taavi@deploy1002> Finished scap: Backport for [[gerrit:917160|Update a/b test code for visual enhancements a/b test (T333715)]], [[gerrit:916903|Enable DiscussionTools visual enhancements a/b test (T302358)]] (duration: 11m 54s)

I've QA'd the AB test events logged to date and confirmed that the numbers appear as expected based on a 50/50 split.

We start logging AB test events on 8 May 2023 for all participating wikis. Here are the number of AB test editing sessions and users that made an edit attempt since the deployment of the test on 8 May 2023 through 16 May 2023.

experiment_groupNumber of distinct editing sessionsNumber of distinct users that made an edit attempt
control152282267
test159112300