Page MenuHomePhabricator

Experiment: Account setup: Instrumentation + A/A test (DE1.3.1)
Open, HighPublic3 Estimated Story Points

Description

INSTRUMENTATION SPECIFICATION

User story & summary

As a Growth team member, I want instrumentation in place and validated via an A/A test so that we can reliably measure the DE1.3.1 experiment before the A/B test begins.

Background

Parent task: T430058: [EPIC] Account Setup (DE1.3.1 - early onboarding)
Experiment details: T432126: Experiment: DE1.3.1 Early onboarding & account set-up

Hypothesis: If we reduce cognitive load and provide more personalized post-account creation experiences for new account holders, then a greater proportion of users will constructively activate by making their first edit.

Details

Events/interactions to instrument:

  • Welcome Survey / Account Setup started (at least one answer completed)
  • Welcome Survey / Account Setup completed (all questions completed)
  • Proportion of accounts with email addresses
  • Homepage impression

Which schema(s)/stream(s) to use: TBD

This A/A test should validate:

  • Instrumentation is working as expected.
  • There are no SRM between groups. (Although this might not be verifiable until an A/B test is started)

Testable hypothesis

If we run an A/A test with identical experiences in both groups, it will show no significant differences between groups because any observed difference would indicate a problem with instrumentation or group assignment (SRM).

Experiment details

Overview
  • Experiment Name: Account setup A/A test (DE1.3.1)
  • Tracking Key: TBD
  • Risk Level: Low risk
Traffic
  • Assign Variation by Attribute: MW User ID
  • Traffic Percent: 100%
  • Variations:
    • A1 (control): default post-account-creation experience
    • A2 (treatment): identical to control (A/A test — no experience differences between groups)
  • Targeting
    • MediaWiki Database Code in:
      • arwiki, enwiki, frwiki, eswiki
  • Expected start date: August 24, 2026

Metrics

Event Timeline

KStoller-WMF moved this task from Needs Refinement to Needs Estimation on the Growth-Team board.
KStoller-WMF set the point value for this task to 3.
KStoller-WMF moved this task from Needs Estimation to Up next Sprint on the Growth-Team board.

Change #1321955 had a related patch set uploaded (by Michael Große; author: Michael Große):

[mediawiki/extensions/GrowthExperiments@master] instrument: collect data for onboarding A/A test

https://gerrit.wikimedia.org/r/1321955

Michael updated the task description. (Show Details)

Change #1324770 had a related patch set uploaded (by Michael Große; author: Michael Große):

[mediawiki/extensions/WikimediaEvents@master] instrument(EarlyOnboarding): enable reader retention metrics

https://gerrit.wikimedia.org/r/1324770

Change #1324770 merged by jenkins-bot:

[mediawiki/extensions/WikimediaEvents@master] instrument(EarlyOnboarding): enable reader retention metrics

https://gerrit.wikimedia.org/r/1324770

Change #1321955 merged by jenkins-bot:

[mediawiki/extensions/GrowthExperiments@master] instrument: collect data for onboarding A/A test

https://gerrit.wikimedia.org/r/1321955

Change #1329533 had a related patch set uploaded (by Michael Große; author: Michael Große):

[mediawiki/extensions/WikimediaEvents@master] instrument: record edit_saved events for early onboarding experiment

https://gerrit.wikimedia.org/r/1329533

Change #1329533 merged by jenkins-bot:

[mediawiki/extensions/WikimediaEvents@master] instrument: record edit_saved events for early onboarding experiment

https://gerrit.wikimedia.org/r/1329533

Change #1338006 had a related patch set uploaded (by Michael Große; author: Michael Große):

[mediawiki/extensions/GrowthExperiments@master] instrument(AccountSetup): don't double-count Homepage visits

https://gerrit.wikimedia.org/r/1338006

Change #1338006 merged by jenkins-bot:

[mediawiki/extensions/GrowthExperiments@master] instrument(AccountSetup): don't double-count Homepage visits

https://gerrit.wikimedia.org/r/1338006

To be resolved when we get close to start the A/B test. At the very latest some time next week as the team is now merging changes for the A/B test which means that the A/A test will stop receiving data next week Thursday.