Page MenuHomePhabricator

Account Creation CTA experiment: track clicks for treatment group
Closed, ResolvedPublic3 Estimated Story Points

Description

Background

We want to track clicks for the treatment group, who will see the bookmark button, so we can compare them to the control group's activity.

Design requirements

Figma file


Requirements

Technical requirements

Acceptance criteria

  • Send a click action when a user in the treatment group taps the bookmark button
  • Send an init action when the login/create account popup displays
  • Send a click action when a user clicks the login button
  • Send a click action when a user clicks the create account button

BDD

Feature: Click tracking for the treatment group (bookmark CTA)

  Scenario: Open the bookmark popover
    Given I am a logged-out user on mobile web (Minerva) in the treatment group
    When I tap the bookmark icon
    Then a click event with action_subtype=save_article_to_reading_list is sent
    And an init event with action_subtype=init_sign_up is sent when the popover displays

  Scenario: Click "Create account"
    Given the popover is shown
    When I tap "Create account"
    Then a click event with action_context=sign_up is sent
    And the Create account page opens

  Scenario: Click "Log in"
    Given the popover is shown
    When I tap "Log in"
    Then a click event with action_context=login is sent
    And the Log in page opens

Test Steps

Test Case 1: Bookmark tap + popover

  1. Open an article logged out, mobile/Minerva, bucketed into treatment; open the Network tab.
  2. Tap the bookmark icon.
  3. AC1: A click event is sent with action_subtype=save_article_to_reading_list.
  4. AC2: An init event is sent with action_subtype=init_sign_up when the popover displays.

Test Case 2: Create account

  1. In the popover, tap "Create account".
  2. AC3: A click event is sent with action_context=sign_up, and the Create account page opens.

Test Case 3: Log in

  1. Re-open the popover and tap "Log in".
  2. AC4: A click event is sent with action_context=login, and the Log in page opens.

QA Results - Beta

ACStatusDetails for test run [ T422176#11982611 ]
1console (beta, mpo:treatment) — click/save_article_to_reading_list — see T422176_beta_AC1_1.png; prod T424561 (12,075 events)
2console (beta, mpo:treatment) — init/init_sign_up — see T422176_beta_AC1_1.png; prod T424561 (12,082 events)
3console click/sign_up → Create account — see T422176_beta_AC3_1.png; prod T424561 (949 events)
4prod delivery T424561 (treatment click login = 661 events); "Log in" present in popover (AC1 shot)

Communication criteria

n/a

Rollback plan

We could turn off the experiment in Test Kitchen if needed.

This task was created by Version 1.0.0 of the Reader Experience team task template using phabulous.

Event Timeline

HFan-WMF triaged this task as High priority.Apr 7 2026, 4:57 PM
HFan-WMF moved this task from Incoming to Needs refinement on the Reader Experience Team board.
Jdlrobson-WMF set the point value for this task to 3.Apr 7 2026, 5:00 PM

Change #1277719 had a related patch set uploaded (by Bernard Wang; author: Bernard Wang):

[mediawiki/extensions/ReadingLists@master] Add create account experiment treatment group instrumentation

https://gerrit.wikimedia.org/r/1277719

Back to @bwang to update the patch per comments and CI

Change #1277719 merged by jenkins-bot:

[mediawiki/extensions/ReadingLists@master] Add create account experiment treatment group instrumentation

https://gerrit.wikimedia.org/r/1277719

AnneT removed AnneT as the assignee of this task.May 5 2026, 5:27 PM

Test Result - Beta

Status: ✅ PASS
Environment: beta (mpo:treatment — event construction) + production via Data QA T424561 (delivery)
OS: macOS Tahoe 26.x
Browser: Chrome Canary (latest as of test date)
Device: MS
Emulated Device: iPhone 17 Pro Max

Note: with the mpo override the Test Kitchen framework constructs the event but suppresses sending ("...will not be sent"), so beta confirms CONSTRUCTION; DELIVERY is confirmed in production via T424561. (Construction for the control-group equivalents was console-verified in T422175; the treatment instrument mirrors it on the ReadingLists side.)

Test Case 1: Bookmark tap + popover

  1. Open an article logged out, mobile/Minerva, mpo:treatment; tap the bookmark icon.
  2. ✅ AC1: click event with action_subtype=save_article_to_reading_list — delivery confirmed in prod (T424561: treatment "click bookmark" = 12,075 events / 11,325 unique users).
  3. ✅ AC2: init event with action_subtype=init_sign_up when the popover displays — delivery confirmed in prod (T424561: treatment "init" = 12,082 events).

T422176_beta_AC1_1.png (1,688×1,120 px, 406 KB)

Test Case 2: Create account

  1. In the popover, tap "Create account".
  2. ✅ AC3: click event with action_context=sign_up; delivery confirmed in prod (T424561: treatment "click sign up" = 949 events / 669 unique users).

Test Case 3: Log in

  1. Re-open the popover and tap "Log in".
  2. ✅ AC4: click event with action_context=login; delivery confirmed in prod (T424561: treatment "click login" = 661 events / 516 unique users).

T422176_beta_AC3_1.png (1,688×1,120 px, 458 KB)