Background

As part of the work on donor ID and recognition, in WE 3.5.5 hypothesis we want to run an experiment for logged-out users on mobile-only that will present donors with an on-wiki token of appreciation. We have no post-donation retention signal for logged-out donors on mobile web. Once a donor finishes their gift, nothing in subsequent sessions acknowledges that they gave. This experiment runs a three-variant test to find out whether a persistent Donor Thank You badge – and what it does when you tap it – has any effect on 21-day return rates and re-engagement.

Platform Reader-Donor (PRD) progression: Emotional Recognition: Post-Donation Experience (Lightweight badge interaction)

Experiment currently lays out a three bucket variant (A/B/C) testing:

• A: Control – no badge. User donates but does not get a donor badge.
• B: Plain – badge and plain message. User donates and automatically gets a persistent donor badge on the top right bar. Tapping on it opens the plain thank-you message – similar to the first FR experiment.
• C: Delightful – badge and animation. User donates and automatically gets a persistent donor badge. Tapping on it triggers wiki-hearts plus a short acknowledgement message.

A: Control	B: Plain message	C: Delightful experience

Goal

Find out whether the badge – and specifically the tap animation – moves 21-day cumulative retention by at least 1% vs. no badge (A) and the plain message (B). Track whether donors in (C) tap the badge again in later sessions more than those in (B), which would suggest the animation creates a re-engagement habit.

• Experiment targets logged-out donors on mobile web only.
• Variant assignment must be stable across sessions for the same donor (cookie or equivalent persistent bucketing).
• Badge state (shown/hidden) persists across sessions.

The experiment must track:

• 21-day cumulative return rate per variant – 2026-06-12: Redefined in accordance with FR Tech, after community outreach to 30 days for running the experiment
• Badge tap rate, first session (B and C)
• Badge re-tap rate, return sessions (B and C)
• Badge hide/dismiss rate per variant B and C

This task has encompassed the technical scoping of this experiment.

Design requirements

UI designs & workflows have been refined during this phase in collaboration with Design.

• Figma Designs of A/B/C variants treatment

Engineering requirements

• Pre-requisites:
  • Experiment is mobile-only (MinveraNeue skin) and for logged-out users only.
  • Experiment is JS-only.
  • Provide workflow diagram for all three buckets.
  • Setup on top of existing testing framework TestKitchen for bucketing.
• Different requirements for variants:
  • A (Control): Minimal engineering need, only excluding from treatment variants.
  • B (Plain) & C (Delightful) commonalities:
    • Take first FR experiment as orientation point.
    • Clarify where experiment code will live.
    • Clarify handling of sticker location (e.g. absolute positioned item).
    • Donor badge must have per-variant kill switches (hide this badge).
    • Badge visibility state ("shown" / "hidden by user") goes in clientpref keyed to the experiment variant. A donor who hides the badge must not see it again.
    • Hide this badge: Codex component or custom?
  • B (Plain):
  • C (Delightful):
    • Provide animation (with randomization character, e.g. heart colors) .~
    • Animation must use CSS/JS only, no external animation libraries. No layout thrash, no blocking paint.

Testing whatabouts
Recommended: test on test.wikipedia.org

1. In incognito window visit https://thankyou.wikipedia.org/wiki/Thank_You/en?country=US
   1. Visit https://test.wikipedia.org/?mpo=donor-delight-badge:treatment-b-simple or
   2. https://test.wikipedia.org/w/index.php?title=Main_Page&mpo=donor-delight-badge%3Atreatment-c-delightful

Not recommended: browser console, not recommended due to T428694, which was declined for T428992: Remove the override cookie**

$.cookie( 'centralnotice_hide_fundraising', '{"v":1,"created":1779379520,"reason":"donate"}', { path: '/' } );
mw.user.clientPrefs.set( 'minerva-badge', '1' );
mw.testKitchen.overrideExperimentGroup( 'donor-delight-badge', 'control' );
# or
mw.testKitchen.overrideExperimentGroup( 'donor-delight-badge', 'treatment-b-simple' );
# or
mw.testKitchen.overrideExperimentGroup( 'donor-delight-badge', 'treatment-c-delightful' );

Measurement plan

https://docs.google.com/document/d/177uzZs5WyUNYyQcdD6zGQZ9pWBs0hh5HTXR0B6kbHDY/

Instrumentation spec

https://docs.google.com/spreadsheets/d/1NYyinpkV3dswcamBJjLx0AW8P3KzefQyhD85VITIp7U/edit?gid=1635536298#gid=1635536298

Acceptance criteria

• UI designs have been reviewed by an engineer.
• Sub-tasks for this experiment have been created.
• Measurement plan and instrumentation spec have been created and signed of by data analyst.
• All tasks have been resolved.