Background
The Android team ran a 30-day experiment of the customizable donation in Italy, from August 21 - Sept 20, 2025 (T399601) to learn if behavior-based fundraising requests are a good fundraising mechanism in the mobile apps. During PM Review there were improvements noted that were not blockers for this initial experiment but would become must have requirements for scaling the feature any further.
For the second iteration of this feature, we plan to adapt it to run alongside a banner campaign, as an alternate follow-up action to "Maybe later" clicks, so we need to make updates to the flow accordingly.
Hypothesis
If we A/B test the customizable donation reminder as an integrated follow-up action to the banner’s Maybe Later action during an active campaign on Android, users who receive the reminder flow will be 50% more likely to complete a donation as compared to those who receive a subsequent banner impression.
User stories
- As someone who clicks on "maybe later", I want to have more control over how and when I'm reminded to donate, so that I can donate when I'm feeling the most value and gratitude to Wikipedia. T408941
- As someone who has setup a donation reminder from "maybe later", I want to see that reminder right away when I'm eligible for it, and have the ability to customize or snooze reminders if desired. T408943
- As a product manager, I want to set-up the donation reminders as an A/B test against the current "maybe later" experience, so I can understand the impact on average donor value over time. T408945
- As someone customizing my amount or article count in the donation reminder form, I want the keyboard to become active right away, so that I can finish the form faster. T408946
- As a privacy conscious person, I want to be able to clear my donation history in settings, and know that it's been cleared afterwards. T408948
Requirements
See user story subtasks
Designs
Audience
- Users who have clicked "maybe later" during an the banner campaign on English Wikipedia
- Limited to those with a device country of Great Britain, Australia, or Canada.
- Includes logged-in/logged-out users
How we'll measure success
15-day indicator check:
- KR 1.2 (indicator) Users who see an article-count-based reminder are 50% more likely to complete a donation as compared to those who receive a subsequent banner impression
- KR 1.3 (indicator) 20% of unique users who see an impression of their reminder complete a donation
3 month analysis
Validation
- KR 1.1 - 5% increase in average donor value for the reminder group donors when compared to traditional maybe later donors over a 3-month period
- KR 1.2 (indicator) Users who see an article-count-based reminder are 50% more likely to complete a donation as compared to those who receive a subsequent banner impression
- KR 1.3 (indicator) 20% of unique users who see an impression of their reminder complete a donation
Guardrails
- GR 2.1 - Total amount raised from reminders over 3 month period is not more than 20% lower than total amount raised from traditional maybe later over a 1 month period
- GR 2.2 - No more than 30% of users who set up a donation reminder turn it off in Settings
Curiosities
- CR 3.1 - Which article count and amount combination resulted in the highest donations completion rate?
- CR 3.2 - What share of donors in each group are making multiple donations during the period? What is the average number of donations being made? (Can be approximated with How many uniques clicked "donate" on their reminders more than once? If needed)
- CR 3.3 - What is the donation rate of the 1st view of the reminder, 2nd view, 3rd view, etc. After how many reminders do we see the highest disable rate in settings?
- CR 3.4 - What share of each group are new donors vs repeat donors?
- CR 3.5 - How does our reminder set-up rate compare to the web banner email-entry rate (% of Maybe later clicks on web banners that submit their email)?
- CR 3.6 How does our reminder follow-through rate compare to the donations/email opens for web Remind me later?
