Background
In 2024-2025 the iOS team built a personalized Wikipedia Year in Review feature T371946. During Q1 of 2025-2026 the iOS team updated the Year in Review feature and is user testing the new additions T382442.
This year we are bringing the feature to Android. This Epic is for the 2025 version of the Apps Wikipedia Year in Review on Android.
There are four main flows with variants depending if users are editors and depending if users are donors.
We have 2 hypotheses for this project:
- If we release Year in Review on Android and A/B test offering engaged users a reward to save a custom reading list, we will see a 1% increase in overall app retention rate amongst readers offered the reward compared to those who were not.
- If we introduce a Year in Review feature on Android that highlights user impact and includes integrated donor messaging, we’ll drive new donation behavior—and we’ll see a 5% increase in app menu compared to 2024.
Requirements
- Data should cover January 1 - Nov 30, 2025
- We should be able to turn the entire feature, including entry point, on and off using the remote configuration file
- Release date on December 1, 2025
- If users are logged-in they should receive the personalized Year in Review slide and fall back to collective slides when they do not have enough editing or reading data depending on if they have EN wiki as their primary language or not
- If users are logged-out they should receive the collective Year in Review slide depending on if they have EN wiki as their primary language or not
- Reading history by category, or mention a few categories they’ve read articles under
- Most visited article, or which articles they visited multiple times.
- Show articles read by geography (You read X articles about Canada)
- Total time spent reading, and article(s) with the most time spent reading
- Update collective (logged-out) data for English users to match the microsite
- Update collective (logged-out) slides for non-English users with data from Shay about the Mobile Apps
- Allow users to share a summary of the entire review
- Follow the user-flow requirements as detailed in this task: T404269: Year in Review Flow on Android for each user type
Nice-to-have
- Give users a reading list of articles from their year at the end
- Shareable Image collage from the articles they read over the year
- T381678: Year in Review - code cleanup and improvements for 2025
- All external links go to device language version of page, and fall back to English
- Do a push notification when Year is Ready
- Give an in-app way to report issues with the feature
- Have links in feature just show link previews, to keep users inside the feature
- Make the survey free-text multi-line
- Improve transition when iPhone user opens YiR in Landscape, improve iPad and Landscape support
- Add visualizations for rabbit holes or knowledge journeys
- T376441: [L] Add article images / slide carousels into slides
- More detailed in-app contributions stats. Instead of just saying “Thank you”, include the number of donations, amount they donated and when
- Include graphs where we can, for example on “day of the week most read”, many users wanted to see the graph of all the days
- Improve editing stats so they cover the entire year / more than 500+ edits. A few users got hung up on that
- Add additional editing insights - we’re getting additional data with the user contributions API that could be leveraged for additional Editing insights (would require legal review)
- List of articles they edited
- Edit tags like minor, suggested edits - could show a breakdown of types of edits
- New pages created - show how many new articles they created
- Timestamp - could do time of the day or day of the week for Editing
- Show which articles they visited that are more niche or lesser read. (The article in your history that has the lowest pageviews (most niche or rare)
- Show insights by language for Multilingual readers and editors (edits on different languages set in the app, reading in different languages)
- Tell users what kind of reader they are: Repurpose design archetypes and let users know what kind of user group they fit into, or classify the reader as a "type", dancer, hunter, etc (only a very few people asked for this in the survey)
Notes
- Users want collective stats and personalized stats intermingled. They like seeing the collective impact, and want to be compared to the collective. Next year we should not remove collective slides for logged-in users.
- It’s important to continue communicating where the data comes from, and why it’s not able to be cross-platform.
- Notes for when writing UI copy
- Avoid saying "this year" or "last year" within copy, to give flexibility of releasing during either December or January. Ideal format is "in $1", where $1 is 2025
- “You read X articles in 2025”. Instead of “You read X articles this year”
- See if there are friendlier alternatives to talk about how the feature respects privacy. "locally stored" or "local storage” might not mean much to users, maybe something like "calculated on your device, privately" - try to make it so non-technical folks can understand
- Don’t have decimals in the collective statistics. For any collective statistics, we want to use variables ($1, $2) in translate.wiki. This requires having integers for numbers
- “Over 4 billion views” instead of “4.2 billion views” for example
Designs
How will we measure success?
5-day indicators for iOS & Android
Funnel check
- KI 0.1 - At least 48% of active app users during the time period see the Year in Review prompt
- Baseline: 48% of iOS uniques saw the prompt
- KI 0.2 - At least 5% click through rate on the Year in Review announcement
- Baseline: 5% of iOS prompt impressions opened Year in Review
- KI 0.3 - At least 36% of those who enter the feature view more than 2 slides
- Baseline: 36% of those who entered viewed more than 2 slides last year
- KI 0.4 – At least 0.2% of engaged users make a donation through the feature
- Baseline: iOS saw 0.2% of engaged users make a donation.
15-day analysis
Android: A/B Test specific analysis
- KR 4.1 - 1% increase in overall app retention among readers who were offered the reading list (B) as compared to those who were not (A).
- KR 4.2 - 10% increase in Reading list engaged retention measured for the experiment group relative to control engaged
- Reading List Engaged = visiting “saved”
- Practical significance OK
- Baseline from Discover: We saw a 64% increase in retention for the Variant Engaged uniques relative to Control Engaged uniques
- GR 4.3 - At least 80% of users rate the Year in Review reading list feature as satisfactory or neutral
- GR 4.4 - No more than 10% of survey respondents complain about privacy concerns related to the reading list (survey feedback)
- CR 4.5 - How did the conversion rate of the reading list compare to the app icon?
- % of impressions of the reading list modal that clicked “Create”
- vs % of slide impressions of the icon slide that turned on the icon
30-day analysis
Validation (the same for both Android & iOS)
- KR 1.1 - 5% increase in donations app menu donations as compared to a similar time period in 2024.
- KR 1.2 - 5% increase in return rate for engaged users compared to baseline of 77%.
Guardrails (the same for both Android & iOS)
- GR 2.1 - Fewer than 50% of feedback about the donation aspects are negative
- GR 2.2 - At least a 80% of users rate the Year in Review feature as satisfactory or neutral
- GR 2.4 - At least 10% of engaged users make it to the contributor slides
- Engaged = viewed at least 2 slides
- If not met: which slides had the highest drop off?
Curiosities (the same for both Android & iOS)
- CR 3.1 What were the main themes in the user feedback from the Year in Review survey? Did we see a decrease in requests for personalization from 2024?
- Baseline: Last year, 60% of free-text responses asked for personalization.
- CR 3.2 - What did the funnel look like for logged-out users? What % logged-in, created an account, and then viewed the Year in Review afterwards.
- CR 3.3 - How do the donations received through Year in Review differ from other app menu donations, and banner donations? Average gift, share of new donors
- App menu donations from a comparable time period (when YiR is not running)
- CR 3.4 - What was the overall donation rate, and which 2 slides had the highest donation rate?
- CR 3.5 - How many people shared their highlights? What was the most commonly shared individual slide?
- CR 3.6 - What share of engaged users with year in review enabled the custom icon (in settings for from within the feature)?