This page contains updates on the Growth team's work to increase new editor activation and retention.
The planning for this work began in June 2018.
The team works on multiple projects at a time, but they will all be discussed on this page.
How to get involved
It is important that our work is grounded in the reality of the communities we hope to help.
If you have thoughts or ideas around this project or our team's work, please comment on this project's talk page.
On Wikipedias with GrowthExperiments enabled, we send newcomers a notification 48 hours after account creation. Previous to this experiment we sent a “Keep going” notification to the small minority of newcomers who had completed a Suggested Edit (about 3%), and a “Get started” notification to all other accounts (about 97% of accounts receive the “Get started” notification). Previous Leveling Up experiment analysis showed that these notifications lead to a significant increase in newcomer activity in the week following the notifications being sent.
Knowing that the majority of newcomers receive the “Get Started” notification, we aimed to make it more engaging and personalized (T400047). Instead of sending the same message to all newcomers who had not completed a Suggested Edit, we created two segments:
“Re-engage” notification: Newcomers who had made at least one successful edit(but not a Suggested Edit) within 48 hours.
(new) “Get Started” notification: Newcomers who had not edited during the first 48 hours.
And we also updated the copy of the “Keep going” notification, but we did not change the targeting. It was still sent to:
New accounts who had completed at least one Suggested Edit within 48 hours after account creation.
Notification experiment findings:
Most newcomers tapped outside the main call to action button. In this experiment both the button and the surrounding area linked to the same destination, but this behavior is important for future tests. It suggests that newcomers may not recognize the difference between primary and secondary links in notifications.
More relevant or personalized messages resulted in higher engagement Overall, when we look at the total Click Through Rate (CTR) for the treatment group vs. the total CTR for the control group, we see that the treatment group had a higher CTR. This supports the idea that more relevant and personalized messages drive higher engagement. So we are releasing the new “Get Started” and “Re-engage” notifications to all Wikipedias that have Growth Experiments enabled.
We actually see a slightly higher CTR for the control group’s “Keep Going” message. Although the results are not statistically different, we decided to revert back to the original “Keep Going” message. We hypothesize that because the original message includes a user’s Suggested Edit count, the message may feel more personalized and engaging.
Verification Email
The first email new accounts receive can be perceived as alarming, especially with language about the user's IP address. This may deter users from confirming their email or engaging further. We aim to revise this email to make it more welcoming and informative, improving the overall onboarding experience and increasing email confirmation rates. The new verification email will be released at English Wikipedia (T396155), and then at Commons and Wikidata soon after (T410971). We will monitor the impact on email verification rates for newly registered accounts before deciding to release further (T400843)
When users attempt to edit while logged out, they encounter a jarring warning message that emphasizes editing as an IP or temporary account. This warning can feel abrupt and discouraging, especially to newcomers. The warning also frames logged-out editing (IP editing / Temporary Accounts) as the primary path rather than encouraging users to create an account.
There is an opportunity to soften the logged-out warning message and guide users toward account creation in a more supportive, motivating way. By improving the tone and visual hierarchy, we can better communicate the benefits of creating an account and make the overall experience less intimidating for newcomer (T407497).
For scaling Add a Link, the team has shared its deployment plan and is working with Japanese, Chinese, and Urdu Wikipedias on configuration details. The Growth team will handle leading indicator monitoring directly to support other analytics priorities (T407448).
For the Revise Tone structured task, instrumentation work is underway and the team is wrapping up engineering tasks necessary for the initial pilot A/B test (T396162).
The Growth team is involved in several research initiatives to help guide our future work, all three research projects are wrapping up and we should have findings to share soon!
Progression System - The research examined motivations, challenges, and feedback on a prototype system intended to help editors build confidence, develop skills, and contribute more constructively over time.
Mobile Web Editing Research - This project combines quantitative and qualitative data, community feedback, and user journey analysis to identify actionable insights to enhance the mobile editing experience.
Newcomers Survey - This project surveys successful newcomers on English Wikipedia to understand their early editing experiences, tool use, and community interactions.
Update 2025-10-03
Kirsten and Benoît present at the Wikimedia CEE meeting
CEE Meeting
Growth team representatives joined the Wikimedia CEE meeting in Greece to share a session about retaining beginners and improving content moderation: an inclusive and sustainable future for Wikipedia contributors. Many communities face a decline in volunteer engagement. Newcomers often leave soon after joining, while experienced editors struggle to manage increasingly complex workflows and overwhelming backlogs. We presented the Contributors Strategy and the different features and workflows that can help communities to address these challenges. We listened to the specific needs of the CEE communities to help guide the Contributors team’s work.
We wrapped up a second round of usability testing and have initial functionality available in patch demo. We have more work to do to polish the UI and UX, and we are working with the Machine Learning team to ensure we have a stable data pipeline of suggestions. Early plans have been shared with communities.
The Search Platform completed work to help support the Revise Tone task: T403593CirrusSearch should allow filtering on page creation and last edit timestamps.
The Editing team completed work to ensure VE cab support the Revise Tone UX: T403072Create proof of concept for launching VisualEditor from a "deep link".
The Growth team continues to work on the onboarding and user flow. We also completed usability testing with the following findings:
When asked to assess a Wikipedia article without guidance, most participants struggled to identify tone issues.
Participants generally viewed the onboarding tutorial positively, describing it as a helpful refresher on Wikipedia standards and useful for beginners. However, some wanted more context-specific guidance tied to actual editing.
While the tutorial helped participants recognize emotional language and bias, it did not sufficiently prepare them for real editing scenarios. When revising text, nearly all participants failed to improve tone meaningfully. Most spent 1–4 minutes per paragraph and relied on external AI tools for validation.
Growth worked with Product Ambassadors and Movement Communications to review Revise Tone suggestions and categorize suggestions to better understand how we can improve the model or how we can improve how we filter out some suggestions. The main learnings are:
Article types
We should include articles about people, articles about sports, and articles tagged with relevant templates
We should exclude articles about politics
Filters
We should filter out references (eg. using <ref>)
We should try to filter out block quotes and inline quotes
We should exclude brand new articles that are awaiting New Page Patrol review
We should exclude articles that are marked for deletion
Analysis
We should be sure to evaluate not just whether the model is identifying tone issues, but also whether the edits newcomers are making are actually fixing them or just changing something that doesn't improve them (or worse, changing them in a way that hides the tone issue without fixing it).
We should evaluate whether newcomers who take the onboarding quiz are able to interact with this task more effectively.
The Growth team plans to release the "Add a Link" Structured Task to 100% of accounts at English Wikipedia on Tuesday, September 2nd (T395524). We have discussed plans and suggested Community Configuration modifications here: "Add a Link" experiment and next steps.
Growth features for Wikidata
After inquiries about if Growth features and Mentorship would work for Wikidata, we have enabled Growth features at Beta Wikidata to allow for testing and discussion (T400937). Although some features, like Suggested Edits, are Wikipedia-specific, the Growth team designed most features to be more wiki-agnostic.
Update 2025-08-22
The Growth team continues to support two complementary research priorities and is eager to learn from the outcomes:
This project combines quantitative and qualitative data, community feedback, and user journey analysis to identify actionable insights that will enhance the mobile editing experience.
The Growth team is progressing on technical architecture, onboarding design, and early user testing of designs. Growth is targeting a Q2 A/B test, with constructive edits by newcomers as the primary success metric.
Recruitment is underway with support from Product Analytics and the legal-reviewed privacy policy in place. Work continues on literature review and discussions about alignment with related ideas. The discussion guide is finalized and newcomer interviews will start soon.
To help developers and designers interact with and build features that are Community Configurable, we have developed an Example extension that helps explain how each configuration field works: CommunityConfigurationExample
Update 2025-08-08
Wikimania 2025 is nearly complete! The Growth team participated in several sessions:
Growth's Product Manager, Kirsten Stoller, presenting at Wikimania
This session served as a listening space where organizers shared their experiences in introducing and training newcomers with the Growth features—tools developed by the WMF Growth team to help newcomers overcome barriers and continue contributing to Wikipedia. Participants discussed common questions organizers receive from newcomers about using Growth features. The session concluded with a brainstorming conversation about possible paths forward and how we might join forces to strengthen newcomer growth across the movement.
This lightning talk explored how Structured Tasks support newcomers in taking their first successful steps on Wikipedia. It broke down how these tasks work, shared data-driven insights on their impact, and highlighted ways communities have configured and improved them. The session included a quick demo of “Add a Link,” examples of how different wikis adapted Structured Tasks to their needs, and insights on why these tasks are a key tool for making editing more accessible and sustainable, especially for mobile-first contributors, by increasing newcomer engagement and retention.
Representatives from the WMF Contributor teams: Moderator Tools, Connections, Editing, and Growth teams
With active editor numbers declining, Wikimedia’s Contributor Strategy aimed to create a clearer, more engaging path for participation. This session, led by the WMF Contributors group with involvement from the Editing, Growth, Moderator Tools, and Connection (formerly Campaigns) teams, highlighted efforts to streamline contributor experiences, offer structured and mobile-friendly workflows, and foster meaningful engagement. Participants learned about ongoing initiatives and shared feedback to help shape a more inclusive and sustainable future for Wikimedia contributors.
We’re finalizing two improvements that reflect community feedback (T389288):
A new rule to prevent linking country names on English Wikipedia.
A configuration option allowing communities to limit the task to newer editors.
Engineering work is nearly complete, and we’re coordinating rollouts.
Our “Structured Tasks” lightning talk was accepted for Wikimania 2025. The session will highlight how tasks like “Add a Link” and “Add an Image” help newcomers make their first edits more confidently and successfully.
Designs are finalized, and engineering has begun on a supportive notification (Echo/email) for new accounts with 0 edits. Early research indicates this type of nudge increases editing activity
New Growth Ambassador
The Growth team and Editing teams are excited to start working with a new English Wikipedia Product Ambassador: User:Sdkb-WMF.
Sdkb began volunteer editing in 2012 and has been active since 2018. Sdkb became an English Wikipedia administrator in February 2024 and contributes to many different areas where there are unmet needs, with newcomer support being a major focus.
Confirmation email
We are exploring ways to simplify and improve the initial account confirmation email newly registered users receive. (T215665)
Spanish Wikipedia increases Mentorship rollout, so that 70% of newcomers receive a Mentor when they create an account on Spanish Wikipedia (T392869). The Growth team will increase Mentorship to 85% of newcomers soon, unless we hear concerns from Spanish Wikipedia mentors (T394867).
Surfacing “Add a Link” on Mobile (Constructive Activation experiment) is still underway on on pilot wikis (fr, fa, id, pt, arz). Initial experiment data is under analyst review; we hope to have a final report to share next month.
We are progressing with work to allow communities to limit "Add a Link" to newer accounts (T393769)
We're implementing an improvement to the enwiki "Add a Link" model to ensure link suggestions more closely align with the enwiki Manual of Style.
Designs are finalized, and engineering has begun on a supportive notification (Echo/email) for new accounts with 0 edits. Early research indicates this type of nudge increases editing activity (T392256)
Communities will soon have the ability to limit “Add a link” to newer editors. This helps address concerns about misuse by experienced editors or overburdening moderators.
The change includes moving the "Add a Link" task to the “Starter” difficulty level and notifying newcomers when they’ve reached the usage threshold.
You can see related designs in the following tasks:
Move Add a link task to the newcomer task difficulty level "Starter" (T393767)
Disable task type X for a user when the configured threshold is reached (T393769)
Send a notification to newcomers when Add a link threshold is reached (T393771)
We're addressing complaints about irrelevant link recommendations. A technical investigation is underway.
Together, these improvements aim to make “Add a Link” more useful, more trusted, and more appreciated by communities. Let us know if you'd like to get involved or provide feedback!
Once an account confirms their email address, we will remove the "email module" from the Newcomer Homepage (T275155).
We are also taking some time to clean up old code and feature flags to decrease the complexity of working within GrowthExperiments extension (T379568 & T379566).
We aim to increase the data cap for the the Impact Module, so the metrics are more relevant to experiences editors who routinely bypass the edit count limit (T341599).
The data on the Mentorship dashboard was not updating for a period of time, due to the associated script causing issues. Growth Engineers have had to work on various changes to improve performance and batch data handling (T391695).
Engineering Maintenance
We have completed several tasks to clean up and remove old code and feature flags (T379566).
We provided the Spanish Wikipedia community with the “Add an Image” data they were seeking (T390075).
Update 2025-04-11
Growth team offsite
As a distributed team working around the world, we only have the chance to get together once each year to work together! This year we met with a few other Core Experience teams in Philadelphia, to draft initial plans that align with the Wikimedia Foundations 25/26 Annual Plan.
We will focus on the WE1.1 Key Result: Key result WE1.1: Increase constructive edits [i] by X% for editors with less than 100 cumulative edits, as measured by experiments by the end of Q2.
i. "Constructive edits" = edits that are not reverted within 48 hours of being published
The Surfacing Structured tasks A/B test on pilot wikis is was launched across six wikis, covering both desktop and mobile (T386029) . While impressions are expected to be higher than in the alpha test, additional work is underway to increase the number of link suggestions without delaying release (T386250).
Growth data
We are continuing work to migrate GrowthExperiments client code to statslib (T359352).
Update 2025-02-27 "Add a Link" rollout and experiment on English Wikipedia
On English Wikipedia, the incremental rollout continues smoothly, with expansion to 15% of newcomers this week (T386029)
The analysis of the English Wikipedia “Add a Link” experiment is in progress. Preliminary data aligns with previous pilot results (indicating that it’s likely that results will show that the “Add a Link” task increases constructive activation on English Wikipedia), however, a final QA of the activation analysis is underway before exact figures will be shared. Once the constructive activation metrics are finalized, Product Analytics will shift focus to assessing the impact on newcomer retention. (T382603)
Mentorship was released to 100% of newcomers at English Wikipedia on February 17, 2025 (T384505).
We will start to work with Spanish Wikipedia to increase Mentorship to 100% (T285235). All other wikis that utilize Mentorship have the feature released to 100% of newcomers.
Update 2025-02-14 "Add a Link" & Constructive Activation Experiments
Next week, we’ll increase the “Add a link” rollout on English Wikipedia to 15% (T386029).
The Growth team reviewed community feedback on the “Add a link” task and documented key themes and potential solutions. As we expand exposure to this task, we aim to incorporate improvements based on insights from experienced editors and moderators.
We are also exploring enhancements to the “Add a Link” models. The Research team is investigating a new approach that ensures more regular training for the models and reduces infrastructure complexity while maintaining language-specific models.
We’re exploring ways to encourage newcomers who leave an edit session without saving to try again. Our idea is to provide a supportive message with a clear next step—such as a Structured Task—to guide them back into editing. This concept is in the early design and community feedback stage (T385541), and we’d love to hear your thoughts!
After a discussion with English Wikipedia Mentors, there was agreement to release Mentorship to all newcomers on English Wikipedia. This week bumped up the percentage of newcomers who receive Mentorship to 75%, and on February 17th 100% of new accounts will be paired with a experienced editor Mentor. T384505
Newcomers often struggle to find a starting point on Wikipedia. The Growth team has developed features like the Newcomer Homepage to provide guidance through suggested edits, mentorship, and impact tracking. However, editing isn’t just about individual contributions—it’s also about collaboration. While experienced editors easily find events like edit-a-thons and writing campaigns, newcomers often miss out.To bridge this gap, the Growth team introduced the Community Updates module on the Newcomer Homepage, making community events and initiatives more visible and accessible to new editors.
The Growth team’s Community Updates module, launching globally on February 11, 2025, helps by making community events and initiatives more visible on the Newcomer Homepage. The feature will be disabled by default, allowing Community Admins to decide how (or if) to utilize it. Trials on Arabic, Czech, French, and Spanish Wikipedias provided insights for future improvements ideas. If your community hosts events, consider setting up a Community Update to welcome and engage newcomers. We’re excited to see how communities use this feature—let us know your thoughts!
Update 2025-01-31 Babel Extension Now Supports Community Configuration on All Wikis
The “Add a link” rollout on English Wikipedia has increased to 10% (T384551). We will start to review results from this A/B test soon (T382603).
In late 2024, German Wikipedia independently enabled the “Add a Link” task and has since been actively discussing its impact. We’ve received a request to provide data on the "Add a Link" structured task for dewiki (T385103).
Update 2025-01-24 Resuming Work on Helping New Account Holders Start Editing
We're continuing our efforts to support constructive activation for new account holders. This week, we’re focusing on preparations for a pilot experiment testing the impact of surfacing "Add a Link" tasks to brand-new account holders. Here’s what’s happening:
Key Updates
Improving the "Add a Link" Task Flow We’re updating the task flow to surface only the specific link tapped on, when the task is initiated from the Read view (T381450).
This change aims to simplify the user experience (UX) and address feedback from patrollers.
Patrollers have shared that when some links in an edit are incorrect, it creates extra work to fix the edit instead of a simple "revert." This adjustment will help mitigate that frustration.
Expanding Link Suggestions We’re significantly increasing the number of link suggestions available for our pilot wikis:
Smarter resource allocation: We’ll store when no link suggestions were generated for a page and avoid reprocessing the article until it’s edited again (T382270). This prevents wasted resources.
More tasks per topic: We’re increasing the minimum number of "Add a Link" suggestions generated for pilot wikis, scaling up suggestions tenfold (T383714).
Addressing Community Feedback
We’re also planning to address feedback from experienced editors about "Add a Link" tasks. Some ideas include:
Adding additional rejection reasons for edits (T384403).
Allowing admins to define rules for linking (or not linking) via Community Configuration (T384405).
Explore more improvement ideas listed as subtasks of T276517.
We’ll continue iterating on these features to support both newcomers and experienced contributors alike. Feedback is always welcome!
Update 2025-01-17 Engineering Maintenance & Add a Link Release
Engineering Maintenance
We kicked off 2025 with a focus on technical priorities that are often challenging to address amidst endless product initiatives and WMF Annual Plan commitments. It's important to balance new feature development with engineering maintenance because a stable, well-maintained code ensures that new innovations are reliable, scalable, and sustainable over time. Without regular maintenance, technical debt can accumulate, leading to system inefficiencies, increased costs, and potential disruptions to user experience. Here are the highlights of what we worked on:
Remove Marketing experiment related code from GrowthExperiments T382499
Create a generic Campaign parameter for the Donor Thank You page T380405
Switch GETempLinkRecommendationSwitchTagClearHook to true at all wikis T379522
Running extensions/Babel/maintenance/migrateConfigToCommunity.php with BabelUseCommunityConfiguration=false results in a fatal error T383904
PHP Warning: Stats: Cannot add labels to a metric containing samples for 'update_mentee_data_seconds' T383477
Migrate GrowthExperiments.Mentorship Module to statslib T374048
Ensure ORES topics are not loaded through legacy CommunityConfiguration code T367576
.jpg)
