Page MenuHomePhabricator

Impact Module: improvements for former newcomers
Closed, DeclinedPublic3 Estimated Story Points

Description

User story & summary:

As a former new editor, I want to the Impact module to remain relevant as I edit more, so that I can continue to use it.

Background & research:

This task is important because ideally the homepage and impact module will remain relevant to editors as they continue to edit and deepen their involvement in the Wikimedia movement.

Due to performance reasons, the impact module surfaces the last 1,000 edits you made. This number is used to determine your longest streak as well.

The impact module main audience is newcomers. We designed this impact module with the average number of edits and days of presence of typical newcomers. This is why the numbers aren't reflecting the edits of very active users.


Example of how the Impact Module edit cap impacts the Impact Module for a highly-active editors:

Screenshot 2025-03-17 at 4.16.01 PM.png (874×590 px, 60 KB)

(Thanks is capped, the Longest streak only applies to the last 1,000 edits, and the Recent Activity graph is limited to only the last 1,000 edits).

Related:

The Wikipedia Mobile apps are interested in potentially using Impact Module data for: T371946: [Epic] Wikipedia Year in Review on iOS App .
However, they would ideally like the upper limit increased to make this data more meaningful for app users.

Potential Improvements:

There are many different improvements we could consider:

  • Increase the underlying query to a higher upper limit: 10,000?

Screenshot 2023-07-11 at 9.55.25 AM.png (690×782 px, 92 KB)

Acceptance Criteria:

Given I'm a logged in editor,
When I visit my Homepage or Impact module,
Then Impact module data includes my last (10,000?) edits

  • Release and test on beta: beta eswiki
  • On one pilot wiki (enwiki): increase the Impact Module data cap to a higher but reasonable limit 10,000 edits / 1,000 Thanks - June 1, 2025
  • Monitor the change, and rollout to all wikis if there aren't performance concerns - June 8, 2025

Related Objects

Event Timeline

There are a very large number of changes, so older changes are hidden. Show Older Changes

Change #1165866 had a related patch set uploaded (by Urbanecm; author: Urbanecm):

[operations/mediawiki-config@master] [Growth] enwiki: Decrease wgGEUserImpactMaxEdits to 1000

https://gerrit.wikimedia.org/r/1165866

Change #1165865 merged by jenkins-bot:

[operations/mediawiki-config@master] [Growth] Move Impact limit configuration to ext-GrowthExperiments

https://gerrit.wikimedia.org/r/1165865

Change #1165866 merged by jenkins-bot:

[operations/mediawiki-config@master] [Growth] enwiki: Decrease wgGEUserImpactMaxEdits to 1000

https://gerrit.wikimedia.org/r/1165866

Mentioned in SAL (#wikimedia-operations) [2025-07-02T12:48:57Z] <urbanecm@deploy1003> Started scap sync-world: Backport for [[gerrit:1165865|[Growth] Move Impact limit configuration to ext-GrowthExperiments (T341599)]], [[gerrit:1165866|[Growth] enwiki: Decrease wgGEUserImpactMaxEdits to 1000 (T398418 T341599)]]

Mentioned in SAL (#wikimedia-operations) [2025-07-02T12:51:16Z] <urbanecm@deploy1003> urbanecm: Backport for [[gerrit:1165865|[Growth] Move Impact limit configuration to ext-GrowthExperiments (T341599)]], [[gerrit:1165866|[Growth] enwiki: Decrease wgGEUserImpactMaxEdits to 1000 (T398418 T341599)]] synced to the testservers (see https://wikitech.wikimedia.org/wiki/Mwdebug). Changes can now be verified there.

Mentioned in SAL (#wikimedia-operations) [2025-07-02T12:58:40Z] <urbanecm@deploy1003> Finished scap sync-world: Backport for [[gerrit:1165865|[Growth] Move Impact limit configuration to ext-GrowthExperiments (T341599)]], [[gerrit:1165866|[Growth] enwiki: Decrease wgGEUserImpactMaxEdits to 1000 (T398418 T341599)]] (duration: 09m 42s)

Trizek-WMF subscribed.

I move this as "not ready to announce" as it is still on the works. Please move it back to "backlog" when ready!

Thanks for adding me as a subscriber on this, Kirsten. I think many more experienced editors will be glad to see this feature. Raising the cap to 10,000 is unlikely to be high enough to produce meaningful data for a lot of folks, though, as many active editors have significantly more edits than that.

Raising the cap to 10,000 is unlikely to be high enough to produce meaningful data for a lot of folks, though, as many active editors have significantly more edits than that.

Thanks for the feedback!

The data cap primarily affects metrics calculated and displayed for the past 60 days. While we know there are a few prolific contributors who may exceed 10,000 edits during that period, the number impacted should be much smaller than the group we were previously “missing” under the 1,000-edit cap.

That said, we are also in discussions with the Data Engineering team about contributor data needs. Ideally, we would have easier access to cached contributor metrics in the future, rather than continuing to rely on the custom Impact metrics built by the Growth team for the Impact Module. So hopefully in the future we can explore adding in other metrics or increasing the data cap further.

Hello @KStoller-WMF,
For Tech News - What wording would you suggest as the content, and when should it be included? Thanks!

Michael changed the task status from Open to Stalled.Jul 9 2025, 7:47 PM

@UOzurumba: Please note that this is currently not enabled anywhere except test-wiki, and that it is fully blocked on T398500: [timebox: 3 days] Impact module: Support larger wgGEUserImpactMaxEdits which is not in our current sprint.

@UOzurumba: Please note that this is currently not enabled anywhere except test-wiki, and that it is fully blocked on T398500: [timebox: 3 days] Impact module: Support larger wgGEUserImpactMaxEdits which is not in our current sprint.

Thank you!

KStoller-WMF lowered the priority of this task from High to Medium.Sep 22 2025, 4:03 PM

Random note: I casually stumbled into the fact that Special:Impact is translated to Especial:Incidencia. Incidencia means incident in spanish. Which is far from the correct meaning for the page. Is this something to rise with the community? Or maybe @IZapico-WMF can make the change if it's not problematic?

Sidebar on translation:

Yes, it is the wrong word, but I sort of see where it came from. Incidencia also means "incidence" in Spanish, which is an epidemiological term which measures the rate or frequency of something and in that sense, where a researcher might write about the incidencia de la influenza in a journal, they might talk about the impacto de la gripe to a group of secondary school educators. I suspect that that (or something like it) was behind the choice of incidencia.

However, it is still the wrong word, because incidencia measures something bad (disease, etc.) and Special:Impact measures something good (streaks, total edits, thanks), and because incidencia only has that meaning in academic jargon, and our terms should be more familiar, common terms for a general public. Even though translators must always be on the lookout for false friend traps, impacto is the right word in this case.

Given the positive outcome from T398500: [timebox: 3 days] Impact module: Support larger wgGEUserImpactMaxEdits I believe we can unstall this task and go for a paced rollout of the limit bump:

  1. Bump the limit in enwiki or similar big site, observe query and job performance and make sure user experience is as expected
  2. Bump the limit in 30-40% of remaining wikis, observe performance
  3. Bump the limit to 100%
  4. Make the new limit the default, this can have its own task
Sgs changed the task status from Stalled to Open.Jan 9 2026, 1:18 PM

meh, the change is suspicious to have caused T414080 so we should tackle that one before proceeding. Maybe even we need to re-think how to store big user impacts as this problem has shown without even bumping the limit.

Random note: I casually stumbled into the fact that Special:Impact is translated to Especial:Incidencia. Incidencia means incident in spanish. Which is far from the correct meaning for the page. Is this something to rise with the community? Or maybe @IZapico-WMF can make the change if it's not problematic?

Sidebar on translation:

Yes, it is the wrong word, but I sort of see where it came from. Incidencia also means "incidence" in Spanish, which is an epidemiological term which measures the rate or frequency of something and in that sense, where a researcher might write about the incidencia de la influenza in a journal, they might talk about the impacto de la gripe to a group of secondary school educators. I suspect that that (or something like it) was behind the choice of incidencia.

However, it is still the wrong word, because incidencia measures something bad (disease, etc.) and Special:Impact measures something good (streaks, total edits, thanks), and because incidencia only has that meaning in academic jargon, and our terms should be more familiar, common terms for a general public. Even though translators must always be on the lookout for false friend traps, impacto is the right word in this case.

Thanks for raising this. As you mentioned here, “Impacto” would be a much better translation in this case.
I’ve filed a follow-up task (T414285) to request changing the eswiki alias for Special:Impact from “Incidencia” to “Impacto”, since it seems that cannot be done via TranslateWiki strings.

I was quoting; @Sgs raised it. ☺

(I thought quoting automatically credited, but I guess not.)

Sgs added a subscriber: Cyndymediawiksim.

meh, the change is suspicious to have caused T414080 so we should tackle that one before proceeding. Maybe even we need to re-think how to store big user impacts as this problem has shown without even bumping the limit.

Not sure if it was a direct fallout of the change here but the mitigation applied for T414080 seems to have worked so we're ready again to give a try to the limit bump to 10K. I plan to do the config change and loop DBAs for awareness.

Change #1254216 had a related patch set uploaded (by Sergio Gimeno; author: Sergio Gimeno):

[operations/mediawiki-config@master] GrowthExperiments: increase edit and thanks query limit II

https://gerrit.wikimedia.org/r/1254216

The more experience a user has, the more they should shift away from using the Impact module and toward the Toolforge Edit counter tool, which contains all of the Impact stats (except for 'streak', which is probably the least valuable of them, especially when one is no longer a newbie).

Note: the tool does have some limits, but they are stratospheric; one million edits is a hard limit. For a typical user, results take a second or two. The #1000 top editor (99k edits) takes about 9 seconds to return complete results. The #25 editor (918k edits) takes over a minute. For a user with > 1M edits, the tool returns immediately but with only one estimated statistic.

I would propose we implement a "shift" in the Impact module, thus:

  • add an Ec tool link below the current stats when the users total edits exceeds X (X =~ 1,000?)
  • replace all content in the module with a 'Please see...' and a link to the Ec tool when total edits exceeds Y (Y =~ 5,000?)
  • drop the Impact module when total edits exceeds Z (Z =~ 10,000?).

So, for their first 1,000 edits, the modules would appear as it does now, and then at thresholds X, Y, and Z it would shift them increasingly towards using the tool.

To help support this, and come up with good values for X, Y, and Z, it would be helpful to see a line graph of number of clicks on the Homepage tab as a percentage of all clicks by a given user (y-axis) plotted against total number of edits (x-axis). This might be useful for other growth apps as well.

Change #1254216 merged by jenkins-bot:

[operations/mediawiki-config@master] GrowthExperiments: increase edit and thanks query limit II

https://gerrit.wikimedia.org/r/1254216

Mentioned in SAL (#wikimedia-operations) [2026-03-18T13:31:50Z] <sgimeno@deploy2002> Started scap sync-world: Backport for [[gerrit:1248095|filebackend: Remove outdated comment]], [[gerrit:1254216|GrowthExperiments: increase edit and thanks query limit II (T341599)]]

Mentioned in SAL (#wikimedia-operations) [2026-03-18T13:33:56Z] <sgimeno@deploy2002> matmarex, sgimeno: Backport for [[gerrit:1248095|filebackend: Remove outdated comment]], [[gerrit:1254216|GrowthExperiments: increase edit and thanks query limit II (T341599)]] synced to the testservers (see https://wikitech.wikimedia.org/wiki/Mwdebug). Changes can now be verified there.

Mentioned in SAL (#wikimedia-operations) [2026-03-18T13:40:36Z] <sgimeno@deploy2002> Finished scap sync-world: Backport for [[gerrit:1248095|filebackend: Remove outdated comment]], [[gerrit:1254216|GrowthExperiments: increase edit and thanks query limit II (T341599)]] (duration: 08m 47s)

@Mathglot Thanks for the feedback, you raise some good points. I've documented your ideas in this task: T420542: Design exploration: Evolve Impact module for experienced editors and patrollers

Your ideas are very aligned with the work that both the Connection-Team, Growth-Team, and the Moderator-Tools-Team are prioritizing next fiscal year. While you may not see immediate changes based on this specific idea, it will inform how we think about the Impact Module evolving as editors gain more experience. This is something our teams plan to explore further as work begins in the new WMF annual plan in July.

Sgs moved this task from Estimated tasks backlog to FY2025-26 Q3 Sprint 5 on the Growth-Team board.
Sgs edited projects, added: Growth-Team (FY2025-26 Q3 Sprint 5); removed: Growth-Team.

The limit to 10K is now enabled in enwiki, giving some days to monitor health signals before rolling out to more wikis.

Change #1259132 had a related patch set uploaded (by Sergio Gimeno; author: Sergio Gimeno):

[operations/mediawiki-config@master] GrowthExperiments: scale edit and thanks query limit to more wikis

https://gerrit.wikimedia.org/r/1259132

Change #1259132 merged by jenkins-bot:

[operations/mediawiki-config@master] GrowthExperiments: scale edit and thanks query limit to more wikis

https://gerrit.wikimedia.org/r/1259132

Mentioned in SAL (#wikimedia-operations) [2026-03-26T13:30:29Z] <sgimeno@deploy2002> Started scap sync-world: Backport for [[gerrit:1259132|GrowthExperiments: scale edit and thanks query limit to more wikis (T341599)]]

Mentioned in SAL (#wikimedia-operations) [2026-03-26T13:32:31Z] <sgimeno@deploy2002> sgimeno: Backport for [[gerrit:1259132|GrowthExperiments: scale edit and thanks query limit to more wikis (T341599)]] synced to the testservers (see https://wikitech.wikimedia.org/wiki/Mwdebug). Changes can now be verified there.

Mentioned in SAL (#wikimedia-operations) [2026-03-26T13:39:47Z] <sgimeno@deploy2002> Finished scap sync-world: Backport for [[gerrit:1259132|GrowthExperiments: scale edit and thanks query limit to more wikis (T341599)]] (duration: 09m 17s)

Change #1268061 had a related patch set uploaded (by Urbanecm; author: Urbanecm):

[operations/mediawiki-config@master] [Growth] Decrease user impact limits back to the defaults

https://gerrit.wikimedia.org/r/1268061

Change #1268061 merged by jenkins-bot:

[operations/mediawiki-config@master] [Growth] Decrease user impact limits back to the defaults

https://gerrit.wikimedia.org/r/1268061

Mentioned in SAL (#wikimedia-operations) [2026-04-06T08:35:40Z] <urbanecm@deploy1003> Started scap sync-world: Backport for [[gerrit:1268061|[Growth] Decrease user impact limits back to the defaults (T422288 T341599)]]

Mentioned in SAL (#wikimedia-operations) [2026-04-06T08:37:18Z] <urbanecm@deploy1003> urbanecm: Backport for [[gerrit:1268061|[Growth] Decrease user impact limits back to the defaults (T422288 T341599)]] synced to the testservers (see https://wikitech.wikimedia.org/wiki/Mwdebug). Changes can now be verified there.

Mentioned in SAL (#wikimedia-operations) [2026-04-06T08:46:31Z] <urbanecm@deploy1003> Finished scap sync-world: Backport for [[gerrit:1268061|[Growth] Decrease user impact limits back to the defaults (T422288 T341599)]] (duration: 10m 50s)

After discussion with @Urbanecm_WMF and discussing the recent incident T422288: Massive performance problems apparently related to the GrowthExperiments extension, we agree that the impact module is not in a stable shape to scale its edit limits broadly.

Thanks, @Sgs.
I agree, and I don't think we should continue to invest more time here, but should plan ahead and work towards a more scalable and stable solution, like: T422563: [EPIC] Impact Module: move to Wikimedia Analytics API