Page MenuHomePhabricator

[Analysis] Evaluate core editing metrics before and after VE default (mobile, en.wiki)
Open, HighPublic

Description

T431858 will make it so newcomers editing en.wiki on mobile will be shown VE by default.

This task involves the work of analyzing core editing metrics in the 2 weeks before and after this change.

Analysis approach

We'll compute and compare each of the metrics below for the 2 weeks before and after VE becomes enabled as the default mobile editing interface for newcomers at en.wiki.

Audience

Distinct registered users with ≤100 cumulative edits, to align with DE1.1 KPI definition.

Core editing metrics

MetricDefinitionLogged-inLogged-outNotes
[DE1.1 KPI] Distinct registered users who make a constructive editProportion and number of registered users with 100 or fewer edits who publish as least one constructive edit on mobile web (edit was not reverted within 48 hours) on a main namespace. Where the denominator in this context refers to users who make an edit attempt (defined as the interface reaching ready). [1]✅ Yes❌ Nonot implemented
Edit completion rateThe proportion of editing sessions on a main namespace that successfully save an edit (saveSuccess) after the editor interface was ready (ready).✅ Yes✅ Yes
Constructive edit rateProportion of published main namespace edits that are constructive (not reverted within 48 hours of being published)✅ Yes✅ Yes
Revert rateProportion of published main namespace edits that are reverted within 48 hours of being published✅ Yes✅ Yes
24-hr Constructive activationProportion of newly registered accounts that make at least one constructive edit in the main namespace within 24 hours of account registration, where constructive means the edit was not reverted within 48 hours of being published.✅ Yes✅ Yes
2nd week retentionPercentage of contributors who return to publish at least one edit (any namespace) during their second week (day 7-14 ) after their initial constructive edit on a main namespace after experiment exposure.✅ Yes❌ NoNot instrumented
Edit abandonment rate (w/ and w/o changes)Proportion of main namespace edit attempts that were abandoned with the user discarding changes (abandon). Proportion of main namespace edit attempts that were abandoned with no changes (nochange).✅ Yes✅ Yes
NOTE: All metrics will be limited to mobile web edits to the main namespace on English Wikipedia.

Event Timeline

ppelberg updated the task description. (Show Details)
Metric update

In T432293#12151365, we updated the definition of the denominator used to compute the "proportion" [DE1.1 KPI] Distinct registered users who make a constructive edit specifies.

This update means that this analysis, and the DE1.1 KPI, will be reporting on slightly different metrics.

@MNeisler and I are making this change for the following reasons:

  1. Tracking distinct mobile page views to edit attempts, as we decided to do in T389404, is difficult outside of an experiment.
  2. Further, enabling VisualEditor as the default only alters the experience of people who initiate an edit. Accordingly, a reader who never taps edit is not affected by this change and wouldn't make sense to include them in the denominator here.

This analysis can begin next sprint, seeing as how it will need to happen ≥2 weeks after the deployment happens, currently scheduled for Tuesday (28 July 2026) via T431858.