T239706 Suggested Edits Data Reporting

The Android team have experimented with a number of ways to alert users about the Suggested edits feature. We would like to assess the relative effectiveness of these methods so that we can prioritise our work for Q3.

01) Product decision: Should we get rid of Explore feed cards in Suggested edits V4?

Data requested:

  1. Count edits and users who edited via Suggested edits Explore feed card in total and per edit type (image captions / article descriptions)

  2. Put these numbers into perspective with total app edits, Suggested edits, users that edit with the app and Suggested edits users

Data from 14 Day Sample

Data used in this report was pulled from MobileWikiAppSuggestedEdits, and shows logged in user data from 2 separate 7 day periods (Group A:12-01-2019-12-07-2019 and Group B:12-20-2019 and 12-27-2019). This allows us to compare user interaction from before and after the "Suggested Edits" prompt was moved from the burger menu to the main tab bar (https://phabricator.wikimedia.org/T231442). For this report aggregate user data will be used.

group  source        
A      feed               440
       menu              1359
B      feed               535
       menu               121
       suggestedEdits    1208
Name: date, dtype: int64

User/action counts

Week 1 12/2019 & week 4 12/2019 Data from MobileWikiAppSuggestedEdits

In [31]:
totals
Out[31]:
source users td_edits ic_edits all_edits
0 feed 975 152 15 167
1 menu 1480 298 245 543
2 SuggestedEd 1280 263 152 415

CTR Rate and Conversion Rate

Compare users by source Menu/Suggested Edits vs Explore cards, using 14 day totals

Compare users who actually make edits (convert), by their source

Percent of all Edits by Source:

14.84% feed

85.16% menu/Suggested