This task involves the work of running an initial iteration of a survey to help us understand the reasons that might explain why people decide to abandon the mobile editing sessions (v1) they initiate.

The goal of this survey is to learn, as noted in T408486, what reasons potentially explain why people who [seemingly] tap out of curiosity end up abandoning before taking any action within the interface.

Decision(s) to be made

• What (if any) improvements will the Editing Team prioritize work on as part of WE1.7 to increase the proportion of people who curiously tap Edit on mobile and go on to publish a constructive change?

Requirements

See: Hypothesis: WE1.7.6. "Exit the editor" survey.

DEPRECATED Quick Survey: Newcomer "Exit the Editor" Survey.
Meta

• Start date: Thursday, April 23, 2026
• End date: TBD

Target audience

• Wikis: en.wiki
• Platform(s): Mobile web
• Max edit count: ≤100 cumulative edits
• Editing interface(s): mobile VE & mobile source
• Account state: logged in and logged out

User experience

When someone who is logged out or logged in [and has published ≤100 cumulative edits] abandons an edit session on mobile web after spending ≥2 seconds in either the mobile visual or source ready state, present them with a quick survey (see specifications below).

Survey metadata

• We will want to group survey respondents by:
  • Whether they abandoning their edit with or without making a change to the document
  • Experience level (read: cumulative edit count bucket)
  • Editing interface (source or visual)
  • Wiki
  • Edit session initiation method (section edit, full-page edit, etc.)

Survey

• Question(s)
  • .
  • .
  • .
  • .
• Response(s)
  • .
  • .
  • .
  • .

Open questions

• 1. How (if at all) does it matter if the same person is surveyed twice?
• 2. What kinds of conclusions are you needing to be able to draw from this survey? Knowing this will inform where we need to strive for a high degree of precisions. E.g. are we looking to run the survey twice (before and after Suggestion Mode) and by extension, needing those two runs to be as similar as possible so as not to introduce confounding variables.
• 3. Will this survey display once per user? Will it display each time someone abandons the editing interface?
  • To start, let's assume this survey will be shown once per user, regardless of how many sessions meet the requirements specified above.