Page MenuHomePhabricator

[Request] Research into suspicious account behaviors
Closed, DeclinedPublic

Description

Please provide all the following information:

  • Context. The cuci_user shows us which accounts have a CheckUser-logged action for them each month (up to 90 days of history). Our public edit data shows us a count of users that edit each month. Our account creation logs show us how many accounts are created each month. The goal of this research would be to look into what is going on with accounts that are active but otherwise don't edit, and to try to understand what a likely-good-faith number of account signups is each month, in comparison to our likely-good-faith number of active accounts. In particular, we should be able to gain a better understanding of how much scripted account activity is occurring on our platform, if we look more closely.

This research could also look into login flows (for example, the gap between the number of MediaWiki-extensions-EmailAuth verifications requested each month and how many are completed).

  • Description. Overall, getting a better handle on good faith and bad actor activity among registered users on the projects.

I've not shared specific numbers or examples here due to sensitivity of the information, but can provide that if/when this project is scheduled. Internal examples:

  • Expected Deliverable. TBD
  • Estimated Effort. Not clear, but probably at least a month of digging and validating results with various stakeholders?
  • Priority This work would support both WE4.2 and WE4.6 and is a priority for us in Q1.

I need this task resolved in:

  • 1 month.
  • 3 months.
  • 6 months.
  • Whenever you get to it :-)
  • Other. Do you have any other questions or comments ?

This also relates to efforts to make sense of fluctuations in our month-to-month account registration funnel.


For use by WMF Research team; please leave everything below as it is:

  1. Does the request serve one of the existing Research team's audiences? If yes, choose the primary audience. (1 of 4)
  2. What is the type of work requested?
  3. What is the impact of responding to this request?
    • Support a technology or policy need of one or more WM projects
    • Advance the understanding of the WM projects.
    • Something else. If you choose this option, please explain briefly the impact below.

Event Timeline

It would good to see if there are precise criteria that we could identify and use to separate accounts into tiers of how likely/known they are to be good-faith accounts. Ideally we would get enough confidence to handle those accounts programmatically differently during authentication, recovery, moderation, and other functions in the wikis.

Based on an internal conversation this task is for Movement Insights and Product Analytics to make a prioritization decision about. I will remove Research's tag and add Product Analytics as a result.

kzimmerman subscribed.

Assigning this to Morten for Q1

Dreamy_Jazz subscribed.

@nettrom_WMF has since left the WMF, so I think we can probably call this declined