Created attachment 16605
new thank process

Attached:

The 2 step process was requested by a very vocal group of editors that kept thanking people they were trying to revert.

That doesn't excuse us from needing to provide validation that the current implementation is needed or correct.

(In reply to Jared Zimmerman (WMF) from comment #0)

If we're seeing a lot of abandons on the second step we need to re-evaluate
the two step process.

Why? To me that would mean that the confirmation is working as designed.

As far as I'm concerned this is a band-aid until we can get to a proper 1-step process with an optional undo. Unless we can understand how users are using the feature now, we can't validate that we're going down the right path. I'm happy to have other suggestions as to how we'd instrument this such that we'd be getting data that you feel would be more useful.

It's been long enough in the making now that I can tell you that I only send 1 thank per 20-25 instances where the thought has crossed my mind because the two step process isn't user friendly. I'm certain that the usage of this moral and community building feature would be much, much higher if it wasn't so clunky and frustrating to use.

Fix the feature correctly so it is one click to thank and if it was an accident, then offer an undo. This really can't be that hard to do. Simply put the thank in a buffer that sends it after 15-30 seconds or on page unload.

(In reply to Technical 13 from comment #6)

Fix the feature correctly so it is one click to thank and if it was an
accident, then offer an undo. This really can't be that hard to do. Simply
put the thank in a buffer that sends it after 15-30 seconds or on page
unload.

bug 69636 welcomes patches!

I'm adding a seealso for bug 69804 which also asks for analytics.

However, I don't understand how analytics will be helpful.
We already know that The Confirmation Step is a bandaid, and is currently needed because the "Thank" link appears directly next to (undo|rollback) - and it's immensely embarrassing/aggravating to Thank a vandal that we meant to Revert - so we can't remove that bandaid until an alternate solution is available.

Unless there is a rationale for adding analytics beyond just "confirm what we already know", then I would suggest that any programming efforts would best be put towards creating the Thank-Undo code (bug 69636) instead of this.