There have been many times when I've gone to type up a new question, clicking the ask question button and start to fill out the form.

But almost half of those times I've successfully found an answer thanks to the awesome suggestions of Similar Questions. Sometimes I just find posts in those suggestions I can't find with Google.

Thinking on this premise, I'm curious to know if the Stack Exchange team is tracking such behaviour and has measurable results they'd like to share?

This is purely for curiosity.

  • 1
    How would you define an "almost question"? How many seconds into the question asking? How many characters?
    – Oded
    Commented Apr 11, 2012 at 18:49
  • Let's say I fill out a title and type my first sentence into the Question box Commented Apr 11, 2012 at 18:50
  • 10
    @Oded I think for DeLonge's purposes, typing out a title and getting the list of suggestions in the sidebar would count. I know I've done that more times than I've asked questions, myself. Commented Apr 11, 2012 at 18:51
  • My point is that without this kind of definition, collecting the info wouldn't be possible. And I doubt the SE team have defined such a thing.
    – Oded
    Commented Apr 11, 2012 at 18:52
  • 1
    Maybe some info could be gathered from the saved draft thing.
    – Mat
    Commented Apr 11, 2012 at 18:54
  • 1
    There might be some value here if it could be tied into question viewing analytics. I've typed up complete questions before, only to finally find a few would-be duplicates in the related question box that earlier searching didn't bring up.
    – Charles
    Commented Apr 11, 2012 at 19:57
  • 17
    I almost upvoted your question.
    – user102937
    Commented Apr 11, 2012 at 20:16
  • What would be the value? Commented Apr 11, 2012 at 20:35
  • 4
    IMO the value to StackExchange would be knowing the effectiveness of "similar question" results and seeing how questions many questions are disambiguated when not using a search engine. I'm sure this can be expanded upon. ;) Commented Apr 11, 2012 at 20:41
  • 3
    If you're curious about the effectiveness of the suggestion feature, then perhaps the following sequence could be tracked: (1) suggestion feature is activated (2) user clicks on one of the suggested links (3) user does not submit a question.
    – David Z
    Commented Apr 11, 2012 at 20:41
  • 1
    I never made it past that point. Zero questions until today. :)
    – fuxia
    Commented Apr 19, 2012 at 1:16
  • If you're entering titles are included; I use that like a search engine, I do it evem when I know the question I'm looking for is there Commented Sep 29, 2013 at 8:36
  • 3
    I use "ask question" + title to search rather than ask quite a bit. That search seems more effective than the regular search feature. Commented Sep 29, 2013 at 17:28

1 Answer 1


We didn't in the past. You could sort of deduce the presence of someone working on asking something but ultimately not posting the question from our load balancer logs and routes that serve the ajax requests going on, but we didn't actively track it.

Now that we're working on an improved interface to help very new users ask questions in better ways, this sort of thing becomes interesting. As we test the idea of breaking up the ask question page into multiple elements instead of one large editor, we will very likely be testing what makes people abandon questions (to the extent that this is possible).

Did they see a related question and opt to just go check that out instead? That's the sort of thing we're going to need to get some insight into.

I don't have full details of how we're going to do this because the implementation details for the 'wizard-y' interface still have to be ironed out, but we're definitely going to be getting numbers, some of which I'm sure we'll have no problem sharing.

  • I tell you what makes me abandon nearly complete (?) questions. When writing them I often realize more and more that they won't be complete unless I double check, debug and add more details (details I have to look up my self at first) or that I already answered them myself while improving them. But I also write those questions offline/off-site in a text editor. ;) I'm very interested what the "improved interface" will look like. Commented Dec 4, 2016 at 16:44
  • 1
    The mandatory link to the article about "complete overhauls" by Joel Spolsky... Oh, I forgot, you already have a live Joel Spolsky to keep an eye on that. Commented Dec 4, 2016 at 17:12
  • 2
    @ivan_pozdeev It's not a complete overhaul, it's an optional "guided" interface for very new users. Like training wheels that can come off later.
    – user50049
    Commented Dec 4, 2016 at 17:20

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .