Full dislosure: It was an identify question of my own that initially turned me onto this site.


  • Music Fans SE should engage in a time-limited trial run of restricting or eliminating "identify" questions.

There are three sections to this post: (1) Two proposals; (2) Justification; (3) Stats.

Two proposals regarding "[identify*]" questions

Proposal #1:

Treat "identify" questions as off-topic for a one-year trial period to observe the effect on overall participation.


Proposal #2

Restrict, for a one-year trial period the "identify" questions that are on-topic. For example:

  1. Require that any posted video be on a publicly accessible site (e.g., not a personal cloud-share site; not a site requiring fees). (Also in the interest of security. I have no interest in downloading unknown files from random sites.)
  2. Require that the above site track views, and that the submitted video have a minimum such number (e.g., 10,000). This will at least suggest some future usefulness.


"Identify" questions have been a point of discussion, contention, and, for some, irritation since the inception of Music Fans. For example:

As shown in the stats below, "identify" questions are

  1. Pushing out other questions;
  2. Producing less engagement and less impact;
  3. Of lower quality.

One concern has been that limiting or removing "identify" questions will reduce site growth. However, as measured by number of questions asked, growth is stagnant even with them (see Stats, below). Evidence suggests "identify" questions are actually suppressing growth.

As a personal matter, I find "identify"-type questions to be a drag most of the time. They often take a substantial amount of research to answer and are generally of limited long-term or popular value. For me, the "identify" questions have become a significant disincentive and have substantially reduced my participation.

I advocate strongly for a trial run of either restricting or removing them.


  1. In Music Fans's first year (Feb - Dec 2015) the ratio of "other" questions to "identify" questions was better than 4:1. However, since 2016, the number of "identify" questions has been holding steady at roughly double, while the number of "other" questions has dropped by more than 50%.
TABLE 1: Questions by year

Year   Is  (%)     Not (%)     Total
2021*  217 (62%)   135 (38%)   352
2020   189 (48%)   203 (52%)   392
2019   236 (59%)   165 (41%)   401
2018   206 (53%)   183 (47%)   389
2017   235 (50%)   231 (50%)   466
2016   231 (44%)   294 (56%)   525
2015** 108 (19%)   472 (81%)   580
*as of Nov. 30, 2021
**first site question asked 2/2015
  1. "Identify" questions are significantly less engaging and less impactful than other questions.
TABLE 2: Question impact/popularity*
Q-type    Score  Views  Favorite  >1 answers/total Q 
identify  1       715   0         263/1445 = 18.2%
other     4      2364   1         555/1703 = 32.6%
*Based on data from SEDE. Queries are below.
  1. "Identify" questions are of lower quality than other questions.
TABLE 3: Percentages of closed questions versus total

Tag        Closed/Total  %
Identify*  50/1463       3.4%
Other      93/4292       2.7%

Queries used for Table 2.

SELECT AVG(ViewCount) FROM Posts WHERE Tags LIKE '%identify%'  
SELECT AVG(ViewCount) FROM Posts WHERE Tags NOT LIKE '%identify%'  

SELECT AVG(Score) FROM Posts WHERE Tags LIKE '%identify%'
SELECT AVG(Score) FROM Posts WHERE Tags NOT LIKE '%identify%'

SELECT AVG(FavoriteCount) FROM Posts WHERE Tags LIKE '%identify%'
SELECT AVG(FavoriteCount) FROM Posts WHERE Tags NOT LIKE '%identify%'

SELECT COUNT(AnswerCount) FROM Posts WHERE Tags LIKE '%identify%' AND AnswerCount > 1
SELECT COUNT(AnswerCount) FROM Posts WHERE Tags NOT LIKE '%identify%' AND AnswerCount > 1
  • 3
    I'll see if I can link other questions important to the track of identification questions on this site over the next few days. They were supposed to be trial only and it never really seemed successful in my opinion, but was near impossible to roll back once it started. I'll also point out another problem with identification questions on this site in your question which is the one answer you got is incorrect, but still upvoted. The questions are set up so only the OP can validate the question, but anyone can upvote the answer making it have issues on a Q&A site.
    – Dom Mod
    Oct 24 '20 at 5:41
  • 1
    This question is tangentially related: musicfans.meta.stackexchange.com/questions/421/… . This question was very important for getting ID questions from off topic to on topic: musicfans.meta.stackexchange.com/questions/203/… .
    – Dom Mod
    Oct 25 '20 at 19:10
  • 1
    About hosting services: musicfans.meta.stackexchange.com/questions/328/…
    – Bebs
    Oct 26 '20 at 7:56
  • 4
    Much of your argument for making such questions off-topic here is based on the premise that even the mere presence of such questions on this site harms it. Any user who feels this could just use SE's "Ignore tag" feature on this or any other tag.
    – Rosie F
    Nov 17 '20 at 12:49
  • 1
    @RosieF That's a fair read, but I'm also saying that they aren't helping -- usage is going down even with them, both in terms of total user engagement and per-question engagement. So I propose a time-limited trial run without them to see if that has a positive effect on overall site usage.
    – Aaron
    Nov 17 '20 at 16:23
  • 1
    @Aaron My apologies, yes. It can’t be abandoned altogether. I did see someone bitterly regret it’s removal on the Movies Stack Exchange site. Many said “It was the reason I first engaged with the site and continue to engage with it”. A large audience will be missed and dismissed, not to mention the considerable amount [closed] dead questions that will be inevitable.
    – cmp
    Dec 23 '20 at 0:40
  • 3
    The reasoning behind ID questions being removed from M&TV was the sheer number of lazy questions. That some well-researched questions may have been lost, as the baby with the bathwater, I think has seen the overall question quality on there go up. I'm still very much engaged there, whereas here I poke my head in once a month if I'm bored. The ID questions imo are generally zero value. Most of the current front page is either 'name this tune' or 'what genre is this?' both of which are equally pointless.
    – Tetsujin
    Dec 24 '20 at 12:39
  • 1
    Is there any way to tell how many ID question askers are likely to return & fully engage with the site longer term, vs fly-by askers?
    – Tetsujin
    Dec 24 '20 at 12:41
  • 6
    The idea of stack exchange as a whole is not to provide value to a single user, but to become a permanent searchable resource for many users. If a question is unsearchable, ie "what's this song? link provided." then it is of literally no value to anyone at all except the asker. Many SE sites have that as a close reason.
    – Tetsujin
    Dec 24 '20 at 13:13
  • 1
    @Tetsujin If we are prepared to see this particular network even less popular than it is already, we are going the right way about it. I have to say it’s very ironic that the person writing this post was first initially engaged to this site by an identify question. Anyway, Reddit exists and will always dominate this platform.
    – cmp
    Dec 24 '20 at 13:39
  • 3
    Stack Exchange is not here to emulate Reddit, or any forum-style site. It is a searchable resource for many users. I see no irony in the question - though I do in the answer to the OP's initial post, which though not the right answer was the only one… & it was mine ;) The question itself was at least a good attempt to provide full information & may possibly be googlable. It didn't just say "what's the song in this video?" which has absolutely zero value as a searchable resource.
    – Tetsujin
    Dec 24 '20 at 13:45
  • 1
    @cmp It sounds like you might be open to setting some tighter criteria for what qualifies as a "good" identify question. I'd be interested in your thoughts on how the criteria could be narrowed to better catch the valuable questions and better filter the "lazy" ones or the ones meant to drive traffic to a YouTube channel, etc.
    – Aaron
    Dec 24 '20 at 16:22
  • 2
    @Clockwork You've hit on what I see as a major downside of "identify" questions. Unlike a book summary, the questions here rarely contain (specific) enough information, and they tend to rely on links that are apt to go bad eventually. Clearly there are enough people who post/read/answer these questions that they've been kept on the site, but for me personally it's a major disinsentive. I mostly just visit now to help manage the review queues and not much else.
    – Aaron
    Apr 4 at 19:46
  • 1
    "help me identify" are the worst questions. They are super low quality, other stack exchange sites BSN then for this reason. They offer no value to anyone else other than the original poster and the suck all of the oxygen out of the site for other questions. Kill them, kill them now. Jun 29 at 7:58
  • 3
    Low value, poor quality growth is worse than slow growth. Jun 29 at 7:59

My reflex is to say yes, goodness yes, these questions tend to be such poor quality they is just noise.

However, on reflection, I think there should be some allowed. I think the criteria should be you must include one of the following (or similar)...

  • an actual audio source for the piece of music (for example a youtube video) including a time code (unless it is throughout)
  • the movie or tv show which you heard it in with timecode or just it is the theme song, or the closing credits etc.
  • an actual image of the sheet music, but not something the OP has produced, an actual source

So this makes the following unacceptable...

  • Half remembered lyrics (something about a lover who leaves early in the morning?)
  • vague thematic descriptions (a syth song with dark themes about nuclear war?)
  • descriptions of music videos (children wearing dolls masks?)
  • OP produced recordings of melody or chords
  • OP produced partial score, chord progressions, or musical description (a descending bass line with a reggae guitar and a woman singing)

FWIW, I have been active on Stack Overflow for years (11k rep) but was never aware of Music Fans. However, I got a song from the 80s stuck in my head but can't remember the band. After fruitless Googling (I remember some lyrics) I landed at Music Fans. I was just about to post a question about the song when I saw this meta post. Now, I'm not sure if I should.

My point is that since the site is still in beta, is it wise to ban a potentially large user base? Maybe you should wait and revisit this issue when the site is no longer in beta.

You may be wondering: if I do ask my question, what will my level of participation be after that? It's hard to say. But at least your numbers will have increased slightly.

Also, it has been mentioned that the Movies Stack Exchange has banned these types of questions. But over at Science Fiction and Fantasy, there are over 18k questions tagged "story-identification" - so it seems to be very successfully handled there.

  • 1
    To your question, I'd say go ahead and post. They are currently on topic, and there are enthusiasts who enjoy these kinds of questions. To the larger issue, I just want to reiterate that my OP doesn't suggest an outright ban. My proposals are 1) a time-limited trial run of a ban and 2) restricting the criteria for these types of questions to be considered "on-topic".
    – Aaron
    Aug 26 at 15:12

You must log in to answer this question.

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