6

Can you get better answers on Stack Overflow when a question is asked at a certain time of the day?

Let's say the question is asked in the morning (GMT), will it attract more attention than a question that is asked at night?

Suppose that is true: What is the best time to ask questions?

On the other hand,

  • How long can one expect to get relevant answers?
  • Is a question after (let's say) 12 hours "lost"?

The question seems to attract no more answers after a certain time, unless the question itself is that relevant that a lot of people keep it active.

  • time is somewhat relevant, but it is more day of week than time of day. Weekends see less traffic than weekdays, but since we have community members in almost every timezone, your post will get seen no matter what time you post. That being said, I think a majority of users are in the US, so posting between -4GMT and -8GMT will get you slightly more users – psubsee2003 Jul 9 '13 at 23:13
  • This is not my impression. And I have asked questions on every day of the week and morning, evening, night, lunchhour, ... but there are probably people out there that have asked more questions and have seen more on time-factor than I did. – Mare Infinitus Jul 9 '13 at 23:15
  • All I said was weekends see less traffic than weekdays, but when you are talking about millions of views per day, your question is still going to be seen. – psubsee2003 Jul 9 '13 at 23:18
  • Sure, there are always some views. But are there more views on a question asked in the morning than on questions asked on the night? When are the most people that give answers online? – Mare Infinitus Jul 9 '13 at 23:20
  • At peak times you'd get more views per second but your question will fall off the front page quicker so I'd expect the total views (and therefore answer number and quality) to remain the same – Richard Tingle Jul 10 '13 at 6:26
12

How long can one expect to get relevant answers? Is a question after (lets say) 12 hours "lost"?

I got an answer on this after 4 months.

Usually, 24 hours is good, though. Simple questions get answered in a matter of minutes (sometimes a matter of seconds). It all depends on the question. Some questions are quite tough and need a bounty to get going.

Can you get better answers on stackoverflow when a question is asked at a certain time of the day?

Let's say the question is asked in the morning (GMT), will it attract more attention than a question that is asked in the night?

Disclaimer: These graphs work on simple SQL queries. They do not, at any point in time, indicate the quality of answers you may receive.

I ran a query, this is what I got:

enter image description here

Looks like the high activity is around 12:00UTC.

However, it seems that the number of answers per question seems to fluctuate randomly.

enter image description here

The views per question isn't terribly interesting either:

enter image description here

Neither is the time it takes (in seconds) to get the first answer

enter image description here

  • Nice, I wasn't sure how easy it would be to query / graph those. Important to note, though, that number of answers and quality of those answers are two very different metrics. :-) – Aaron Bertrand Jul 9 '13 at 23:53
  • @AaronBertrand Of course. #include<aaronsanswer> and all that. I didn't feel like repeating all you said, decided to serve up some cold hard data :P – Manishearth Jul 9 '13 at 23:56
  • Oh no, that's not what I meant, I just wanted to point out that the graph showing answers per question - even if it did show some kind of pattern - does not address the "better" part of the OP's question, just the "more" part. :-) – Aaron Bertrand Jul 9 '13 at 23:58
  • @AaronBertrand Added a few more graphs, and a disclaimer. Also, you're a DBA mod. You should know how to SQL (more than me at any rate). Shame on you :P – Manishearth Jul 10 '13 at 0:38
  • Touché. To be fair, it's not because I couldn't figure out the queries, just that I'm unfamiliar with the schema and what data is exposed. :-) – Aaron Bertrand Jul 10 '13 at 0:44
  • @AaronBertrand Ah. The schema is actually presented on the side (not too hard to grok) when you edit or write a query. Most public data is exposed, excluding full revision information, and review queue data (except for suggested edits) – Manishearth Jul 10 '13 at 0:47
  • This question took about 6 months! I offered a bounty on it and then it got an answer. – Cole Johnson Jul 10 '13 at 1:01
  • Well wasn't really talking about grok either, but rather complete unfamiliarity. I know of the site and I've looked a couple of times, but never really used it. – Aaron Bertrand Jul 10 '13 at 1:05
  • @AaronBertrand I know, I know, no need to defend yourself :) Just saying that it's not too hard to learn and extremely useful once you do. The data I pulled out of it for today's Physics.SE close reason discussion was really helpful. – Manishearth Jul 10 '13 at 1:16
  • +1 - Sweet graphs! That is cool data to look at. – Travis J Jul 10 '13 at 2:25
  • Very interesting answer! Thank you very much! – Mare Infinitus Jul 10 '13 at 5:42
7

Stack Exchange is a global site. I'm sure that while a majority of the user base is in the Western Hemisphere time zone, I'm also sure that there is a wide distribution of when answerers are online and active - on their lunch break, after dinner, after the 11 PM news, etc. You also need to be careful about assuming that a good answer will posted at or around the time the question was posted. I often mull a question for a while before posting and if you think about that logically that should be true as well. Also when I am most active is not always the time that I have available to formulate answers. I often read a bunch of questions and note the ones I want to think about and come back and ask later.

I don't know that there's any way to draw anyManishEarth drew some relevant statistics from data.se, but "better" is quite subjective. Just because an answer to a question posted at 1100 UTC got 15 up-votes while an answer to a similar question posted on a different day at 1800 UTC only got 5 up-votes does not mean that the former was a "better" answer. And more answers does not mean better answers. Now, I don't know if they keep statistics on active volume / sessions vs. individual questions and answers - I'm sure they have some of that information somewhere, I just don't know if it is exposed to us.

If you tend to participate in a few tags frequently, you may have noticed users that often produce high quality answers. So one thing you may wish to do (on a much smaller scale than your question implies) is to observe when they post most of their answers. Then you at least have a slightly higher chance at getting answers from people you know give good answers. Keep in mind though that their answers posted at 1800 UTC might be on a question they first read at 1100 UTC.

Possibly more relevant

One thing I have observed as having an affect on the quality of answers is the quality of the question. If you ask a poor question, you get poor answers. Time of day is irrelevant. And regardless of the quality of the question, on SO at least, you get no shortage of answers in either case. More on that in a minute.

There is also the "rush" syndrome. People seem to be obsessed with posting the first answer (for badges like Enlighted, aka FGITW), or as early as possible (for a better chance at a share of the up-votes).

One other issue is the badgering I often see for the OP to accept someone's answer. We took away the accept rate in part to alleviate this badgering (which, admittedly, did not always). But still this can occasionally pressure OPs to accept answers before a large majority of the potentially interested population get to see them.

Now, the problem with all three of these behaviors as I see it is that much less attention is paid to questions that have answers. And this is particularly true for questions that have up-voted and, most importantly, accepted answers. But the folks most likely to spend time and effort crafting a good answer are likely to gravitate to questions where they are likely to have the greatest impact: those with few / no answers and certainly without an accepted answer. The theory about that is that if an answer has been accepted, the OP already has a solution and probably doesn't care much about new solutions, and if there are multiple answers, the competition might be harder.

On a site with millions of users, all questions do get additional views in spite of these facts. But I suspect that a lot of the factors that influence whether you get the best possible answer have less to do with time of day and more to do with dumb luck / coincidence.

  • As seen in Manishearth's answer, there are high activity times. But you are right, the question itself has probably the most impact on the quality of answers. – Mare Infinitus Jul 10 '13 at 5:44
5

Time for questions is not an important thing to care for because there is people for all around the world here and every country have a different time from the other and some people sleep on the day and some on the night so there will be people answering and asking questions like 24 a day so time is not a big deal

the big deal is how the answer hardness goes on if it not hard you will get an answer in a no time if its hard it either need a discussion on the chat or along explanation in a long answer that will be edited several of times

You must log in to answer this question.

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