What is meta? ×
Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 127 Stack Exchange communities.

I crunched some numbers I got from SE's site directory on daily visits:

                Sites  Daily Visits (%)
Stack Overflow      1        2.8m (82%)
Launched Sites     31        595k (17%)
Public Betas       42       22.4k (>1%)

Perhaps this chart will help you visualize the information above:

Pie chart

As you can see, SO easily beats the entire SE network in daily visits. All of its 73 sites.

Is it possible that this is happening because this Q&A format was designed for programming questions and is not that well suited to other topics? Or perhaps it's a Q&A format that makes a lot more sense to programmers than other people?

I'll admit that when I first used SO, it just made too much sense to me. My sister is interested in many topics covered by the newer SE sites, but for whatever reason she doesn't find this system appealing. It also didn't appeal to other non-programmer friends I introduced SE to.

Considering many users on the newer SE sites come from SO (and are thus programmers), this might further mask a deficiency in appealing to non-programmers. I realize I'm asking this to an audience mostly of programmers, but is it possible that our Q&A format is not that well suited to a more general audience?

EDIT: Yes, it's not entirely fair to compare SO to the rest of the network because it's much older.

But SO had 231,070 questions one year after its launch (2008-07-31), and 282,028 questions one year after the public beta started (2008-09-15). Many other SE sites have been around for much longer, and have nowhere near this many questions.

The 2nd and 3rd most popular sites in the network, ServerFault (2y 8mo.) and SuperUser (2y 5mo.) have 201,606 questions between them right now. This is less than what SO had in one year.

enter image description here
(from the launch of the private beta)

I crunched more numbers. Here's a spreadsheet with the number of questions after one year into a public beta.

Stack Overflow had 282,028 questions one year after public beta. 22 other sites combined to 195,031 questions after one year. Here's another delicious pie:

enter image description here
(12 sites omitted from legend)

share|improve this question
It is probably because programmers spend more time on the PC and in the internet than most other people. –  Marcelo Dec 1 '11 at 20:27
Given our bias, this would probably be better asked to the communities on the Cooking, Bicycles, or Home Improvement meta sites. (Of course I know that this is the right place to ask such questions, I'm just saying it would be nice to get non-programmer input.) –  Bill the Lizard Dec 1 '11 at 20:29
@Marcelo There are people who spend their whole day on Facebook (and most programmers I know, myself included, hate Facebook) –  NullUserException อ_อ Dec 1 '11 at 20:34
I totally forgot about Facebook and its' games. If that is the case, we should do a Stackexchange Facebook game. Success! –  Marcelo Dec 1 '11 at 20:37
Whatever, those other sites are primarily there to make it possible to keep crap off SO. ;) –  tvanfosson Dec 1 '11 at 21:01
I've read this twice and I'm not sure what the purpose is. Are you saying we should change the format on SE2.0 sites? –  Michael Mrozek Dec 1 '11 at 21:25
@MichaelMrozek Yes. It's clearly not working nearly as well as it did for SO. –  NullUserException อ_อ Dec 1 '11 at 21:26
@NullUserExceptionอ_อ I'm not convinced that "SE 2.0 sites are smaller = Q&A format is not working". –  Anna Lear Dec 1 '11 at 22:38
@NullUserExceptionอ_อ you need a chart of the questions per user ratio. That might give you better information. –  amanaP lanaC A nalP A naM A Dec 1 '11 at 22:46
+1 for lots of cool charts and graphs. :D –  Andrew's a Unitato Dec 2 '11 at 1:32
@NullUserExceptionอ_อ - You're missing data for Electrical Engineering - Public beta 2010-09-29, 12410 questions after 12 months (putting us just above Programmers) - Admittedly, we were seeded with some 800 questions and a couple hundred users from an SE 1.0 site. –  Kevin Vermeer Dec 2 '11 at 14:10
add comment

2 Answers

StackOverflow had a huge headstart on all the other sites, not only in time but also in the starting audience that Jeff & Joel brought from their blogs. I don't think you can gain any easy conclusions just from traffic data, the age factor will lead to a significant distortion. The other sites just haven't had enought time to even have a chance to close in on StackOverflow.

If you want to do a better comparison, I'd start with comparing the relative traffic growth. See if the SE 2.0 sites are growing at a rate similar to SO. I still think that this comparison would tell you more about the initial community and the quality and demand for the content they produced, and not necessarily much about the general applicability of the Q&A system to the specific topic.

There are certainly topics where the Q&A system works better, one factor that differentiates SO in my opinion is the easy verification of answers on SO. In most cases I can copy the code from the SO answer and check if it actually solves the problem. On many other sites it is far more difficult to determine if an answer is good.

share|improve this answer
I agree. This entire analysis feels like it's starting from a flawed premise. –  Anna Lear Dec 1 '11 at 20:47
@AnnaLear It's not that flawed. SuperUser and ServerFault have been around for over two years now, but they have less questions combined than SO had in one year. –  NullUserException อ_อ Dec 1 '11 at 20:57
OK @AnnaLear, I posted two edits with more reasonable comparisons. –  NullUserException อ_อ Dec 1 '11 at 22:20
@NullUserExceptionอ_อ Not that flawed is still flawed. :) I do not see how "other sites are growing slower" is anywhere near "this Q&A format is not working". SE is building sites of high quality even if they're small at this point. We have the CHAOS team to help promote SE sites. I'm active on several SE sites, including some non-programming ones and I think the model works quite well. When it doesn't, the sites are closed or don't make it to beta in the first place. –  Anna Lear Dec 2 '11 at 2:30
add comment

It's not the format... it's the site itself.

The reason StackOverflow was initially so popular was that it was innovative - there simply wasn't a way to get good programming help for free that covered such a wide variety of topics. There was a hole in the internet, and StackOverflow filled it. With the other sites, this isn't so true.

I suspect most sys admins are used to following newsgroups to get their questions answered, so they don't need to go to Server Fault. There's more "computer help" sites than you can shake a stick at, so there's a smaller pie left for Super User. Many of the other StackExchange sites are experiments too; they cover ground already covered by other sites, but are approaching it in a different way.

So it's not that Stack Exchange's format is well suited to programmers. It's that programmers needed a site with a format that was suited to them. The format is just as well suited to any other discipline, but those disciplines simply didn't need one like programmers did.

share|improve this answer
There are probably more programming help forums than I can shake a stick at too. –  NullUserException อ_อ Dec 2 '11 at 0:17
@NullUserExceptionอ_อ There are many programming forums, yes. But they don't cover such a wide variety of topics individually. If you like Java and C# you probably have to be a member of two forums to do it, without Stack Overflow. –  corsiKa Dec 2 '11 at 0:19
Well, there was a certain hyphenated site that had the position SO has now, but SO still managed to eventually surpassed them fairly quick. –  NullUserException อ_อ Dec 2 '11 at 0:22
That's exactly the point though. The best option before wasn't a good option at all. I would imagine most other disciplines can't say the same. –  corsiKa Dec 2 '11 at 0:51
add comment

You must log in to answer this question.

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