What is meta? ×
Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 134 Stack Exchange communities.

Hi ! It seems to me that there are fewer people talking online on stackoverflow.com as sooner.

The number of views is decreasing, the number of answers logically too etc.

Is it just a seeming or reality ?

EDITED:

Alright look at this: http://www.alexa.com/siteinfo/stackoverflow.com

Look at the Graph 'Daily Reach' with max settings

share|improve this question
    
caught in a landslide –  random Feb 18 '10 at 7:44
3  
Your question needs some serious parsing. –  random Feb 18 '10 at 7:48
1  
@random no escape from reality –  alex Feb 18 '10 at 7:54
    
So the feature request is for?? More users?? –  o.k.w Feb 18 '10 at 8:08
    
No, look at this: youtube.com/watch?v=tgbNymZ7vqY with max settings. @ale –  random Feb 18 '10 at 8:10
    
@random you could scar someone for life with that... –  alex Feb 18 '10 at 8:24
    
Where's the close reason for "Huh..?" –  Ether Feb 18 '10 at 18:27

2 Answers 2

up vote 3 down vote accepted

I don't think there's a problem. From the figures actually on the page:

            Reach  Change
Yesterday   0.291   +81%    Change in Reach over the trailing 1d period
7 day       0.203   -13%    Change in Reach over the trailing 1w period
1 month     0.224   -7%     Change in Reach over the trailing 1m period
3 month     0.2421  +50.9%  Change in Reach over the trailing 3m period

I'm suspect of the "Yesterday" figure as it's early morning now (UK time) so there's little "today" traffic to compare against.

The weekly and monthly figures are down, but again the former could be skewed by the time of day.

However, the three month figures show an increase, and surely it's this figure that's the important one.

More importantly, given that the graph is measuring the % of internet users who use Stack Overflow this number can't be expected to keep on rising. There will be minor (and perhaps not so minor) fluctuations in the reach on a weekly or monthly basis, so these figures should be treated with some scepticism.

share|improve this answer

Let P(A) be the probability that a given internet user has the Alexa toolbar installed, and let P(S) be the probability that a given internet user is part of Stack Overflow's core audience.

Here in Germany a poll is usually considered statistically valid if about 1,000 out of the 80 Million people have been asked. So let's assume 0.001 % (or one in 100,000) of people have to be asked for statistic data to actually have a meaning.

Alexa states that

it is difficult to accurately determine the ranking of sites with fewer than 1,000 monthly visitors

In other words (we're of course simplifying a lot), Alexa considers their statistics for a particular site to be reasonably accurate if 1,000 people per month have been "asked". Combining this with the above number means that the number of internet users Alexa knows about is

N = 1,000 * 100,000 = 100,000,000

According to the ITU, 22% of all people were internet users in 2007 (the latest data they seem to have). Let's assume 25% out of 6 billion, which is 1.5 billion.

So we can deduce that one in 15 internet users has the Alexa toolbar installed, hence P(A) = 1/15.

I seem to recall (but correct me if I'm wrong) that on some Podcast, Joel and Jeff mentioned that they estimate SO's worldwide potential audience size somewhere between 4 and 8 millions. Let's assume 6. So 6m out of the 1.5b internet users are (potential) Stack Overflow audience, P(S) = 6 / 1500 = 0.4%

The Alexa page you refer to gives SO's reach as about 0.2%, in other words, the probability that a user that has the Alexa toolbar installed is a SO visitor is P(A|S) = 0.2%.

According to Bayes' theorem,

         P(A|S) P(S)   0.2% * 0.4%
P(S|A) = ----------- = ----------- = 0.012%
            P(A)           1/15

So, 0.012% of the 6 millions SO audience, or 720 people, have the Alexa toolbar installed. Because of the technical background of the people here, who are a lot less probable to install a spyware addon into their browser, the actual number is probably less.

According to a reliable expert, 14% of people know that people can come up with statistics to prove anything. Thus, beyond any reasonable doubt, I have proven that the Earth is a disc. Now what was your question again?

share|improve this answer
3  
.... --- .-- .. ... -... .- -... -... -.-- ..-. --- .-. -- . -.. ..--.. –  random Feb 18 '10 at 10:16
4  
@random: [... - .- - ..- ... -....- -.-. --- -. -.-. . .. ...- . -..] –  balpha Feb 18 '10 at 10:25
    
Your explanation is really amazing ! I understand what's your point. All those statistics are senseless. –  PaN1C_Showt1Me Mar 22 '10 at 7:31

You must log in to answer this question.

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