# Site statistics

This meta post will track our stats on Area 51, fortnightly (14 days) or a bit longer if I forget and someone else updates it:

Here's the Quantcast estimate of site traffic, with the mean number of visitors per month, a sparkline of weekly visitors for the last six months, and highest and lowest weekly visitor count in the last six months: (the chart doesn't always show, so just click through on that link if not, to see the detailed stats)

You can help build the site, by being active on meta and in chat; by asking questions, answering questions, and by sharing links to particular questions and answers across social media: use the "share" link on posts to generate a URL that can earn you bronze, silver and gold badges.

Date      Days in  Qs/   %     200+  2k+   3k+    Total  Ans   Visits  Qntcst
beta   day answrd  Usrs  Usrs  Usrs   Users  Ratio  /day    p/mth
-----------------------------------------------------------------------------
2016-07-03   810   3.6   89%    230    28   20   3 899   1.6   1 629   42 630
2016-05-18   763   1.6   90%    226    28   19   3 679   1.6   1 813   46 012
2016-05-02   747   2.7   90%    226    28   19   3 614   1.6   1 930   44 210
2016-04-18   733   1.9   90%    225    28   19   3 533   1.6   1 793   38 729
2016-04-04   719   2.1   90%    222    28   18   3 452   1.6   1 404   37 517
2016-03-19   703   2.4   89%    218    28   17   3 387   1.6   1 688   42 487
2016-03-04   688   2.5   90%    213    28   17   3 317   1.6   1 834   45 275
2016-02-19   674   1.9   90%    211    28   17   3 246   1.6   1 662   49 579
2016-02-04   659   3.7   90%    208    28   17   3 163   1.6   1 928   52 334
2016-01-21   645   2.9   91%    203    28   17   3 014   1.6   1 696   39 346
2016-01-05   629   1.9   91%    196    27   16   2 884   1.5     982   31 980
2015-12-05   598   2.0   91%    195    26   16   2 746   1.6   1 383   43 738
2015-11-21   584   2.2   91%    192    26   16   2 670   1.6   1 866   49 956
2015-11-05   569   1.8   91%    189    25   16   2 579   1.5   1 929   47 510
2015-10-20   553   2.7   90%    189    25   16   2 497   1.5   1 815   42 379
2015-10-06   538   2.0   91%    186    25   16   2 424   1.5   1 617   40 791
2015-09-21   523   2.4   90%    185    25   16   2 370   1.5   1 654   39 243
2015-09-07   509   2.4   90%    184    24   16   2 308   1.5   1 446   35 847
2015-08-24   495   2.0   91%    183    24   16   2 238   1.5   1 418   34 100
2015-08-06   477   2.4   91%    178    24   16   2 168   1.5   1 370   39 705
2015-07-21   461   2.0   91%    173    24   13   2 091   1.5   1 501   47 040
2015-07-07   447   2.8   91%    162    23   13   2 019   1.5   1 516   42 934
2015-06-23   433   1.6   91%    160    23   13   1 947   1.5   1 403   38 155
2015-06-08   418   1.6   91%    159    23   12   1 897   1.5   1 622   38 439
2015-05-25   404   1.6   91%    158    23   12   1 840   1.5   1 422   38 126
2015-05-11   390   2.1   91%    155    23   12   1 806   1.5   1 420   36 803
2015-04-27   376   2.0   91%    153    22   12   1 761   1.5   1 256   31 936
2015-04-11   360   2.0   91%    151    21   11   1 703   1.5   1 054   27 445
2015-03-27   345   2.4   91%    150    22   10   1 658   1.5   1 083   26 421
2015-03-09   327   1.6   91%    145    22    9   1 582   1.5   1 058   26 495
2015-02-23   313   1.4   91%    141    21    8   1 539   1.5   1 030   27 690
2015-02-09   299   1.9   91%    138    21    8   1 488   1.5   1 023   27 035
2015-01-26   285   1.7   91%    137    21    8   1 438   1.5     898   21 421
2015-01-10   270   2.6   91%    133    20    6   1 367   1.5     674   20 031
2014-12-10   238   2.4   89%    117    16    5   1 242   1.5     758   24 845
2014-11-25   223   3.2   88%    112    12    4   1 178   1.5     833   20 390
2014-11-12   212   2.1   86%    102     8    2   1 055   1.5     496   11 737
2014-10-30   197   1.5   86%     97     7    1   1 020   1.5     479   14 166
2014-10-14   181   1.8   85%     92     7    1     986   1.5     527   13 002
---- Panda 4.1 Google update boosts our rankings
2014-09-30   167   1.9   84%     92     5    1     917   1.5     342    9 856
2014-09-16   153   1.7   84%     90     5    1     875   1.5     367    9 412
2014-09-02   139   1.8   85%     87     4    1     837   1.5     355    8 284
2014-08-18   124   1.4   85%     86     4    1     778   1.5     175    4 503
2014-08-04   110   1.1   84%     83     4    1     730   1.5     162    3 533
2014-07-21    96   1.4   84%     82     4    1     691   1.5     148    3 240
2014-07-05    80   2.6   82%     78     4    1     645   1.5     137    3 061
2014-06-20    65   2.1   82%     77     4    0     602   1.5     107    2 421
---- End of the "Honeymoon" surge of new-site traffic
2014-06-04    49   3.6   90%     67     3    0     549   1.4     110    7 267
2014-05-17    31   5.1   87%     62     2    0     485   1.5     423   14 292
2014-05-03    17   6.7   83%     48     0    0     366   1.6     375    8 472


Note that a dip in Qs/day and Visits/day is normal for the first few weeks of a new beta - the initial surge in the first week or two is exceptional: after that, site-building is a steady, cumulative process after the dip.

Moderators have a view of the site's statistics, but they cannot share the specifics. So, I'm starting this post to keep track of the evolution of our publicly available statistics over time. Please feel free to update it every now and then.

I use the Area51 page as source for those stats, in addition to the Quantcast figure.

Questions per day is a rolling 2-week mean. Visits per day is a rolling two-week median.

Number of users who posted 2 or more upvoted answers per week, from this SEDE query:

The thing is, these stats are only crudely indicative. They're crude, but they're all that can be shared publicly. They are not yes/no indicators for graduation. They're crudely indicative, that's all.

For graduation, we need to build a body of expert content, and a community to maintain and expand it. It's much harder to quantify that - there are some stats that can be manually generated by the SE developers, such as the number of high-rep users who visited in the last few weeks, and so on. And there are publicly-visible indicators such as the liveliness of this meta site, and our chat.

So, this post is just a broad-brush view of where we've come from, and is emphatically not a direct measure of how close we are to graduation. And whatever the indicators stand at, the message is still the same: if you want this site to thrive, keep attracting new expert users, keep adding expert content, and keep maintaining the existing content, through sharing, editing, voting, flagging, closing, reopening and deleting wherever appropriate.

Users with at least 5000 rep now have access to limited site analytics available at /site-analytics that let you explore statistics on posts, votes and visits for our site.

• +1 @EnergyNumbers: Awesome, thanks, really nice; though bit puzzled why you'd post stats after saying, "those numbers are a very small part of the story, and are a legacy of Stack Exchange 2010. It's about the community. Those stats were a first stab at measuring proxies of the community. They're not very good proxies." What's the dealio? – blunders May 4 '14 at 2:26
• @blunders fair point: I've edited in an extra 3 paragraphs at the end to explain. – EnergyNumbers May 4 '14 at 11:48
• Thanks, guess my take is that: (1) SE has internal stats for sites leaving beta that are neither shared with the mods or the community; (2) SE, not the community, has the final say in if a site leaves beta; (3) SE has not published a DIY guide to growing a SE beta site; (4) the stats are very important, but only if they correlate to a healthy site; (5) SE refuses to publicly define healthy, even though it's possible; (6) unlike the main SE sites, pulling data is hard to do; (7) Managing goals, actions, results, etc via meta is time consuming. – blunders May 4 '14 at 15:20
• @blunders I suspect the graduation criteria are deliberately qualitative, even privately within SE. Yes, SE Inc is the only one that makes the graduation decision. "Healthy" is a qualitative thing, and it's still early days for this sort of site, so, like the famous quote about pornography: although it may be impossible to define, we know it when we see it. Setting up new standard stats is hard - it takes valuable dev time - and it's hard to see the business case for it at present. (NB though I say "we", I mean we generally - I am not part of StackExchange Inc) – EnergyNumbers May 4 '14 at 15:33
• Being "deliberately qualitative" does not mean that the sources and methods of those deliberations are unexpressible. Further, find it very hard to believe that there's much volatility internally within SE that they're unable define sources and methods for themselves. There's a reason why they don't publicly state the recipe for a Q&A, guessing is intellectual property related. – blunders May 4 '14 at 15:57
• @blunders - here's an alternative explanation: it's constantly evolving, as they learn more. And publicising the previous site's target would mean it couldn't be used as the next site's target, because the users would end up chasing narrow proxies, thus distorting them too much to be useful. – EnergyNumbers May 4 '14 at 16:01
• Really, from experience, fortnightly updating is sufficient. Updating after 3 days adds noise, not signal. And the chart really is best with weekly averages: it smooths out the noise, to better show the signal. – EnergyNumbers May 7 '14 at 14:11
• I made the updates in part because a single row of data is not as useful as two. At any rate, appears we have very different opinion on how to get stuff done, so I'll leave the maintenance of this page to you. – blunders May 7 '14 at 14:32
• I have been posting links to my Facebook and Twitter account, and even have links to here on my ResearchGate account. – user889 Nov 4 '14 at 9:12
• Peculiar that % Answered has gone up by 10%-points but Answers per question has remained constant. – gerrit Jan 20 '15 at 21:11
• @gerrit it may be that "% answered" means "% of questions with at least one answer with a positive net score". In that case, this would increase, without a corresponding increase in answers per question, if there has been a decline in the proportion of questions that only have answers at net 0 score or below. – EnergyNumbers Jan 21 '15 at 4:07