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.

Has Stack Overflow released any regular information about how much traffic the chat feature gets? When it first came out, it seemed like there was a lot more activity and interest, but whenever I wander in there these days, it seems pretty empty, which makes me want to go to chat even less.

Clearly the team put a lot of effort into it, and it's a great product that seems to have gone to waste. I'm curious from the team's perspective has the chat feature met its goals and lived up to its expectations?

Would it make sense to surface chat data more frequently into the main site? Maybe a sidebar widget with the top most recent chat messages on the home page? Or if I'm looking at a question tagged PHP, show recent chats in the sidebar from the PHP room. Stuff like that

share|improve this question
8  
You can't measure the pleasures of procrastination with traffic stats alone. –  Grant Thomas Jun 14 '11 at 18:53
1  
@Mr. Dissappointment I'm so filing that statement for future use. –  squillman Jun 14 '11 at 19:04
add comment

1 Answer

up vote 8 down vote accepted

Has StackOverflow released any regular information about how much traffic the chat feature gets?

Since the features you speak of already exist on all sites but Stack Overflow itself, I assume you're talking about chat.SO in particular:

enter image description here

  • Unique users: The number of known (in particular, logged in) users who generated an event in the given month. This is quasi-equivalent to the number of users who were ever in a room during that time.

  • Total hours: The accumulated amount of time, in hours, spent actively participating (if you will: man hours spent on chatting) during the given month (for some definition of this, so take the values with a pinch of salt).

  • Number of messages: The average length of a piece of string in all t-shirts produced in the given month, measured in nano-foobars.

share|improve this answer
    
ooh, where did that graph come from? Did you whip that up just for me, or did that get posted on a blog at some point or something? –  qntmfred Jun 14 '11 at 19:14
3  
@qntmfred Just for you :) No, we had that graph handy (we love data porn). –  balpha Jun 14 '11 at 19:20
    
figured i'd just point out that the growth in that chart roughly matches the overall traffic growth of SO in same period –  qntmfred Jun 14 '11 at 19:28
    
@balpha care to comment on some of the other higher level questions? Is the team satisfied with chat? Is it meeting the goals they had when they made it or do they perceive it to have missed the mark in some ways? –  qntmfred Jun 14 '11 at 19:30
1  
@qntmfred: I think it did, yeah. The whole chat was very much an experiment, with not too high expectations (in the beginning at least), and I'd say the expriment was successful. Btw, we still have the option of activating the chat ads on Stack Overflow (like on the others sites); I assume that would cause a huge boost. –  balpha Jun 14 '11 at 19:37
1  
How about now telling me how many times I've clicked on "chat" by accident. –  Anthony Pegram Jun 14 '11 at 19:40
1  
@AnthonyPegram: Your browser doen't seem to send the X-User-Actually-Intended-To-Click-Somewhere-Else HTTP header, so I'm afraid I can't. –  balpha Jun 14 '11 at 19:45
    
Are total hours counted by actual message commitment, or time spent "physically" inside chatrooms? –  Grace Note Jun 14 '11 at 19:48
    
@balpha: That figures. Truth be told, though, I often do click "chat" on stackoverflow when I intend to click on meta, and "about" on meta when I intend to click on main. This is mainly due to link placement (allowing that it could possibly float, I don't know) and the rapid switching back and forth. –  Anthony Pegram Jun 14 '11 at 19:52
1  
@GraceNote: Actual messages. As I said it's a pretty crude definition. If you want to know: The given month is broken up into 10-minute intervals. If a user has not spoken (i.e. posted a message) during a particular interval, they are considered not active during this interval. If a user spoke exactly once during this interval, they are considered active for 5 minutes during this interval. If a user spoke more than once during this interval, they are considered active for the whole ten minutes of the interval. –  balpha Jun 14 '11 at 19:52
    
Huh, that's a very interesting algorithm. Works well to catch... certain someones. That sates my curiosity on the system. Thanks. ♪ –  Grace Note Jun 14 '11 at 19:53
add comment

You must log in to answer this question.

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