Joi Ito's conversation with the living web.
Chart of weekly server requests since I started using Movable Type on June 25, 2002. I wish I could plot traffic vs. hours blogging. It would probably be a flat line...
I have never published my site stats before, but my March 1 analog (Analog is no longer. Here are some other tools.) run just finished, I'm sitting in a hotel room with nothing to blog and nothing to do, so might as well...

These stats exclude accesses from any of my own networks. ;-)

Program start time: Mar 1, 2003 06:34
Time of first request: Jun 25, 2002 11:56
Successful server requests: 4,185,420 Requests
Successful requests in last 7 days: 292,830 Requests
Successful requests for pages: 628,324 Requests for pages
Successful requests for pages in last 7 days: 61,250 Requests for pages
Distinct hosts served: 101,928 Hosts
Distinct hosts served in last 7 days: 14,541 Hosts

Some interesting referrer stats:
Radio newsAggregator 45,215, NetNewsWire 28,650, Google 21,453

Update: If you have an opinion about whether posting stats is in poor taste, please read the comments and let me know your thoughts.


Interesting to note how trolls cause spikes... What's that old saying? Any publicity is good publicity? ;-p Would rather get spikes for telling funny jokes...

(The previous comment links to a site talking about how it is in poor taste to post your stats...)

Hmmm... Interesting point of view. Actually, I think I saw BoingBoing post their annual stats saying it was tradition... (Not sure if it was BoingBoing)

This also highlights an interesting point that relates to the discussion about "is this a public place, or is this my living room." I guess that if I were a broadcast medium or a blog that didn't allow comments, I might agree with you. And in retrospect, I realize how posting my stats could be seen as poor taste. My view was that since comments section of my blog is quite an important part of my blog, I allow html in my comments and the names of the people posting in the comments section are in the same size font as my name (Howard's suggestion), and comments are googled properly, I think my blog is a community more than just "me." On the other hand, I did call this my "living room" and the name "Joi Ito's Web" is pretty self-centered.

Still, I feel like this blog is a group effort by the people participating in the conversation and I had a feeling that most people who participate here would be like to know how many people were "participating in the conversation"...

Anyway, I would be sincerely interested in whether other people feel posting stats is poor taste. It really never occured to me.

BTW, I would LOVE to see Doc and Dave's stats and I definitely would not feel inferior if they romp my stats because it makes me feel better to know that so many people are part of "the community".

Maybe it depends on your relationship with power law curves...

At risk of being criticized again for posting stats, the comments on the weblog far exceed the number of entries.
Entries: 550 Comments: 1697

It seems the spikes didn­­'t have an accelerating effect on your hit count, though. These people came and went but didn't think of coming back.

I think Zeldman was talking less about sites with comments--like yours--than he was about sites that are write-only.

One of the things that can be very deceptive in the world of participatory computer-mediated communication--from mailing lists to weblogs--is the size of the invisible audience.

Posting a comment to your blog may seem to many people like a personal and conversational act. And on some levels, it is--you've always made me feel very much like an honored guest in your virtual house, ever since my first venture into commenting (on your Diet Coke post...). But I made my comments knowing full well that thousands might read them (and even then, I didn't realize just how many thousands until you posted these stats).

An occasional post that reveals the size of the audience we're part of is, IMNSHO, a good thing. Not only does it help potential posters to understand impact, it also helps all of us to understand just how far-reaching blogs like yours are.

Thanks for coming to my rescue Cory and Liz. Liz, that's EXACTLY the way I hoped that our community would feel and why I call this a conversation. I think it's also important that the comments on my blog end up become part of the static html file instead of some pop-up window or on a separate site. I think that the comments are an important part of the content on the site and would like people to feel that way when they write here.

I think Zeldman's point is only valid when the discussion is solely about numbers. Numbers are boring, and don't tell much of a story by themselves. I don't see any problem talking about statistics aside from raw numbers, and looking at patterns over time is definitely where the interesting stuff happens.

Take for instance this entry, I am fascinated by the abnormal peaks in the graph. The y-axis is meaningless, but the change in Y over time is what I'm looking at. What happened in late August of last year to cause the first bump, and what coincided with the peaks later on (looks like one in Nov, Dec, and Jan)? Controversial posts? Traffic spikes from getting press?

Also keep in mind everyone seems to use a different metric. I don't really care about hits or requests, since that counts every image on a page and doesn't translate well to a size of audience. I'm interested in pageviews over time on my own project, while others I know swear by distinct IPs over time as the only barometer of audience size, but overall nothing is a perfect measure of readers, only an approximation.

Yes. The peaks.

The little bump coincideds with my trip to Aspen for Brainstorms, a conference. I posted my notes and the links circulated among a small number of people. The traffic is so low at this point that a few people dropping by shows the surge.

The first big Nov peak is when the online diary community in Japan ripped me apart on the anonymous slashdot like site in Japan called 2ch for trying to push blogs as something new when many people had been doing online diaries in Japan for a long time. They had some good points, but it was a troll-mob scene for awhile.

The second peak is when I launched my moblog.

The third peak is when I visited Silicon Valley last time and went to SuperNova and met a lot of folks face to face.

It's kind of interesting how meeting other bloggers face to face creates a flurry of activity and an brief increase in traffic as people check out your site and big blogs such as Dave Winer's Scripting News link to you more frequently when you cross paths.

Since you mentioned you'd be interested in stats of the Doc weblog: There was a mention of numbers there last tuesday.

I discuss the relation between traffic stats and freedom of speech on my blog (I trackbacked that to this topic). I close with:

So, my short answer to the discussion: Never mind. Post whatever you want, including stats. It's a free world.

Be proud babeee you are top stats in my book :)

there is no shame in seeing progress :)

I don't mind sharing my stats, even though they're way under yours or boingboing's.

I don't really think a lot about building attention. I do keep playing with the site, and with various blog toys, partly as practice with various web tools. And I keep blogging away because I like to write/communicate. Everything I post goes to the small email list that used to receive occasional writings I did as 'cyberdawg' years ago, soon after I left FringeWare.

First started reading Joi Ito when I heard about this enthusiastic new blogger who took the ball and really ran with it from...Justin last July? Yeah, this looks like the reference. But stats like this are hard earned with best intentions, well deserved with best wishes. Troll tromped favorite project, emergent democracy, but was thinking about it just today when the sky cleared and the sun came out. In the USA, there's no real experience of democracy. That's why it's good, important. Will work harder to keep little end of conversation going.

I think that stats are a good way of showing the wider audience the site/conversation gets. The comments are a very good way of showing who is actively taking part in the conversation. However, I have been reading this site for a few months now, along with other blogs, and very rarely do I make a comment. Therefore the stats indicate the 'lurkers' that do not feel like participating.

Hi Andy. Well, thanks for commenting. What do you think it is that keeps people from feeling like participating?

There are several reasons that I can think of:
1. The viewer does not have the confidence in their own opinion.
2. The viewer feels that they cannot voice their opinion well enough in comparison to other viewers.
3. The comment that the viewer wishes to express has already been conveyed in a more eloquent way.
4. The viewer does not feel they can comment on a subject that they know little about.

Read into this what you want (some or all may apply to me).

I we should try to do something about #1. People should feel confident that their opinion matters and #2 should be ignored as long as the substance is interesting and it isn't rude. I think #3 and #4 may be good reasons if it lowers the noise. For instance, I get much more noise on my Japanese blog because I don't put as much attention and I THINK people are less inhibited about posting noise. Interesting phenomenon.

5. (maybe) Participation is just plain more effort. That doesn't mean people are lazy, but they only exert themselves to speak coherently and originally about the few things they are most interested in.

6. (hopefully) You've inspired entire blog entries which people don't want to write twice.

7. Monitors are shaped too much like television sets, subliminally suggesting passive receptiveness rather than interactivity to users.

I think David's number 7 is the main reason. I'm still fairly new to blogging, but have spent many years screwing around with communities and the biggest hurdle seems to be people don't feel the need to be interactive. They are so used to the experience being passive (working off line, reading web pages, etc.) that unless they are doing something in a specific "location" (email, messaging, etc.) they feel that they are just spectators.

The only way to really circumvent this that I've found so far is to increase the amount of noise in the group. Or, more specifically, change the dynamics of the group so that limitations on subject matter are very minor or non existent. Which allows people to more easily "get to know" the people involved and feel more apart of the community.

I'm not saying that Joi should do this, obviously, I'm just tossing my findings out into the mix. Acctually, I'd love to hear of any other similiar theories.

Interesting. This came up in the dinner discussion with Larry last night as well. How do you get people to engage. One reason for a representative democracy is so that you can have experts who spend time to focus on issues that the public doesn't have the attention for. One aspect of direct democracy which is essential is to draw out the public from the "silent majority" and get them to learn, think, debate and decide. How.... I hope there is another way than just increasing the noise. ;-)

I guess I should have pointed out that my noise comment only seems to work given a community of a certain size.

Back at the end of '98 I found myself in a position where I was involved in an open community at it's inception. I used it as an opprotunity to test some of my theories. The two main theories were the noise theory and a theory on the importance of open real life gatherings (anyone from the community, reguliar or lurker, was free to come). The gatherings are good for pulling lurkers into the mix who are not used to dealing with online communities.

In the end, I found that once the community hit a kind of critical mass, things fell apart. And the group switched from a very open and friendly group to a very cliqueish and closed group.

So, when I gave my theory above, I was talking purely about a community of around this size or smaller.

The problem with increasing noise, obviously, is the old signal-to-noise ratio problem. As you increase participation, how do you make sense of the results? How do you extract the relevant information? How do you keep people from feeling as though their voices are lost in the crowd? Direct democracy in a room of 100 people is very different from direct democracy in a country of millions.

ps Joi, your comments cookies seem to be directory specific, so information on a comment from "03/02/" won't be automatically entered on a page in the "03/03" subdirectory...

Liz. Made some changes. Let me know if the problem persists.

Again, I admit that the noise theory does not scale. I was not considering large groups when I came up with the theory or when I posted about it here.

Leave a comment

4 TrackBacks

Listed below are links to blogs that reference this entry: Joi Ito's Web site stats.

TrackBack URL for this entry:

test... Read More

There is some discussion at Joi Ito's site about whether it is in bad taste to disclose traffic statistics. I Read More

Templates for my MT-based courseware, and step-by-step instructions on how to implement it. Not for the technically fainthearted; basic MT template and plug-in installation knowledge is assumed. Read More

Okay, I think I’ve “gone about as fur as ah c’n go” for this first version of the courseware. I’m ready to call it version 1.0, I guess, with all the caveats that go along with that. You can see it in action on my fall cou... Read More