Joi Ito's Web

Joi Ito's conversation with the living web.

I get this feeling that the diffusion of new services and technologies such as blogs and social network services are not normal. Normal diffusion patterns are sort of bell curves that track mass media attention and other factors including effort on the supply side. With social network systems, there seem to be regional explosions of users. Orkut now has more Brazilians than Americans and I have yet to hear a good explanation of why. There are very uneven proportion of bloggers in different regions. The last I looked, Poland and Iran seem to have an unnaturally high number of blogs. Does the digital word of mouth nature of social software make it's diffusion faster (viral) and non-linear? Is the diffusion of other technologies in these markets and segments that are networked also change?

Is there anyone doing good work in this area? I'm sure marketers have their theories on this. I can imagine anthropologists and sociologists also studying this. What is the right way to study this?


"Linked" by Barabasi should give you an answer. Maybe Power Law Distribution is the answer and not the Bell Curve."The Power Law Distributionof a scale-free network predicts that most nodes have only a few links, held together by a few highly connected hubs".

I am also curious about the unnaturally high number of blogs in Poland. I remember you mentioned it when we met in Stockholm as well. Where does one see that there are many bloggers from a specific country?

Power-law works a lot differently online than most people would suspect, and small, not large, communities and cliques dominate the internet. The strength, organization, and evangelization of certain communities makes a huge difference in their growth.

One of the things I did for LiveJournal was create and promote specific communities that were beneficial for LiveJournal's goals. Promoting our developer-oriented communities, for instance, to attract people who might contribute to our codebase. Likewise, I encouraged the growth of international communities for specific countries, as we were working on translating the interface and documentation into numerous languages and I wanted to have communities that could be used to jumpstart the translation projects. It had the added benefit of making a very noticeable difference in site growth in those countries to have communities geared towards those nationalities, however.

This can be seen especially in LiveJournal's Russian community. Two of LiveJournal's larger contributors in the past were Russian, and they helped to form and promote a very large, active, technically-minded Russian community on LiveJournal. Today, there are 67491 Russian users on LiveJournal, and another 7604 Ukranians. Compare that to 46287 Australians and 6929 from New Zealand. Language is clearly not a huge barrier, especially if the system supports multiple languages. There are numerous Russian communities on LiveJournal, and one person wrote a paper on the subject.

I still encourage communities like this, btw. My latest community on LiveJournal, insideiraq is a community for over 50 LiveJournal users who are, were, or will be in Iraq soon. Most of the members are soldiers and contractors, with a few Iraqi expatriots. Much of the conversation takes place in "friends only" posts only visible to members, so it's a kind of back channel to the conflict.

Another notable difference for LJ has been how early LJ emphasized comic strips. When Clay Shirky was writing articles about LiveJournal and power-law a year or so ago, I sent him the following email which might be of interest to you:

Hi Clay... I wanted to pass this along, as I thought it might be useful for your work: (link no longer live, but available via )
It's a directory that was made independently of LiveJournal that shows all of the RSS feeds LJ imports currently (~2700) ranked by the # of readers.

The interesting thing about this to me is that it shows some clear categories of use. Unlike most such listings, the top 10 are all syndicated humor, with only The Onion not being a comic strip. I suspect that I had an influence on this, because long before LJ ever supported RSS syndication, it still had friends list syndication, and I went way out of my way to actively recruit indie comic strip artists to join the site and post their strips online, so when syndication came along, it was second nature to do this.

The top readership is Calvin and Hobbes with 2754 readers, and by the time you get down to ranking #10, Doonesbury, you're all the way down to 926 readers. The next 50 feeds are so are very similar to the kind of feeds you'd see on other sites, and go from #11 having about 900 readers down to about 70 readers at #60.

At that point, though, something interesting starts to happen. LiveJournal users (and former LJ users) who have external weblogs start to show up in the listings. Refriedgringo, for instance, is seen as having 74 readers, which is actually a lot less than some of the LJ accounts out there. In comparison, you get people Like Doc Searls only getting about 7 readers.

In other words, although there does seem to be a power law-distribution going on here, things in practice are a lot flatter than you'd suspect, because the web is made up of numerous communities of users of various interests, locations, languages, etc. It's quite possible within a huge community like LiveJournal for many, many people to have more RSS readership than even sites like Slashdot or Wired, and even if you get an ouside site at some point and have people add your external weblog as a "friend", you can still have as many readers within that community as Dave Barry, Salam Pax, or Lawrence Lessig... not to mention the readers that you get from other webloggers.

So, yes... there are superstars, but their stardom is spread pretty thin, with a whole lot of room for people to garner attention, at least within their circles. This indicates to me that most weblogs still tend to be focused towards specialized interests... the conditions are such that an Indian technologist could potentially have more readers within their community than Slashdot, so in that sense, anyone can play with the big boys. And, as I pointed out earlier, the most widely adopted feeds are humor, and half of those are independents who aren't nationally syndicated.

Perhaps the one thing which we can best agree upon is that which makes us laugh. ;->

I think academia is still a little bit slow in this area. It would probably take time before some of the best schools in the West will formally study and research blogs and its impacts on our society. However, some of the schools have started to take advantage of them.

There have always been quirks in uptake of products by apparently similar countries and communities. There are loads of examples. Barney the Dinosaur was massive in Ireland but never really caught on in the UK. The British Indian community is said to generally prefer Huggies over Pampers nappies. The Germans consider David Hasselhoff to be a talented singer.

You should look into the Bass model . A simplified version boils down to 2 parameters where one of them is moouth to mouth adoption rate. My guess would be that this parameter depends on the community/nation adopting a new technology.

C U!

-- Mario Valente

just read chapter 3 of ED, now I feel a little stupid. I'm sorry.

Don't feel stupid Hans! ;-) I "restarted" reading Linked after you mentioned it.

great, joi. Do you have an erdoes number ?

Why that many Brazilians? First, Brazilians are social animals. Second, there are a lot of Brazilians. Third, no other social network was big in Brazil before Orkut, not even Friendster. Fourth, when the number of Brazilians start to swell there was a kind of campaign for people to invite friends and make the numbers even higher. There was media coverage, on newspapers and TV, too. Put all of those together and you can see what happened.

The pervasive proximity enabled by instantaneous communications is a retrieval of oral forms ("secondary orality" in the words of Walter Ong). This can be clearly seen in weblogs, wikis and social networks that take on characteristic effects of an oral society, despite the fact that they are written with characters and must be "read." (As Marshall McLuhan would say, the extreme acceleration of written language by electricity results in a reversal from literate form to oral form.)

The relatively greater acceptance of these capabilities by emerging societies, compared to Western societies, can then be understood to be oral societies adopting modern oral forms. Unlike Western societies that have to cope with the disruption of the reversal, "eastern" societies do not have to reframe their collective cognition.

"an unnaturally high number of blogs?" and no one (from poland) comments this?
not so simple.
it's like in mobile phones market. Japan, Korea and Poland seem to have an unnaturally high number of mobile phones. True?

most of those blogs (in Poland) live short. most of them are "diaries". but not all.
yes, we like conversation, we also like computers, web (and mobile phones).
we don't like OFFICIAL MEDIA - and maybye that's the reason for unofficial conversation - i.e.blogs.

-->Mark Federman: true.
it's a question of acceptance. but with Brazilians it's also true. 40 mln of people -- compared with 8 mln in Czech Republic...
25% of Poles have internet acces. most of them via telephone cable.
but we have FASTEST growing number of internet users in Mid-Eastern Europe.

its young, educated and (a little) frustrated society. easy to speak.
(Ito-san if You are...)

here is a blog (= polish blog from Japan!) of my friend, (living now, and studying in Japan):
in polsh, but he is a smart guy...

and here is a blog of well educated man from Poland (now in USA, i think)
this page is also in english:

and remember:

4 TrackBacks

Listed below are links to blogs that reference this entry: Technological diffusion patterns.

TrackBack URL for this entry:

From a Joi Ito today post: Joi Ito I get this feeling that the diffusion of new services and technologies such as blogs and social network services are not normal. Normal diffusion patterns are sort of bell curves that track mass media attention and ot... Read More

This is actually mostly pretty old, and mostly for my own benefit. Mimi Ito's article on how mobile phones affect social interaction amongst teenagers, via Many2Many, boingboing, Joi Ito's blog, and probably zephoria's blog, too. Joi Ito on technologic... Read More

Off to Netherlands in a few minutes. Among the things that I haven't had the time to fix/finish before leaving: * buying a new battery for the Titanium, since the old one has been dead for several weeks now. So in these days I'll be equipped with not s... Read More

Off to Netherlands in a few minutes. Among the things that I haven't had the time to fix/finish before leaving: * buying a new battery for the Titanium, since the old one has been dead for several weeks now. So in these days I'll be equipped with not s... Read More