Sunday, October 23, 2005

The Blog Reed-ing Law of Networks?

So I was doing my weekend blog-reading, and came across Fed Wilson’s blog on Reed’s law – and he pointed me to Bubble-Generation and Nivi, neither of which I was familiar with, and who had excellent posts. I have subscribed to their blogs and if I continue to find them inspiring, I am sure that they will point me to other posts and blogs that I might find interesting, some of which I will subscribe to. This is the power of the blogosphere network, and in this network, the ‘peer-produced’ content is not so much the blogs themselves as the links to the other blogs – they determine the value of the network!

All those bloggers could have been there happily blogging in their sub groups, but I would have probably never found them (Certainly not via a google search – what would I be searching for?) had it not been for Fred’s post.

All networks are formed of sub networks, and blogs are a perfect example of this. A group of mutually-admiring friends may be blogging for ever and ever – no thousands of such groups may be blogging for ever and ever with out really adding to the value of the network exponentially. So the main issue I would have with Reed’s law is that it does not account for the connection between the various sub groups. And that is where I think the real value lies (not the fact that they can be formed). Reed and Metcalf calculate a theoretical value for a network (and its subnetworks) but as these sub-networks become more and more specialized, what matters more is the actual interconnectedness of the sub-networks.

Take ebay’s example from Bubble-Generation: Could it be that ebay has turned into a series of unconnected sub-networks which don’t add value to each other exponentially, thus accounting for their lack of “network scale economies”? Are ‘car sellers’ and ‘antique dealers’ on ebay simply two completely disjoint sub-groups adding little value to each other? The fact that I would go to to find either does mean that there are branding economies of scale in the traditional (pre-internet) sense, but not necessarily anything exponential (unless ebay-car-dealers want to barter for antiques.) (Interesting that Reed’s original paper, which is admirable in its vision, cites ebay as the prime GFN.)

I would argue that the blogosphere is the ultimate social network, in the way that traditional(!) internet social networking sites can never be (unless, of course, they totally open up to each other.) Each social networking site adds a sub network with structured links between each other. (eg User question: “What school did you graduate from?? We’ll try to connect you to people from the same school who answered this question.”) Compare that to the blogosphere network (and the web in general) which is all about unstructured network connections via a variety of tags, search possibilities and most importantly direct linking to other bloggers. (After all, it is those links that made Google’s search algorithm possible!)

So, back to reed’s law – I wonder if there is way to mathematically model the connections I mentioned in the first paragraph above. Or perhaps there is a way to measure the value of that interconnectedness using empirical data on the way that bloggers and blog readers travel from one sub network to another and sometimes create hard connections between these networks, making them even more valuable. For example, let’s assume that bloglines (and/or other rss readers) allowed us to see the historical data on their users’ subscriptions (in the way that sourceforge allows researchers to study the behavior of the open source community). By mining this historical data, we could then see the progression (over time) of the various blogs the users are subscribed to. We can then correlate that with the various hyper-links and trackbacks that each of the subscribed-to bloggers has created to the others over time, and see to what extent these links create added readership and subscriptions. Since each subscription by a user adds value to that user by definition – why else would they subscribe? – we can get a sense of the actual practical value of the blogosphere as a “Group-Connecting Network”…

And there we have it: A new “blog-Reed-ing law of networks”…


Note: I realized that I have been talking about the value created in general, not the monetized value accrued to a single company. Of course with open networks such as the blogosphere (or open source communities for that matter), extracting $S from the value created is not necessarily straight forward… except if you recognize the potential of that value early on and build new monetizable value on top of it!

Sunday, October 09, 2005

Declining Price of Software / Web 2.0

Somewhat inspired by all the Web 2.0 Conference discussions, here is my take on Web 2.0:

  • The price of creating new web based applications is plummeting, largely due to the Open Source Software and the Open Source movement's cultural influence.
    • Think of what it takes to build some pretty sophisticated functionality using 'free' open source infrastructure (like LAMP or RoR), and support it with minimal resources. Compare this to the money you would have to spend to buy and support a Solaris/Oracle platform prevalent in the late nineties for example, and the point is made.
    • What's more, the culture of sharing freely (which I think has been enhanced by the success of open source) allows for more and more new offerings which are free... while the companies creating them figure out their revenue models, which in turn, do not need to be outrageously high to cover their low costs!( I liked this related post on ning.)
    • This creates a virtuous cycle of low cost tools enabling low cost apps, some of which are low cost tools...
  • Plummeting costs of supplies inevitably result in experimentation and innovation
    • This is not the first time this has happened. I go back to the chart I had used before, which continues to amaze me. (Credit goes to Kurzweil for thinking of a way to quantify this, and Larry Augustin who first got me thinking about it in this way.) It's pretty simple grade school economics - lower the cost and demand goes up! And this is on a logarithmic scale.

    • The form this innovation is taking is that you get a lot of small groups (2-10 people say) that are creating very interesting software with minimal overhead. They sustain low burn - sometimes, these are people who became somewhat independently wealthy during web 1.0, and themselves may have low overhead lifestyles, and thus have a longer time frame before they need to monetize their creation. The idea is that they can sustain the burn and get ad-sense revenue or something while the site becomes more popular and then sell it for a few million to one of the internet giants
    • The low cost infrastructure also means that web 2.0 businesses can go after smaller segments of the market to become real businesses (and potentially expand from there) - thus the long tail.
    • It is a great environment for creative destruction. And before the destruction comes, you have a growing and immense supply of low cost applications being built on the web.
  • I think the cost issue is a critical piece of web 2.0 (as I see it.) Clearly there are some other pretty important common themes.
    • Broadband: Certianly, the web is 2.0 because it is relying on the infrastructure (actual mbps) and user base (network effects) that came on with web 1.0.
    • Openness: ... as in open source, or APIs. It's all about linking content, code in innovative ways.
    • Participation: I think a critical central theme to web 2.0 is the concept of participation by users, whether it is to contribute to the infrastructure (bit torrent), to the marketing (skype) or in the O'Reillian architecture of participation (with tagging, sharing, amazon reviews etc.)
    • Revenue model: So all this is somewhat enabled by the micro revenue models specifically enabled by google ad-sense and Amazon affiliate programs... I recently talked to abusiness school student who created a site before bus school, and the site is "self sustaining" while he does his bus school and gets another job. Now how cool is that???
That's all for now - the question I really need to be thinking about is: Is all this good for venture capital, and investing in general?? And for how long? When will the destructive part of 'creative destruction' start to rear its necessary head?

Tuesday, October 04, 2005

Tim O'Reilly - Visionary

Reading Tim's recent article, I remembered his presentation at OSBC last year when he was talking about his ideas on the architecture of participation. He went on about the difference between Amazon applying principles of participation, as opposed to B&N which did not... and then continued to talk about how mapquest (and the other mapping services at the time) had not figured out a way to do that and so they were not able to create value. It marked me that he should single out mapping the way he did. Why pick on mapping? It just seems so visionary for him to go on about how mapquest was failing when he could have picked so many other examples... visionary because 'Google maps' is, for me in any case, the poster child of Web 2.0, not only because it helped to popularize Ajax, but also because of its use of APIs, which has made it a staple of so many mashups today.