KernelPop

Wednesday, December 14, 2005

PushWrite

Couple of Kernels have popped. It seems that most of the focus on read/write development are still focused on building better 'pull' technology to navigate the sprawl of peer-based publication. To date, I have found nothing that places value on 'push' technologies that deliver optimized, customized content from this same sprawl.

I see a real opportunity on this front. Part of this is personal bias. I value 'push' media as a way to reveal items of interest that I might not have found on my own, or might not have time to look for.

In the context of Umir Haque's take on emerging Micromedia Economics (massive PPT on the topic, my basic/simple discussion), the relative scarcity of Attention increases (vs Production) as the Media value chain shifts towards peer-based production. This creates opportunities to add value in distribution by building tools that economize Attention. This is widely observed, but remains focused on 'pull' methods (as far as I can tell). I don't know if this is because search is in vougue, or if hyper-efficient pull approaches will completely replace push methods of distribution in a micromedia environment. I think its Google hype.

In fact, I think that personalized push distribution would represent greater distribution efficiency than any pull approach. The 'find me quality' techniques and mechanics would actually be almost the same as search, only everything is preprocessed and presented to the user without any user action (though a feedback mechanism is the keystone).

Some of this already exists in various forms. Google news alerts are a good example of the functionality at its most basic form. In a micromedia environment the key will be to leverage the information expressed in peer networks -- feed subscriptions, user posts & tags, and expressed preference -- to optimize relevency & merit. The good news is that most of this information already exists in all of these web2.0 services.

Examples of Value in Push Media

Exhibit A: Yahoo! LaunchCast. I love the personalized radio station. It is not perfect --some things it chooses to play I hate -- but I have also found music that I love that I may have never found in a retail situation, and certainly never heard on commercial radio. This is all driven by user expressed taste, and network preference clustering (or something along these lines) that personalized value based on media attributes (like genre in the case of music) and aggregations of peer taste expressions.

Exhibit B: Satellite Radio. I love the heavily niched push in this environment for many of the same reasons I love . I actually use Satellite & Y! LaunchCast together -- listen, learn, seed, listen, repeat -- to deliver myself a very fresh, current and high-quality music environment (though my wife would disagree).

I've had debates with people about how best to spend $12 per month on music -- 99c a song and an iPod, or Satellite+Y!LaunchCast. It usually comes down to how much you travel, and how you get to work. iPod fanatics will talk about the vast array of tools that they have at their disposal, which includes user playlists and other bundles of people with similar tastes. They value a certain degree of push also to help them keep fresh. The value of push is in the time savings to idenitify high quality content.

Exhibit C: Feed subscriptions. Once you've found a source that you like, you add the feed to a collection of things that you want 'pushed' to you -- through some RSS Reader or a web-based service like My Yahoo! I've been juggling my feeds around to make sure that I don't miss anything that might be relevent across a number of sources. I want this stuff pushed at me.

But I have hit a limit. After adding more and more feeds, I am hesitant to add new things. Its just too much. A personalize push engine built from expressed taste, would be a great way to maintain bredth without all the noise.

Where 'push' becomes so important -- queue the web2.0 background music -- is at the edges.

0 Comments:

Post a Comment

<< Home