Apparently provoked by Weinberger's blog on Shirky at PopTech, we seem to be having a reprise of the power law brouhaha, with A-lists and C-lists exposed and waving about. Meanwhile, Jarvis puts up a thoughtful post on blogs as personal media, and (to my mind) Shirky runs a bit off the rails trying to map authors' intentions onto log-log paper.

May I suggest that we have broken this camel's curve's back? Overloaded it with explanatory tasks? We all know we've been oversimplifying, right? Just to take one example, go back to Adamic and Huberman's original work on power laws. They used real session logs. We use links as attention surrogates. They explicitly looked at the differences in power law coefficients among differing site populations (and discovered that pr0n sites are more 'democratic' than .edu's - go figure). We try to keep reminding ourselves that the One Great Power Law Curve of A/B/C listers is really drawn from multiple, overlapping communities of interest, with different characteristics, but the metaphor is not helping us - it's concealing what it should be revealing.

Naturally, I have one alternative to propose, with perhaps some more explanatory power in directions I want to explore later. Consider this post to also be a bleg to others to pound on, critique, and clarify the metaphor, and especially to any lurking graphics or Mathematica gurus to help illustrate it - why will become clear.

I said one of my gripes with the Power Law was its inability to represent 'in frame' the landscape of interests. Let's fix that first. So imagine a flat X-Y plane - any floor will do - that's a map of topics in two space. Not just randomly assigned, but clustered by similarities. We can discover those empirically by clustering blogs and stories based on links, or dumping the full text of blog posts and news stories and looking for word co-occurence patterns, or both. For the techies reading, I'm talking about something like a Kohonen map. Make it as big as you like, but finite. To illustrate a little, let's stand on the map at 'environmentalism'. Somewhere near we can see 'Lomborg', 'ANWR' and 'Kyoto'. A little further off in the direction of the latter two is 'fuel cell' and in the mid-distance behind is 'Saudi Arabia.' 'Anime' and 'Reality TV' are lost in the haze, in an unknown direction.

There are gross simplifications here, too. Let me pause and acknowledge them: This is a lot more than a two dimensional problem, but I've squashed it down for the sake of visualization, and this has side effects. You might suppose that 'Edward Abbey' and 'Wallace Stegner' should be nearby as we stand on 'environmentalism', but instead they're stuck with a blob of 'Western authors' somewhere else. Also, let's remember the map is not the territory. We're not pretending this one represents what's in any particular person's head. It's a space of externalized symbols, in which we all make the gestures of linking and browsing, supposing that we're representing the Web medium.

Now, up off the ground, and let's add a Z (vertical) axis. This axis is attention. Attention is finite, and conserved. If you're reading Instapundit, you can't also be reading Due Diligence (dammit). Let's give every reader / viewer in the audience for our medium 100 poker chips, and ask them to put the chips on the map areas representing how they would ideally like to spend some time and attention this month, available content permitting. Ready? Go! We now have several tens of billions of poker chips, piled on top of the map, making an irregular terrain. Here's a very modest sized pile on 'venture capital'. Somewhere close by is a much larger tower around 'Microsoft.' Away in the distance is a stupendous peak of chips burying the entire area of 'war' and 'politics'. Remember these concepts are fuzzy, so we don't actually have neat stacks of chips, they are messy piles that slope off to recognize the differences in actual individual appetites.

Again, some of the technical will recognize this as similar to a 'fitness landscape.' I'm going to call mine an 'attention landscape', cause that's what it's built from. On this landscape, we can recognize some familiar patterns. The noticeable peaks, towers and piles are going to correspond to specific power law curves, representing niches of interest. Assuming our landscape is actually the blogscape, that's Glenn Reynolds up in the stratosphere on top the big peak, and there's Scoble on the Microsoft tower, and the Venturebloggers and I are kicking sand at each other on the VC hillock. (Again, note the simplification - none of us really occupy a single fixed point - topic drift, and subject bundling, are everywhere.)

Now we can talk a little bit about media strategies. If you are following a plan to get as high as you can, on the biggest peak you can get to, you are a broadcaster by intention. If you pick a congenial nearby tower, and try to get to the top of it, you're a niche player by intent. If you forget the climbing, sit down where you damn well please, and scribble away, you are a vanity publisher (or a diarist). Good so far?

OK, since this is my imaginary landscape and I can play God, let's change it to represent the print medium. Away with you, Prof. Reynolds. Cue the special effects, and pour in the water. That's enough. Now let that water represent the fixed costs of entry in the print medium, restated as the amount of attention you need to monetize to survive. You'll notice most of the landscape is now underwater. So are the business models of those on the small peaks. As for the vanity types in the valleys, the rich ones are dropping change at Vantage Press, but most of them are piling up their works in the closet.

Now, back to the blogscape again. Drain most of the water - fixed costs are way down. Even many of the smaller hills are dry - making money - if only indirectly as a form of virtual advertising. The diarists are still underwater, but the snorkel is pretty cheap. There's also a lot of life on the lower slopes of the big peaks, more microclimates and ecologies to exploit. If you're a scribbler, you've got a much better chance of occupying a location close to your tastes, and not drowning, even if you never climb the peaks. If you've been a reader, the costs of changing role are much less. Personal media.

Did you spot the big handwave? With luck, I got you to buy into the notion that this metaphorical landscape is solid and fixed, like a real one. It's nothing of the sort, of course. Since this post is getting too long, I'll explore that fact in the next one.
9:42:19 PM    

Awww, couldn't have happened to more deserving people.
9:33:59 AM