Democracy is Not the Issue Here, Part 2

Continuing our nascent foray into social theory blogging... with another ridiculous wall of text:

I concluded my first post by claiming that "money buys noise". But first, let's take a closer look at Digg and other aggregators, and then onto "long-tail" theory and its supposed rebuttals.

Digg, unlike many sites, is straightforwardly democratic: users vote various pages up or down. For popular items there is a feedback loop: items with more positive votes are displayed more prominently, and thus have a greater opportunity to receive votes. It's negative voting that presents the most problems, however, as it allows controversial items to be denied significant consideration. As a result, we would expect the average successful item to have a broad base of appeal be quickly digestible (no Digg item that takes more than one sitting to process will have any success). I have no extensive experience with Digg, so I'll leave the final analysis to others. The governmental analog is a kind of collectivized democracy, a cross between classical Athens and Soviet Russia... with a nod to the immediacy and frivolity of the 24-hour news cycle.

As an adjunct to anarchy, Digg is not terribly troublesome; as long as there exist Diggers who are exploring the Internet in depth, uncommon items will have at least have an opportunity to rise. As a content aggregator, Digg is supposed to make interesting content easier to find; the complexity of the interlinked web is reduced to a single, linear space. The argument against this is that it reduces an infinite homesteader-style prairie into a crowded tenement (with a penthouse on top). It certainly does make things easier-to-find, though... Ah, there it is. If you like what comes to the top, Digg improves the signal-to-noise ratio of the Internet. But if the things on Digg are so common that you've already seen them, you aren't going to stick around (or do you? and do you vote them up or down?). If the aggressive explorers abandon the site, this should produce a vicious cycle of ever-more-limited content and ever-less-intrepid content locators.

Of course, it should not surprise anyone that popular things are more popular than unpopular things, but there has been a great deal of sturm and drang over just what the coefficients should be in this equation. Chris Anderson of Wired is credited with coining the term "long-tail" theory, to describe the phenomenon where, on the web, it is possible to profit from selling a large number of unpopular works. For example, let's look at this counterpoint from the Wall Street Journal, that constant friend to egalitarian theories:

"By Mr. Anderson's calculation, 25% of Amazon's sales are from its tail, as they involve books you can't find at a traditional retailer. But using another analysis of those numbers -- an analysis that Mr. Anderson argues isn't meaningful -- you can show that 2.7% of Amazon's titles produce a whopping 75% of its revenues. Not quite as impressive."

I hope what you're taking away from this is that the glass is clearly half-empty, and that Amazon sell a lot of books -- millions of titles in fact. Amazon sells over 25 million different titles, 2.7% of which is about 675,000. Quick, how many books fit in the average bookstore? This is not exactly the New York Times best-seller list. And it fails to consider the most important question: how did these books become part of the privileged 2.7%? How many of them were at one point part of the tail? And, and one more thing... Are they any good?

Back to that in a minute. I have someone to be angry at, the writer of this Slate article :

But according to Elberse, that sort of anecdote is the exception. The reason? We're not very adventurous. Elberse examined the rental habits of customers at Quickflix, a Netflix-like service in Australia. She found that no group of customers exhibited "a particular taste for the obscure." Sure, a small number of customers regularly rented films from deep in the catalog—but they tended to be people who watched a lot of movies generally and so had much more "capacity" for venturing into the Long Tail. And still they chose a lot of hits: The most widely traveling Quickflix customers picked only 8 percent of their rentals from the least popular of available titles and 34 percent from among blockbusters.

Science journalism is doomed. We aren't given anything to tie those percentages to. We're given a lot of arbitrary mosts and leasts. But let me pull a Wall Street Journal here: there is some portion of Australian mail-service film-renters, for whom two-thirds of the movies they watch are not "blockbusters." And of course, that's ignoring that such a service appeals to people who watch a lot of films (because it is monthly), and encourages them to watch films indiscriminately (because they can get as many as they want). But let's just look at that "particular taste for the obscure" line. Of course there won't be such a group... if I take the sum of all people who watched a film of (let's say) 92% obscurity, I am actually taking the average of several different obscure tastes. Where do I end up? the center! magic! Whereas if I take the group of all people who have watched rented no "blockbusters," clearly, I have such a group. Furthermore, movies are probably the most homogenized major media.

You want a long tail success story? Talk to someone whose business model didn't exist fifteen years ago. Talk to a guy selling T-shirts for a living, while providing content for free. Talk to some guys who can raise millions of dollars because they make funny pictures about video games. Yes, webcomics are our cultural standard-bearers. God bless 'em.

Wait, actually, let's look at films... why are there films that everyone has seen? Because people are only aware of movies that are showing in theaters, which are large, expensive buldings to establish and maintain. Theaters want to be assured that the movies they show will attract an audience. They want a professional movie. This requires cameras, technicians, stagehands, extras, and a whole infrastructure that simply is not required to write Finnegan's Wake. They want professional actors, preferably famous, so that you don't need to spend any effort convincing people you have good actors. They want a professional marketing campaign, so that no matter how lousy the movie is, people will see it before word-of-mouth sinks it. By the way, did I mention money buys noise?

The businesses of advertising is almost wholly based on this. Word of mouth is the gold-standard form of buzz, because by definition it comes from known, trusted sources. The hyperlink structure of the web is modeled on academic citation, which is just a more rigorous version of word of mouth. In recent years, as advertising audiences have become more sophisticated, cynical, or perhaps just overstimulated, advertisers have gone to greater and greater lengths to inspire trust (hint: you should not trust them). Of course, people know you shouldn't trust advertisers, which has inspired disguised advertisements or "guerrilla marketing," the goal of which is to inject advertisement into a normally trusted or at least independent source -- paid bloggers, press releases disguised as journalism. This tactic is also favored by spammers and computer viruses, e.g., sending the virus to everyone in your e-mail address book. The citizens of the Athenian democracy knew it as the "Trojan horse" tactic. Taken to its logical conclusion, "guerrilla marketing" eventually produces the total breakdown of society as we know it -- by gradually voiding every source of information beyond your immediate social circle. Hooray?

Okay, that's a little far-fetched, because we have things like the Better Business Bureau and people are still furious when you violate their trust. And maybe I'm being a little hasty in autamatically categorizing advertising as "noise" (shameless falsehood, if you prefer). Advertising can contain information: the name of the product, various facts about it. You may be interested in the product. But you probably aren't -- and you can't trust those facts without independent verification (information that is part-right wastes far more time than bald-faced lies). At least you shouldn't trust them, because businesses are sociopaths.

Not that I think the average businessperson is a sociopath -- possibly a greater percentage than the human average, due to selection bias -- but large, publicly-owned companies (the kind best positioned to put money into advertising) are large, complicated entities with highly abstracted chains of responsibility. Most include in their charter a clause which promises to "maximize shareholder value." This encourages company executives to put stock price above all other concerns, for fear they may be sued otherwise. If a corporation is to be considered legally a person, it must be conceded that rarely is anyone in position to act as a conscience. Now, corporations are not evil: they are very specifically sociopaths -- for my purposes, this means having no empathy but stil having ability to feign empathy if it provides advantage. In a system that does not actively prevent it, sociopathic businesses will be the most successful, as they have no moral restrictions to their actions.

Now, suppose you are a sociopath in a position of power, and that your products occupy a dominant position in the marketplace. Will you focus on providing accurate information to consumers? Rationally, you should attempt to build arbitrary brand loyalty and discourage the use or even awareness of your competitors. You do this by, respectively, exaggerating differences in your product (to prevent commodification) and disparaging all attributes of your competitor. Both of these are noisy. In fact, it is in your interest to generally increase the amount of noise in all channels. As the established player, you are in favor of the status quo, so any free flow of accurate information has potentially negative consequences on your market share. So, for example, by taking control of one channel of information and spreading doubt and uncertainty about other channels, you can easily make it far too difficult for people to bother.

As examples, which just occurred to me in force, I present to you A) the demonization of bloggers by the mainstream media, and B) the demonization of the mainstream media by Republican talk radio types.

Epiphanies aside, this should demonstrate that long-tail theory is not necessarily a fact of human existence -- it's a fact of a capitalist economy with significant information inefficiencies. At the very least, this should have a severe effect on the rate of change in the "hump" of the popularity curve, which in turn would flatten the long-term shape of the curve.

Democratic systems on the Internet are almost certainly more homogeneous than anarchical/networked systems, but we can't say by how much unless we have a way to remove external stabilizing/noise-creating forces from the system.

One more quick thought: for the general benefit, it is necessary to construct systems of laws that do not advantage sociopaths. To the extent that a system allows money to be accumulated by sociopaths, and moreso to the extent that the system allows money to be converted into power, that system is primed for failure (not collapse, necessarily, but failure to "promote the general welfare," as the Constitution puts it). This suggests, at minimum, that all political campaigns should be publicly financed and that there should be strict oversight of political advertisement. On the more general point of "noise-pollution" in information channels, the solution is more difficult, but should involve tighter truth-in-advertising laws, a rigorous and even-handed fourth estate, and most importantly a cultural movement towards intellectual curiosity, critical thinking, and the absolute refusal to tolerate being lied to.