In ur filters, fixing ur tubes

1 December 2008

After being outed today as a posterboy for Andrew Orlowski’s “League of the Clueless” I had an interesting and sobering conversation with a colleague in the music business. In his particular genre (drum & bass), the explosion of cheaper (or pirated) tools to create music, the availability of MySpace and similar services allowing artists to bypass their labels, has led to an end to the market being the mediator of quality. In a world where anyone is a publisher and with mere clicks & links enough to make anyone an Internet star, the result is that (as he put it) “a wave of shit” dominating the drum & bass scene.

Let’s make clear that before the Internet came along and ruined it, the music industry was far from rosy. The market is not a pure mediator of quality, in the music business, and it of course has been undermined by the vast and seemingly unstoppable amount of illegal downloading) but has the absence of market mediation and the growth of online really ruined things?

The radical lowering of transactional costs associated with the rise of the Internet has been elementary in “Organizing Without Organizations” (as related very well by Clay Shirky in his book Here Comes Everybody), creating collaborative efforts and new forms of media as worthy and high-minded as Wikipedia to the more lighthearted lolcats. There is now a vast amount of user-generated content – although the really important thing is not that it is generated by ordinary people, but that it is mediated by ordinary people, through the interactions they make with it.

People’s interactions with content can be minor or major; for the Wikipedia article it can be as simple as correcting punctuation to revamping an entire article. For a YouTube video, it can be as simple as viewing, a bit more in rating or favouriting, further more so by commenting or forwarding to a friend or blogging about it. Positive feedback like this can lead to steadily self-reinforcing loops – a few people see it, it’s picked up by a blog or other site, so more people see it, which means some more blogs pick it up, which means more people see it, etc. This positive feedback is the cornerstone of how the popularity of free (as in beer) content is mediated online (and I’ve been the beneficiary of this very effect quite recently).

But just as markets can become faulty or unbalanced, so can these positive feedback loops. It’s very easy to spot herd mentalities and groupthink on community sites such as Digg. Digg has a typically North American, male, libertarian-dominated demographic (no coincidence that this also the traditional demographic of early adopters online). The result is a site dominated by Apple fanboyism, FAIL or lolcat memes, lists and mild misogyny. An awful lot of crap, to be honest.

The much-vaunted Long Tail is becoming tighter and more inequal, squeezed by these systemic biases and turning Web 2.0 into a less inclusive space. As these systemic biases are those of the early adopters, we face the risk of subsequent adopters turned off from the very start. This is the basis of the nightmarish exaggerations of the like of Andrew Orlowski and his fellow traveller Andrew Keen.

So much for a democratic web. Is this the natural consequence of what happens we rely user-mediated content, that what comes up is domination by the most vocal minorities? I don’t think so. First of all because we’re all still in a massive upheaval in how media are produced and distributed and this is not the end of it. We’re in the middle of it and don’t let anybody say they’re an expert or guru in such a changing and chaotic field (further exacerbated by an economic downturn). As we are only in the middle of a changing field, the status quo will not last long, especially as people from non-traditional online demographics move in.

Here’s a personal story, which will lead on to some suggestions. All too often I will see a link, click on it, and see that after a few seconds of reading or watching the content, it’s rubbish. I won’t get past scrolling the first page, or the first few seconds of video, I quit the tab and move onto something else. But how will anyone else know that? Interactive controls on a piece of social media are generally positive – it records a hit. If people like it, they will Digg it or favourite it. But there’s not much I can do to record a negative opinion. I can’t undo my hit on the page. I could leave a comment saying “this is shit” (or some such) but it will get ignored, or worse still, recorded as evidence it is being “talked about” and this is further evidence of its quality. There is often a “mark as objectionable” control but this is for copyright violations or inaccuracy, not lameness. The only negative feedback I can usually give is a low star rating – but when was the last time you decided looking at a video or blog post based predominantly on its rating?

Back to Shirky. In a really interesting talk at Web 2.0 expo about “information overload”, which he re-terms “filter failure”:

This isn?t a design problem, it?s a mental shift. It?s understanding we?ll have information overload just the way fish have water. A quote: ?when you?ve faced a problem long enough, maybe it?s not a problem, it?s a fact.? Some of this can be solved with programs: look at Digg, Google, etc. But a lot of it can?t. When you?re getting deluged with information the question you need to ask yourself is ?what filter broke??

In the case of some parts of the web, the filter for crap is broken. It’s broken because if we don’t like something we tend to ignore it – partly because it’s the easiest thing to do (why waste time marking down something as crap when you can just move on to the next awesome thing in near-zero time?) and partly because there is no capability to make your impact fully known. So the solution? Here’s three:

  1. We can come up with active systems to fight crap, which is the Wikipedia way – adopting ever-more Byzantine content policies and actively combatting systemic bias.
  2. We can rely on filters that combine with some financial or market-based control (such as Metafilter’s $5 membership fee).
  3. Better still – we come up with new, better filters that take their cue from our inaction rather than our action.

The first requires a lot more work than may make it one’s worth. The second may work but micropayments and subscription services have largely failed up till now except in a few rare cases. The third option is the most attractive from a purely technological point of view – but possible implementations such as monitoring whether you scroll down a page, or quit watching a video early, are intrusive and unacceptable if a third party is in control rather than the user. Squaring the circle of controlling you privacy, while making sure your inaction counts is the key to making that kind of filter work.

Of course, none of these three filters are mutually exclusive, and therein might lie the best solution – but there’s been comparatively little work done in the field yet, I believe.

One Response

Interesting. It’s something I’ve also been thinking about of late, and without wishing to go all pretentious on yo ass, it reminds me of John Rawls’s Theory Of Justice from my political philosophy class. Essentially he puts forward the premise that justice stems from fair systems, which sounds pretty obvious, but the opposite is found in a system like Digg. One of Digg’s greatest errors, in my opinion, was forming the system of social networks that it did. Its ‘friend’-based economy created a clique-based system which meant that people blindly gave Diggs to anyone on their list regardless of quality. Delicious dealt with this slightly better, insofar as one still has ‘friends’ but you have to actually devote some time and effort to relinking, whereas Digg makes it too easy. Alongside the ease and the social-network based warping, it’s cliqueyness has been amplified.

Anyway, I think this links in with your notion of faulty filters, but I would go so far as to suggest that regularly the problem is actually built into the system from the start. Your example of any comment, positive or negative being looked at as ‘buzz’ is prime example. What we need, as you say, is a fairer system in which people’s disappointment can be registered. I don’t know what that system would look like, and I think you’ve outlined the problems with a ‘this is pants’ button, or detecting inactivity (but I’d add to that that it has the problem of being easily gamed by overly-sensitive types as a means of ad-hoc censorship).

I’m going to stop there, or my response is going to be longer than your original post, but suffice to say you got me thinking – thanks!