Skip to main content

Crowdsourcing Curation: The Social Graph as Gatekeeper

I've written before about the compromise we tacitly agree to when amateurs take over the roles formerly held by professionsals. The Internet promotes this takeover by lowering the cost of production and transmission to near zero for nearly every user, for everything from words (blogs) to pictures (Flickr) to video (YouTube).

As Clay Shirky put it so well: As freedom to produce increases, average quality necessarily goes down. For example: Thanks to Flickr, we now have access to a mind-boggling array of beautiful pictures, but that's partly because we simply have access to a mind boggling array of pictures, period. Some of these, of course, are beautiful; but there are a lot more of Aunt Bettie's 43rd picture of a bundt cake than of an Annie Leibovitz Rolling Stone cover.

It is at this point that many people interject: "This is the problem with the internet! It's full of crap!" Many would argue that without professional producers, editors, publishers, and the natural scarcity that we became accustomed to, there's a flood of low-quality material that we can't possible sift through on our own. From blogs to music to software to journalism, one of the biggest fears of the established order is how to handle the oncoming glut of mediocrity. Who shall tell us The Good from The Bad?  "We need gatekeepers, and they need to be paid!"

All of this is true, to an extent. We do need ways to filter and discover content. And just because we can produce something and transmit it, doesn't mean that it's worth consuming. Luckily, the Internet not only gave us the means to produce and transmit on our own, but to curate as well. We do it every time we e-mail, share, "like", tweet, or buzz a link. For example...

How did you find this article that you're reading right now? I'm pretty sure it hasn't been published in the New York Times. No professional editor or publisher made a determination for you as to this article's quality, aside from the author himself. Did a friend send you a link? Did you see it posted to Facebook, Twitter, Buzz? The Internet has enabled us to build our social graph, and in turn, that social graph acts as an aggregate gatekeeper. The better that these systems for crowdsourcing the curation of content become, the more accurate the results will be.

This social-graph-as-curation is still relatively new, even by Internet standards. However, with tools like Buzz and Digg 4 (which allows you to see the aggregate ratings for content based on your social graph, and not the whole wide world) this technique is catching up to human publishers fast. For those areas where we don't have strong social ties, we can count on reputation systems to help us "rate the raters". These systems allow strangers to rate each other's content, giving users some idea of who to trust, without having to know them personally. Yelp has a fairly mature reputation system, where locations are rated by users, but the users are rated, in turn, by each other.

Reputation systems and the social graph allow us to crowdsource curation. I'm not ready to argue that these systems are up to replacing individual human curation (yet), but they're getting better every day, and I think they are well on their way.

A good example of manual curation vs. crowdsourced curation is the competing app markets on the Apple iPhone and Google Android phone operating systems. Apple fans complain that the Android marketplace has too many low-quality apps for any given task. They complain that it's hard to find an "official" or "sanctioned" app. On the other hand, Android fans criticise Apple for limiting their choices. They don't want to be beholden to the whims of a select few.

Apple is a monarchy, albeit with a wise and benevolent king. Android is burdgeoning democracy, inefficient and messy, but free.

Apple is the last, best example of the Industrial Age and its top-down, mass market/mass production paradigm. They deal with the big head of the curve, and eschew the long tail. They manufacture cool. They rely on "consumers", and they protect those consumers from too many choices by selecting what is worthy, and what is not.

Google Android is building itself as a platform for bottom-up innovation. Their marketplace publishes first, filters second, utilizing little more than the rankings of the community. They release tools like the App Inventor to spur creation by amateurs. They can do this because they rely on their audience to curate for itself. This process makes it messy (for now), but imagine a time when the ratings in the Android Marketplace aren't just from the community at large, but from your tailored social graph?

Consider for a moment that the Internet will only continue to increase our freedom and ability to produce content, whether they are smartphone apps, videos, music, pictures, or words. Can you imagine if Apple had to approve your videos for posting on Youtube, where every minute, 24 hours of footage are uploaded? There's no way humans could keep up! The traditional forms of curation and gatekeeping simply can not scale to meet the increase in production and transmission that the Internet allows. 

Crowdsourcing is the only curatorial/editorial mechanism that can scale to match the increased ability to produce that the Internet has given us. As the former "consumers" become "producers", we're going to see better and better implementations of reputation systems, and better integration with our social graph, because they are the only mechanisms that are feasible, but also because we just love to share. Twentieth century mechanisms for curating/editing are built for a top-down, mass-production age, and they can't keep up with us, the former audience, as we make the leap into production.


  1. Excellent post. Based on what I've read (don't own an iPhone), Apple's excellence on the apps is overrated, BTW. Because the process of responding to user community feedback is so centralized and bureaucratic, the correction of bugs tends to be extremely slow.

    Cf. the argument that the most effective way of aggregating and using counter-terror intelligence would be to remove security classification and open everything up so it could be acted on by the rules of Raymond's Bazaar. The enormous amount of intelligence collected is unusable because nobody's in a position to even know what's there or put individual items together. Even the people with the top clearances are forced to sit in a locked room and forbidden to take notes.

  2. I should've added this little gem:

    "Whether this revolution in the reading habits of the American public means that we are being inundated by a flood of trash which will debase farther the popular taste, or that we shall now have available cheap editions of an ever-increasing list of classics, is a question of basic importance to our social and cultural development."

    --Harvey Swados, 1951, re: the paperback.

    via Clay Shirky, _Cognitive Surplus_

  3. 9 months later, and this post is still relevant! thanks, its thought provoking. Ruti

  4. Thanks Ruti! Glad you found the post useful.


Post a Comment

Popular posts from this blog

The Re-Opening Experiment

We should remind ourselves that, this Memorial Day weekend and the weeks that follow, we are subjects in a grand experiment to see how good we are at social distancing as stay-at-home orders are being slowly lifted. The state's stay-at-home order was never meant to keep you, individually, safe from infection. It was meant to keep hospital's safe from being overwhelmed by too many of us needing them at the same time. In Michigan, the daily new cases of COVID-19 are higher today than they were when we locked down in late March. We are testing whether or not we can open up (with all of our new precautions and protocols) without spiking the rate of spread, but make no mistake: it *is* an experiment, and we *are* the test subjects. Please don't get careless as things start to open up. We need to get our economies back on track, but we are still a long way (and a vaccine away) from being out of the woods. Stay vigilant, folks. Wash your hands. Wear a mask. As has always been the

VMWorld Wednesday

Today I noticed three things: 1) All the good sessions ran today. 2) Lines for everything! 3) You can't do back-to-back sessions all day without burning out. Today's sessions were not to be missed, and everybody knew it because lines starting forming 45 minutes before some sessions. VMWorld has been on their toes, however: I didn't miss any session that I wanted to hit, and the most popular sessions from Monday and Tuesday got added back to the schedule on Wednesday and Thursday so everyone would have a crack at them. This is some very nimble work for a conference this big. Well done, VMWorld! Here's the photolog: My morning run takes me down to ferry building and up the Embarcadero. Here's the view at sunrise. This lovely scene is the hallway in my hotel. Creepy, but swank! Lines! Today was the day of lines! This was the line first thing in the AM for the Labs. More sidewalk art outside of Moscone South. Bean-bag Alley - where people and devic

A Christmas Present

What more could I ask for? Wine. Innovation. A blow against government over-regulation. A story about a penniless Yugoslavian immigrant. Capitalism. And whooping some French ass. All from Merry Christmas to all, and to all a good night.