3.03.2010

Crowdsourcing: Art vs Model

i've been reading about it a lot lately on blogs, on agency sites. it's all across twitter and bouncing around the webs through other forms of chatter. and, recently, more than a few agencies have popped-up, citing crowdsourcing as their sole motivation and business model.

this is a problem... for a lot of people and for a lot of reasons.

let's start a little over a year ago.

aaron koblin and his friend takashi kawashima came up with an interesting idea: create a piece of digital artwork that was made by the collective efforts of 10,000 double-blind and anonymous participants rather than the artists themselves. the overall philosophy isn't novel. it's derived from a movement among industrial artists who chose to remove themselves entirely from the production process of their work. they would draw up the plans and send them to a factory to fabricate and assemble the work. days, weeks, months later a complete piece of artwork was delivered as ordered to the show. the artist didn't even show up to direct the hanging of his piece. he would leave assembly instructions, likely in the box. the point behind it all was to reduce art to a process, not a person.

koblin and kawashima succeeded magnificently.

they managed to get approximately 10,000 people (some people participated more than once) to collectively draw a one-hundred dollar bill, pixel by pixel, and then assemble it without even knowing what they were making. this process is itself an adaptation of a computational method-made-human called a mechanical turk. k + k write: "the project explores the circumstances we live in, a new and uncharted combination of digital labor markets, 'crowdsourcing,' 'virtual economies,' and digital reproduction." a new, more sophisticated twist on an interesting philosophy, indeed: art without the need of an artist... now made mass via technology.

it worked for a few different reasons:
1) k + k were making art. the great thing about art is that, for the most part, it serves itself and nothing else. the purpose behind its design is usually manifested as much in its creation and existence as it is in its conception. and so, the actions of the participants were both the means and the end.

2) the compensation matched the contribution. each artist had to digitally recreate a pixel. a single pixel... and a rather simple pixel at that. for that effort, each new pixel rewarded its creator with $0.01. one pixel, one penny. that's about as fair as it gets.

3) the process was double-blind. the participants had no idea what they were making or who else was making it with them. that keeps everything honest: the artists and the art. no one's trying to outdo anyone else. there's not a single winner because there's no contest. and the whole thing flies below the radar until well after its completion. there's no undue influence, incentive, or awareness anywhere in the process.

so, the verdict in this case appears to be that crowdsourcing can work; and, it can work pretty well.
and a lot of people noticed.

as a result, the premise of crowdsourcing quickly crept into a lot of the digital tools and applications we use. wikipedia might be the definitive monument to crowdsourced behavior: an entire encyclopedic catalogue sourced for free from users all over the world. many gps systems crowdsource their traffic data from the actual whereabouts and flow of their users in transit. google has been mining people-powered computing for a while, too: we can rank and respond to the accuracy of search results. and now, google is using humans to translate mottled text in books via captcha.

i'm the first to say that all of these services are great. but i'll also point out that an important aspect to these models is that the input and the output are on equal footing: everyone can participate and everyone can benefit. wikipedia is free and captcha translation will wind up powering google scholar, which is also free. and, where the gps systems are concerned, those paying for it are the same ones powering it.

true, at some point the paradigm narrows. not everyone actually winds up participating via contribution. but think of the model as an hourglass shape turned on its side. the narrow point in the middle is the moment of action. it's narrow because, naturally, not everyone will participate. but the wide part on the left represents the fact that everyone can participate. and the wide part on the right indicates that everyone has access to the output. equal footing.


and then there's that value proposition again: free. it doesn't get more equitable than that. (again, with the gps example, it's definitely not free, but everyone who winds up paying to use it gets the same amount of access.)

now, something funny starts to happen. agencies and brands get wind of this thing... crowdsourcing. and it quickly transforms from a method driven by meaning to an exploitable premise; it gets reduced to its pieces and not its purpose. they see "cheap," "content," "volume," "audience," and "engagement." coincidentally, these happen to be all those happy metrics that make a great campaign on paper. but what they're missing is the reason... that really inspirational piece of artwork or that indispensable digital service that makes a point simply through its existence.

so then what do we get?

we get a commercial, a promotion, and sometimes... a product.

it works like this:
brand x spends a lot of money upfront to generate awareness for a contest. this brand advertises across media vehicles, platforms, and silos: "participate and you may be entered to win your very own superbowl commercial." then, as awareness builds, entries are submitted via the purpose-built promotional site. the entries keep coming. perhaps there is a gallery to view said entries. finally, after the deadline is announced, a winner is chosen. at long last, a crowdsourced spot airs during the superbowl for 30 seconds. exeunt stage left.

the problem is that this dynamic upsets the value proposition put forth earlier on; in this case it is in fact completely lopsided. users are asked to participate and provide input with no guarantee of an upside: i.e. you get paid (k + k), or you get access (wikipedia, google). this creates two issues:
1) with little-to-no reward, the value of the incoming work--for the most part--matches that of the potential for an upside: paltry.
2) since the lackluster--or perhaps ephemeral--reward has a likewise effect on the output, its ultimate longevity is also significantly diminished.

what's more is that this lopsided dynamic not only influences the perceived value of the output, but simultaneously impacts its overall benefit. in these types of situations a whole lot of people are persuaded to pitch-in when almost no one really winds up benefitting. well, to be honest, a few people/entities do see some positive results. the brand that orchestrated the contest receives a low-cost piece of content for their commercial, and the contest-winnter(s) gets his (their) 15 minutes of fame. other than those two parties, i can't be too sure that anyone else really benefits from a cleverly crowdsourced commercial.

so what we end up with is an ad-like object possessing low perceived value that has little staying power and only benefits a few people.
i'd say that constitutes a pretty shitty paradigm... for everyone involved.



it's little wonder then that we don't see the output of these contests sticking around for very long or brands dipping into the crowdsourcing well over and over again.
this type of crowdsourcing system simply can't sustain a matched correlation between volume, quality, and benefit for any significant period of time.

there's one model left: the crowdsourced agency.
and i personally believe that this scenario has the potential to be one of the most dangerous crowdsourcing variants. dangerous bad, not dangerous awesome.
the reason being fourfold:
1) it over-commoditizes an already threateningly commoditized market
2) it transforms the agency into a broker instead of an incubator
3) it makes us price ourselves out of the market
3) it once again upsets the hourglass dynamic by creating imbalance: a lot participate to benefit a few.

these days, agencies seem to be wading through particularly rough waters when it comes to justifying their need + cost to clients. new and different agency models are popping up right and left in order to try and solve the issues plaguing our industry: agencies need to be more nimble, some clients only need compartmentalized competencies, too big = too much overhead, does your agency get digital?, retainer payments versus time sheets... all of these are different ways of answering the overwhelming concern of clients that agencies costs are disproportionate to their benefits. so if an agency begins to crowdsource its solutions, it begs the question as to why an agency is even needed in the first place. suddenly, we've cut ourselves out of the picture entirely. as @eyecube points out, "If you’re going to crowdsource, why does the client even need a middle man...?"

at best, however, a crowdsourced agency is merely a broker. and, for the most part, all that a broker offers is access in exchange for a reasonable rate of return. in this form the actual agency (philosophically speaking) of an agency is entirely negated: it has no need for a strategic department, a creative department, or an account department. all it needs are a few well-connected former department heads that are social media savvy; people who know what good work looks like so when it lands on their desk they can pass it along.

the benefit of an agency like this is that it is MUCH cheaper than any of its competitors. but that may also be its downside. it all has to do with an economy of scale. this is the clever trick that crowdsourcing has pulled; it makes the cost of solving a problem cheaper by efficiently expanding the audience of those who can solve it. suddenly, you've engaged one million people to try and solve a ten thousand dollar problem. on average, that breaks down to one penny per person. now we've reset client expectations with a slippery slope... an insanely cheap one at that. why would a client pay a freelance rate for a senior creative when he can pay a fraction of the price for ten times as many people? why gamble with the cost and know-how of a small group of people with narrow, but finely-honed skills (an agency) when you can pay one person to manage the output of thousands? what we've done is essentially price ourselves out of the market while simultaneously create demand for a widening net of problem solvers at a cheaper rate.

finally, the crowdsourced agency model creates an unbalanced system. i'll try to be quick with this one since it has been laid out earlier. essentially, a crowdsourced agency asks for work from a lot of people, and the benefit of that work only gets passed on to a few people/parties: the agency, the brand, and the person(s) whose work was selected. that in turn creates some advertising. and, well, ads don't normally benefit too many people... and if they do, not for very long.


granted, all of this can be turned on its head. an agency can contract one million people to make one million tiny pieces of something amazing and long-lasting, something with purpose and meaning that serves a need and strikes a chord. it's not impossible, and it requires a perfect storm of agency vision/leadership combined with a the most amazing, brilliant, and trusting client. it could happen. but could it happen every single time? day in and day out? unlikely but i've got my fingers crossed.

(i've got a few thoughts on how/why/when a crowdsourced model can actually work, but i'll save that for another ridiculously long post)