Distributed decision-making

I spent a chunk of this afternoon at Bush

in 30 Seconds, a website from the people who brought you moveon.org.

The purpose of the website is to find a 30-second ad which can then be run in

Bush’s State of the Union speech. Between November 24 and December 5, anybody

could make their own ad, save it as a QuickTime file, and submit it to Bush

in 30 Seconds for consideration.

The organisers expected about 300 entries; in the end, they got over 1,000.

And even if there are lots of eager volunteers, it’s hard to approach 1,000

Bush-bashing ads and keep a fresh and open mind with respect to each one. Many

were excellent, but choosing between them was going to be very arduous.

So they didn’t. Instead, they basically decided to use the same kind of web-based

system that was initially popularised by Hot

Or Not. Once you register at the website, you’re shown a sequence of no

more than 20 ads, chosen at random. (Of course, you can vote on just one or

two if you like.) You then rate each ad on a range of criteria, ending up with

an overall grade. Even if each person only votes for a handful of ads, pretty

quickly the total number of votes will add up, and, with luck, a handful of

spots will emerge as the clear favourites.

I looked at 20 ads in total, and some of them were truly appalling. Others

were very good, however, like Bushopoly,

featuring a Monopoly set; If

the Bush Administration Was Your Roommate (pretty self-explanatory, but

well executed); and a wonderful little spot called Bush

Doesn’t Tip, featuring his former beer vendor at the ballpark in Arlington,

back when Bush was an owner of the Texas Rangers.

Three out of 20 is a pretty good hit rate, I think, and I can only imagine

how good some of the other ads are that I haven’t seen. I guess I’ll find out

when the 15 finalists are announced. More importantly, I have a lot of faith

in this process: it’s a lot more reliable than shutting a bunch of people in

a room with junk food and asking them to choose between hundreds of different

entries all of which start blurring into each other after a while. Moveon.org

isn’t being particularly innovative here: TriggerStreet.com

has been doing a similar thing for the past two years with screenplays, cutting

out the Hollywood bullshit to try and find the very best product.

Now, Jeff Jarvis seems to be proposing

something very similar for the World Trade Center memorial competition.

Don’t trust a small jury to find the best submissions, he says: "Viewing

5,201 entries is a daunting prospect. But by the time Web viewers get finished,

they’ll have whittled that to, oh, a few dozen."

But a Hot or Not / TriggerStreet model wouldn’t work in this situation. When

people are competing on the quality of their television ads or their own individual

pulchritude, a popular vote is a good way to measure quality. The ability to

read, understand and judge an entry for the WTC memorial, however, is far more

difficult, and in any case there are often very good reasons to dismiss memorials

which might look good at first glance.

Jarvis, in fact, doesn’t propose any kind of randomising device which would

ensure that all the 5,201 entries got a reasonable amount of scrutiny. Just

put them all up on the web, he says, and we’ll do the rest. But we won’t. His

own entry, and those of a few other people with strong web presences or other

brand names, would get quite a lot of discussion. And the vast majority of web

browsers, not wanting to jump into such a huge pool at random, would seek out

guides to the more interesting designs – guides which, pretty much by

definition, would not have been written by people who’ve actually gone through

all the entries individually. Take a design without significant traffic being

driven to it: the chances of its being discovered and acclaimed are actually

pretty thin.

That said, putting all the entries online is not a bad idea. The simple act

of doing this could, as Jarvis, says, be positive:

What it does is open up the process, allow all of us to feel involved and

to help point to those designs that touch us and speak to us. There are bound

to be surprises there.

In addition, this also meets the jury’s fine goal of displaying all the proposals

as a memorial in and of itself. The heart and soul that went into those 5,201

entries will be, I guarantee you, inspiring.

And if it should happen that one or two of the entries do start getting a lot

of popular support, then at the very least the LMDC can start wondering why

that is, and whether certain elements might not be incorporated in the final


In general, though, the concept of using the internet to whittle down an unwieldy

number of entries to something more amenable to straightforward "which

of these is the best" comparison is surely an idea which is only taking

off. Imagine if we could have had something similar in the California gubernatorial

race: rather than everyone simply voting for Arnold because he had the name

recognition and the momentum, all those tiny individuals might actually have

been in with a chance. (Of course, you’d need to implement something like Single

Transferable Vote or instant-runoff voting in order to make this worthwhile.)

And if it worked in California, imagine what could happen in a presidential

race! Of course, if the WTC memorial is too controversial for such tools, then

a political election is certainly beyond the pale. But just imagine… maybe,

some day, there will be a way for voters to rank a large number of candidates

based on something other than campaign money and name recognition. Then, a very

minor candidate could win an election just by being ranked in the top 10 of

most voters’ lists. Choire Sicha for


This entry was posted in Politics. Bookmark the permalink.

4 Responses to Distributed decision-making

  1. Jonas Cord says:

    I don’t think Jeff is implying that the online popularity of memorial designs should determine what one gets built – rather, I think it’s just a way to narrow the entries down from over 5,000 to something people could conceiveably consider.

  2. greg.org says:

    A lot of good points, FS. I took Jarvis’s idea to be more about the formation of links and the discovery process than about the popularity contest aspect.

    Like how you pointed out the three moveon.org ads you saw that you liked, which I might never have been exposed to when I rate them.

    The picks and finds of webcelebs would certainly get more attention than the picks of others, but since the 5201 entries themselves all go up at once, there’s not an automatic kottke or metafilter among them. Of course, numbers 1-20 will be seen by far more people than those in the middle…

  3. Brian McConnell says:

    I submitted a design in the WTC memorial competition, and agree with Jarvis’ basic idea. While deciding what was built based on a popular vote would have been a bad idea, a well designed system for exploring and rating the 5,201 proposals would have been useful in identifying common themes that appeal to the public.

    With this information, the jury could have focused their attention on the proposals that would be likely to appeal to the public, rather than guess what the public _might_ find appealing.

    The decision would have still been up to the jury, but at least they would have had some idea what the public was expecting from the memorial.

  4. Are you want to buy uggs?There are many good ugg boots :












Comments are closed.