I've been excited about the Since Sliced
Bread (SSB) idea contest ever since
EchoDitto's Michael Silberman
explained it to our ORGware development team at dinner the
week before it launched. The SEIU's innovative contest, which
ultimately put $200,000 on the line, was sure to demonstrate the power
of
engaging the wisdom of the crowd to find the best and the brightest of
the common-sense ideas out there.
Well. The Law of Unintended Outcomes has not been repealed,
and the devil remains in the details. More than 22,000 ideas were
submitted and a zillion comments written about
them. That's the good news and
the bad news. How could it be bad news? The crowd that shared its
wisdom with SEIU became so invested in the process that, like
SlashDotters discussing
Linux sockets APIs, they had strong
opinions about the system's design. Although clearly shell-shocked and
reluctant to mix it up with its disgruntled members, the SSB blog
acknowledged the problems by putting up a guest post by SSB member
Scott C. last Wednesday: Best
Revolt Since Sliced Bread.
This is incredibly exciting for us at the Open Resource Group,
because these are the issues we spend every day fretting about, as we
design a system that will handle a contest like this as reliably and
elegantly as your email client does its complex SMTP/POP3/MIME
Dance.
Today I'll review the reviews and finger the probable cause,
which I think of as a "filtering slope." Next time, I'll describe
ORGware's specific mechanisms for teasing the best thinking and passion out
of any size community so that the community's sponsor can spread their
cash around for maximum impact. In fact, a contest like this should
generate more cash than it costs. If the Dean campaign and Katrina
taught us anything, it's that people will stand in line to back up
their passion with hard cash.
It was supposed to be one of
the Service Employees International Union's leaps into bottom-up online
consensus building, but the community blowback at the Since Sliced
Bread project that broke out this week has all the
appearances of being
an online
revolt.
Since Sliced Bread is a $100,000 contest inviting people to send in
ideas to improve the lives of working people in America. As described
by SEIU on the site: The contest encouraged ordinary Americans, policy
experts and economists to enter fresh ideas on how to create the kinds
of jobs that allow people to raise families, obtain affordable health
insurance, pay for college and save for retirement."
Designed by the web technology firm EchoDitto, some of
whose employees
are former colleagues of mine from Howard Dean’s presidential
campaign, Since Sliced Bread's architecture appeared in many respects
rather open and bottom-up oriented. Anybody could send in his or her
proposals. Visitors were encouraged to participate in the community
blog . . .
. . . But only a few days into the voting process, things started going
terribly pear shaped. In a nutshell, the big contention is that the
judges picked a bunch of rather unfresh, and tame ideas.
"NO VOTE FROM ME! All
these ideas suck. I wouldn't pay $5 for any of
them. What a waste of time," went one commenter's response. "I too am
very disappointed in the lack of originality and diversity in the final
selections," wrote another. "Three selections out of 21 involve
national health care, which may be a great idea, but is hardly
original."
These comments came in
response to Andy Stern's call for appreciation of the ideas that were
chosen after the initial blowback: "I confess -- I'm a bit surprised at
the hostility meeting the 21 ideas announced yesterday morning," he
wrote. "Let's take a minute to appreciate the work of the 21 people who
are finalists -- they are amazing ideas that deserve discussion and
consideration."
"If this is what is considered 'amazing' we really, really have sunken
to new depths," one commenter responded. Hundreds of commenters offered
similar sentiments of disappointment at the 21 finalists.. . .
. . . The debacle
illustrates what can happen when a top-down organization like a major
labor union tries a "web 2.0" approach without fully preparing itself
for all the implications of empowering a network or community; that the
participants in fact expect to have power. But how could the union have
avoided this outcome? There are few examples to guide how to get
something like this right, much less offer a definition of what "right"
is. Trailblazing is a trial by fire process. And just because SEIU got
a little burned shouldn't be cause for them to stop trying.
The Real Bun in the Oven
The best lesson that the SEIU can get out of this is that the
overarching interest of the SSB community is to finish creating itself
and to perpetuate the thrill of having a place to check in at night and
when the boss isn't looking. For most of SSB's members, this is their
first rush of impassioned online involvement, and the feel of the place
is just like Dean For America during the salad days 31 months ago. As
the contest draws to a close, their union hall is scheduled for
demolition. I suggest that abandonment is the subtext and underlying
grievance behind the explicit complaints.
As for those explicit complaints, we can blame them on the
"Filtering Slope," a term I just coined. In one fell swoop, the 22,000
ideas were narrowed to 70 by a "group of diverse experts,"
from which the official judges voted for the final 21 candidates, to
then be voted on by the members.
If SEIU actually wanted new,
worthwhile ideas it would have structured the contest differently.
There would have been much more filtering on the way up, much more
asked for than 175 words, and more structured opportunity for
collaboration. As some of the SSB commentators noted, giving the whole
$100,000 prize to one person leads people to hide thoughts, isolate
their creativity, and maximize the gain for themselves. Furthermore,
the judges may have been picked for their legitimacy to the community
in the abstract, not for their legitimacy as new ideas testers.
(The SEIU added two runner-up $50,000 prizes after the
original $100,000 announcement.)
SSB's Black Box
That 3,000:1 reduction gear, by which 22,000 entries became 70, is the
part that has the members upset. It's probably an artifact of the
system and is the weakest part of the design because it builds in a
potential scaling problem, and doesn't engage the wisdom of the crowd
to do the smart filtering, as Zephyr points out. I can't remember if
Mike Silberman forecast the response when we had dinner in mid
November, but I'm sure it was much greater than expected. Might the
questions look better to SSB's members if the "panel of experts" had
more time to sift through the questions? Perhaps. In any event, If
there had been "only" 2200 entries, the filtering would have been more
like a slope than a cliff.
Zephyr comes to the problem with real experience in this kind
of design work. She feels that SSB could have been framed better:
Unintended Consequences?
The money offered by SEIU to the best idea submitted creates a huge
opportunity -- $100,000 is enough for people to send in loads of
suggestions, so when the SEIU crows about what its doing, I’d
love to see comparisons of other $100,000 contests with low barriers to
entry. I can just feel the muscles in my brain start to stretch when I
think of 175 words I could write for the chance to win $100,000.
There’s something both charming and disturbing about bribing
people into citizenry.
But what strikes me most is that SEIU asked for something it didn't
actually want, if that makes sense -- what it wanted was more
membership and participation, fresh ways of expressing new ideas, fresh
variations on old ideas, so it asked for something it thought would
generate that.
Or at least that's my read on it. If the SEIU actually was trying to
come up with new ideas to solve something, it would have acted like
people act when they are trying to solve something –
“Here’s what
I’ve been thinking (about, say, the questions of hidden
tariffs and free trade and labor and isolationism) and here’s
where I keep getting stuck, and I’m not sure I’m
thinking about it the right way.”
Or, if the problem the SEIU was
trying to solve was find out what’s important to people--
“Look, we’ve
done polls, and while these are helpful they tell us only so much
– we’ve done some deliberative polling too but
that’s too expensive – this project is a first
round in our effort to find out what people dream about when they are
asked seriously, not just what they say when they are answering the
phone a 8 pm on a Sunday night. We’ll do more rounds, making
sure it’s more inclusive. This is a collective
brainstorm.”
Taking it from the Top
Zephyr starts her post by describing one person's
project to solve the impasse. Brian
S. Julin has built a balloting system
to harness the judgment of the crowd to sift through all the
suggestions by comparing them. Each ballot contains 10 of the 22,000
SSB ideas, each with its own row of 10 radio buttons. You're asked to
rate each idea from 0 to 10, like this, after following each idea's
link:
Done this way, Brian believes that 7040 ballots will catch
most of the ideas, though clearly not most of the entries. He's the
first to say that it's an imperfect system with sharp edges, but he
demonstrates how a community like this will seek ways to improve its
systems to protect itself and grow.
It's Deja Viewed All Over Again
Brian's 10 radio buttons per idea is a lot like my issues voting
proposal from November, 2003. But in a general sense, Brian's
fix is SO
reminiscent of the ad hoc systems development at the Dean campaign.
Every day, we'd realize that there was something that should be
improved and we'd try to come up with a way to deal with it. Generally,
it was impossible to do so within the constraints of a campaign. I
experienced a great deal of personal frustration because I was happy to
try to help with workarounds, though I wasn't qualified to do so. It
was just that there was no one else to volunteer. As both my readers
know, I have a history of volunteering for low-probability missions.
And that's why I call myself a former Senior Advisor to the
Dean Campaign. I was older than everybody there and no one took my
advice. My only real claim is that I did put in the hours and made some
useful introductions.
ORGware, meet White Horse
I'll use this as a case study next time to describe how our
decreasingly vaporous ORGware solves SSB's problem by hitting the Install button.
