I wanted to wait ‘til the February 3rd polls opened to post this, because I wanted it to be a post-mortem and not a vivisection. What follows is a long musing on the Dean campaign’s use of internet tools, but it has a short thesis: the hard thing to explain is not how the Dean campaign blew such a huge lead, but rather why we ever thought that lead actually existed. Dean’s campaign didn’t just fail, it dissolved on contact with reality.
The answer, I think, is that we talked ourselves, but not the voters, into believing. And I think the way the campaign was organized helped inflate and sustain that bubble of belief, right up to the moment that the voters arrived.
Take this as an early entry in a conversation everyone who was watching Dean’s use of the internet should contribute to: what went right? what went wrong? and what to do differently next time? We should do this now because ‘next time’ still includes a passel of primaries and then, most importantly, the general election. If we have the conversation now, we won’t have to wait til the few uncontested House races of 2006 to see if we learned anything.
Two caveats at the beginning: first, the stupidest thing I’ve said on this issue was in Dean: (Re)stating the Obvious:
“…the most salient characteristic [of the campaign] is the style of engagement, including the use of social software.” Dumb.
The most salient fact of Dean’s campaign was Dean himself. Whatever conversation we have about the use of internet tools, Dean himself was the most important factor in the losses.
The analysis here is not concerned with the candidate, however, but with the campaign. So with additional caveat that a political campaign strategy is so absurdly multi-variate that certainty is impossible, those of us who care about the use of the internet in politics need to talk about how to use those tools better than the Dean campaign did.
Mental Models
Howard Dean had the best-funded, best-publicized bid to be the Democratic nominee; he was so widely understood to be in the lead that the inevitability of his victory was a broad topic of discussion. (Google “Howard Dean”+inevitable if you need independent confirmation.) Even the people disputing the posited inevitability burnished the idea; no one bother debunking the idea of, say, Kucinich’s inevitability.
I’ve had a hard time processing his Iowa and New Hampshire losses because I’d spent months hearing about how well he was going to do. It has taken me two weeks to decide that my mental model — how could such a successful campaign suddenly do so badly? — was the problem.
Dean’s campaign was never actually successful. It did many of the things successful campaigns do, of course — got press and raised money and excited people and even got potential voters to aver to campaign workers and pollsters that they would vote for him when the time came. When the time came, however, they didn’t. The campaign never succeeded at making Howard Dean the first choice of any group of voters he faced, and it seems unlikely to do so today.
If this thesis — call it the ‘mirage’ thesis — is too strong for you, consider its cousin, where the campaign was doing well until the last few days. In this version, one New Hampshire voter in three dumped Dean after no event more momentous than a third-place showing in Iowa (rarely known to track to New Hampshire elections) and a little hootin’ and hollerin’ in the concession speech (to use Sharpton’s memorable phrase). Not one Dean supporter in three, mind, one voter in three.
In this view, Dean’s support was real, but so thin and vulnerable that a mere political pin-prick was enough to cause the whole thing to collapse. Call this the ‘soap bubble’ thesis; the only difference between it and the mirage thesis is that in some other version of the election, if Dean had done everything perfectly, he could have performed well. I leave the likelihood of a primary race going perfectly as an exercise for the reader, but neither model suggests a campaign prepared for the real world.
Durham NH, December 9, 2003
We must take care not to re-write history so that we ‘always knew’ what was going to happen. I, like many, believed Dean would win the early primaries right up to the moment he cratered. To keep myself honest, I’ve spent some time reading Dean coverage pre-Iowa, and have come across an article that, if you want to understand the Dean campaign, should be required reading.
It is a Washington Post piece by Hanna Rosin called People-Powered (free registration required) and sub-titled “In New Hampshire, Howard Dean’s Campaign Has Energized Voters.” It’s about campaign volunteers canvassing for votes in New Hampshire in early December, and it’s not particularly broad or critical — rather it is an unintentional preview of the subsequent failure in that state, by presenting the techniques of the campaign in terms of motivational culture.
The piece is long, and almost every paragraph has something interesting in it. I can’t really do it justice here, but I’ll quote a couple of passages so you can get a feel for it:
The story goes on:
These passages, I want to emphasize, are from a piece that presents Dean as an all but certain winner, and yet the description now sends another message entirely: Less about Howard Dean than about James Moore, less like a political meeting than a support group, not so much about pushing Dean as they are about engaging people in conversation…
Several times Rosin comes right to the edge of predicting the events two months hence: those women were happy to have that nice young man come and listen to their stories, but in the end they weren’t going to vote for Howard Dean. She even provides a comparison to self-help books, and then finishes the sentence before she finishes the thought: self-help books sell well, but mostly don’t work.
The exit polls from New Hampshire were quite instructive. Kerry beat Dean in every demographic and psychographic group, including “voters under 30”, with but three exceptions: people for whom the Iraq war was the most important issue; people who labeled themselves not just liberal but very liberal; and people who’d made up their minds more than a month earlier.
So where did his 30 point lead go? Why would someone say they would vote for Dean if they weren’t actually sure? I believe that last category contains a clue as to Dean’s collapse in the polls – Dean lost among voters who waited til January to decide who to vote for, which is to say almost everybody.
Prior to January, “Howard Dean” was pronounced “Anybody but Bush.” The thing Dean did spectacularly right was to pick a fight with the President, a hugely polarizing and therefore energizing figure, on the issue most Democrats wanted to keep quiet about. Even if you’d been only been following politics casually, you would have known that Dean was the person who had most directly challenged Bush. For any Democrat whose primary motivation was not a bundle of particular policy proposals but the chance to send the current President home, Dean was the man of the hour.
In this view, the change in the poll numbers in January reflected not a transfer of votes from Dean to Kerry but rather from the general to the specific. Voter’s polled as to their choices last year were not bound by their answers, and nor had most of them bothered to sort out the candidates positions from one another. (My wife and I, both deeply interested in the primaries, couldn’t always remember all of their names.) A couple of weeks before the primaries, though, voters in those states started to have to make some real decisions, transferring their sense of “Anybody but Bush” to a specific Democratic candidate. And sometimes that candidate was Howard Dean. But mostly not.
How Did We Get Here? What Should We Do Next Time?
So how did this collective delusion of Dean’s front-runner-hood happen? And what if anything did the use of the internet contribute to it?
Here are a number of effects that I think led us to the false conclusion that Dean was, if not inevitable, than at least tipped to do very well. The bad news is that these effects, taken together, swamped what might have been a better-run campaign. The good news is that most of the effects are easy to recognize in retrospect, and therefore may be easy to defend against in the future.
– Novelty Campaign
The first and most obvious effect was novelty. Although there are already people running around claiming that the Dean campaign wasn’t really an internet campaign (on the grounds that such a campaign will, by definition, be successful), for those of us watching for the use of the net in politics, Dean was our guy, and we should remember that. He and Trippi and many of the Dean staff put the internet to the best, most vivid, and most imaginative use it has ever gotten in any national campaign.
That, of course, is a story in itself, and the press treated it as such. The NY Times, normally scrupulously balanced at election time (they gave front page coverage to a profile of Carol Mosely Braun) wrote article after article, including a cover story in the Sunday magazine, on how the Dean campaign had managed its “Come from nowhere” movement, the by now familiar story of MeetUp and MoveOn and internet donations and Dean weblogs.
The story took on the characteristics of a firestorm, where the original fire pulls more oxygen in, fueling the flames higher. The perception that Dean was first the strongest challenger, and then the frontrunner, was part cause and part effect for those stories.
We don’t need to worry about this in the future, I think. The press has a way of running fast epidemics, where an idea virus runs its course quickly, leaving everyone inoculated in its wake. The problem we will now have to watch for is where a candidate that makes innovative use of the internet will be cautioned about what happened to Dean.
– Support isn’t Votes
Other than this one-off effect, though, there are a number of more serious issues to contend with. The first of them is the difference between signs and the thing they signify. (Steven Johnson wrote about this over the weekend as well.)
Getting people together in the real world is hard – the coordination cost of any gathering runs into the inertia of modern life at every turn. (Robert Putnam in one sentence.) For many of us, the first time Dean appeared on our radar was when 300 people showed up for a Howard Dean MeetUp in New York City in early 2003. This was unprecedented, and Dean himself took note of it, coming down from Vermont to speak to his supporters.
We were right to be excited about this MeetUp, but wrong about the reason, because MeetUp was founded to lower the coordination costs of real world gatherings.
The size of the MeetUp in NYC was as much a testament to MeetUp as to Dean — it’s a wonderful tool for turning interest into attendance, but it created a false sense of broad enthusiasm. Prior to MeetUp, getting 300 people to turn out would have meant a huge and latent population of Dean supporters, but because MeetUp makes it easier to gather the faithful, it confused us into thinking that we were seeing an increase in Dean support, rather than a decrease in the hassle of organizing groups.
We’ve seen this sort of effect before, as when written correspondence on letterhead stopped being a sign of a solvent company, thanks to the desktop publishing revolution, or with the way email to politicians matters less than telegrams, because email is cheaper and easier to send. As we get the tools to make such gatherings easy, we need to concentrate on the outcome of those gatherings, rather than assuming strength simply by looking at the number of attendees.
– Fervor Isn’t Votes
Margaret Mead once said “Never doubt that a small group of thoughtful, committed people can change the world. Indeed, it is the only thing that ever has.” Generations of zealots have tacked these words up on various walls, never noticing that the two systems that run the modern world – markets and democracies — are working right precisely when they defeat these attempted hijackings by small groups.
Voting in particular is designed as a repudiation of Mead’s notion. In the line at the polling booth, the guy with the non-ironic trucker hat and nothing other than an instinct for who he trusts cancels the vote of the politics junkie who can tell you the name of Joe Lieberman’s Delaware field manager.
In Is Social Software Bad for the Dean Campaign?, I suggested that Dean had accidentally created a movement instead of a campaign. I still believe that, and this is one of the things I think falls out from that. It’s hard to understand, when you sense yourself to be one of Mead’s thoughtful and committed people, that someone who doesn’t even understand the issues can amble on down to the local elementary school and wipe out your vote, and its even harder to understand that the system is designed to work that way.
You can ring doorbells and carry signs and donate and stay up til 4 in the morning talking with fellow believers about the sorry state of politics today, and you still only get one vote. If you want more votes than that, you have to do the hardest, most humbling thing in the world. You have to change someone else’s mind.
Internet culture is talking culture, so we’re not used to this. In our current conversational spaces, whether mailing lists or bulletin boards or weblogs, the people who speak the loudest and the most frequently dominate the discussion.
Imagine if a mailing list had to issue a formal opinion on the issues discussed, and lurkers got a vote. The high-flow posters would complain that the lurkers votes would not reflect the actual discussion that took place, merely the aggregate opinions of the group, and yet that is how the primaries work. Talking loudest or most or even best means nothing.
– Effort Isn’t Votes
Here’s a catchy phrase: “Design, Create, Produce to Elect Governor Howard Dean for President.” That’s the slogan atop DeanMediaTeam.com; can you spot the error? (and we’ll let the fact that Dean is not currently Governor slide.) Designers don’t create votes, voters create votes, one per, and its votes, not stylesheets, that do the electing.
It is natural for a campaign, attracting so many eager young people, to oversell them on the effect they’ll have, when the truth is so rough – you’ll work 80 hour weeks while sleeping on someone’s sofa, and in the end, your heroic contribution will be a drop in the bucket of what’s needed.
So a little pep talk now and again can’t hurt. However, you can go too far, and too far is when people begin selling one another on the idea that their work for the campaign has a direct effect on getting their candidate in. Someone at Dean HQ should have sent out a bulletin to the staff – all effects are indirect. Money, bell ringing, blogging, advertising, none of it will get Dean elected unless it convinces voters to elect him. If you can’t point to ways your work is getting votes, you’re not helping. In particular, if you are preaching to the converted (see “Fervor Isn’t Votes”, above), you’re really not helping.
– Money Isn’t Votes
This is the big one. Dean’s internet strategy was a curiosity until there was money involved, and when it got involved, it got involved in a big way. We mustn’t forget how enormous a change this was – an upstart politician blew past all the favorites and even exited the public funding system because he got enough money soliciting donations from the internet, a few bucks at a time.
If none of the rest of it, the MeetUps and weblogs, had ever happened, but Dean’s campaign had still done this, its place in political history would be assured.
And yet money does not in fact buy votes. As candidates like Michael Huffington and Ron Lauder have shown, you can be very rich and still very lose. In Dean’s case, though, the effect was compounded by two other effects from above. By moving campaign donations online, they made it much easier to donate, so much easier in fact that raising millions from individuals was never the sign of strength we thought it was. (Support isn’t votes.) Like MeetUp, a lot of what the campaign achieved was by lowering the threshold to contributing, which helped create a false sense of strength.
The other effect was that last fall, when Dean announced his desire to top the fundraising list, a lot of us gave him money, self included, as a vote for that method of fund raising, without that meaning anything about whether we’d vote for him.
We were donating to the use of the internet as a tool, in other words – in the same way that the voters heard “Anybody But Bush” when Howard Dean was mentioned, a lot of us heard “Contribute to the use of the internet for politics” when the collection plate came around. (Novelty campaign.)
This won’t happen again – there will never be a second candidate to use the internet first. However, as with other shows of seeming strength, we need to be careful not to equate the advantages of online fund raising with electoral success in the future. When you change the game, the old rules are useless for figuring out who’s winning.
– Sometimes Votes Aren’t Even Votes (Depending on Who’s Counting)
None of the above needed to be fatal. Having a high profile, even if not for your policies; eager supporters, even if a little too confident; and good fund raising, even if it didn’t translate directly into votes, could have been a pretty good place to start, and after losing in Iowa and New Hampshire (which, let’s remember, Clinton also did), it could have provided a cushion to catch the fall, and to help bounce back as a more appealing candidate.
The moment for me, and I think for many of us, when we realized that Dean was sunk was on Wednesday after New Hampshire, when the press reported that he’d spent most of his $45 million war chest already. The obvious question, “How did he think he could do the rest of the campaign on a few million dollars?” has an obvious answer: “He thought he’d raise more, when Iowa and New Hampshire anointed him frontrunner.”
This was a fatal flaw in the campaign – they believed their own press. Dean was so out of touch that he had not prepared a concession speech in Iowa, a state where his third place finish was so bad that if he’d gotten every single Gephardt vote as well, he would still have been in third place, and would still have been double digits behind Kerry.
This is the question within the question. Out here, we had an excuse (albeit a flimsy one) for believing Dean was the frontrunner: it’s what we read in the papers. But campaigns don’t just use the pollsters, their field operations also keep their own numbers. And for Dean to blow all his cash and then not even prepare for anything other than victory means their internal numbers predicted certain victory as well.
Back to Hanna Rosin’s story in the Post:
But how does he know that? He is, by his own admission, a newcomer to this game. How does he know that a middle-aged New Hampshire lady is a “definite supporter?”
Consider his incentives – he is out proselytizing for a guy he’s dropped the rest of his life for, campaigning on stories that are more about him than Howard Dean, and getting a reaction from his listeners that is closer to revival tent than political kaffeeklatsch, and then there’s that daily quota of 180 to fill. What’s he gonna say? “None of them think Dean is any good”? “I tried my best, but I couldn’t convince them”? “They said they were going to vote Dean, but I think they were just being polite”?
Amateurs and zealots both have strong incentives not merely to misrepresent reality, but to actually misunderstand it. If you’re on a mission to change the world, you have an incentive to believe it’s changing.
Furthermore, if this is your first campaign, and you are doing it not because you want to be in politics but because you want to make the world a better place, what happens if you mark a bunch of people as “definite supporters” who really aren’t? Nothing, really. You have no professional reputation on the line – if you put your thumb on the scale, even if only unconsciously, no one will ever call you on it, because by the time the actual votes are in, no one will remember how many “definite supporters” you said you had.
For Dean to have spent everything on a momentum strategy so secure that he didn’t bother contemplating even second place, he must have heard from his field operatives that things were just great, super, you wouldn’t believe how many “definite supporters” we converted today.
It’s not clear what to do about this – skepticism is a hard virtue to put into software. Maybe a reputation system, or a market for accuracy, will convince eager amateurs to be careful about the difference between politeness and voting intent. But this is the hardest problem, because many of the other forces listed here make it easier to recruit and use the efforts of amateurs. Unless they face pressures that dissuade them from happy talk and mutual affirmation, however, they will end up convincing their candidate to misallocate precious or even irreplaceable resources.
Affinity Over Geography
A number of people, disputing the idea that the use of the internet had anything to do with the gap between Dean’s predicted and actual support, have advanced the “internet minority” thesis, as in “The internet is used by a minority of citizens”, or, in its more regionally biased version, “Who in Iowa has computers anyway?”
With national internet penetration at roughly two-thirds of households, it’s long since time to retire this canard. More people use the internet than read a daily newspaper. More people use the internet than vote in general elections, much less primaries. Iowa and New Hampshire both have better than 50% penetration (as does most of the country except the antebellum south.) Furthermore, one of the commonest uses of the internet is getting daily news. The internet is now, and from now on, a political media channel.
That’s a sideshow, however, compared to the internet as an organizational tool. The main effect of the internet on politics is as a lever, not a hammer.
Culture matters, and since the 1970’s, anyone who has looked at the cultural effects of the internet has picked the same key element: the victory of affinity over geography. The like-minded can now gather from all corners, and bask in the warmth of knowing you are not alone.
And yet as wonderful as this effect can be, it carries pitfalls. Liberal judges become more liberal on panels of all Democrats, as do conservatives with Republicans. Support groups can become maintenance groups by accidentally reinforcing the normalcy of deviant behaviors. (One of my former students presided over the removal of YM magazine’s “Health and Beauty Tips” BBS, because it had become a place for girls to swap tips on remaining anorexic by choice.)
Voting, though, is the victory of geography over affinity. Deaniacs in NYC could donate money and time, blogging like mad or tramping through the cold to talk to a handful of potential voters, but they couldn’t actually vote anywhere but NYC. Iowa was left up to the Iowans.
The easy thing to explain is why Dean lost – the voters didn’t like him. The hard thing to explain is why we (and why Dean himself) thought he’d win, and easily at that. The bubble of belief, which collapsed so quickly and so completely, was inflated by tools that made formerly hard things easy, tricking us into thinking that getting votes had become easy as well — we were all in Deanspace for a while there.
It was also inflated by our desire to see someone get it right, a fact that made us misunderstand the facts on the ground – we suffered the same temptations as the campaign workers to regard our fellow citizens as “definite supporters”, even when we ourselves were supporting a movement rather than a campaign.
It’s been a shock, but it doesn’t have to be a fatal one. Lowering coordination costs and making it easier for citizens to create media and distributing fundraising to the masses are all good things. This year, however, to the surprise of many of us, pasting those things on to relatively traditional campaigns has worked better than the Dean campaign’s organic strategy did. The biggest difficulty for whatever version of next time comes around will be remembering not to believe our own PR.