There have been a few stories lately (Here is one ) on how the Obama campaign kicked the Romney’s campaigns butt in terms of data analysis. There is another, long article, in the current edition of Technology Review.
For instance, anyone donating to the Obama campaign with data analyst in his or her job title got a job offer. The Obama campaign was able to get DVR logs (without names associated) by people they had identified as likely to be swayed. They did all sorts of experiments on messaging, and found, for example, that self-identified Republican women were more likely to switch to Obama based on messages about women’s issues.
Basically the new data analysis allowed the Obama campaign to individually target people and not groups. That is why they could use cable, as noted in the linked article. They could also make very accurate predictions about likely outcomes, using even more precise methods than Nate Silver did.
First, what is your opinion about this? Is this a game changer for campaigns in the future? It seems that they abandoned mass blanketing of their message, with excellent results.
Second, several articles noted that people in this area are predominantly young and Democratic. Sure Republicans can (and did) hire companies to do this, but will truly dedicated workers give the Dems an advantage.
I’m not saying this is going to guarantee Dem victories - bad candidates will lose no matter how much data there is. But it might give an edge.
This appears to be the story of this campaign, like Sarah Palin being a twit was the interesting story of 2008.
So did Romney. Obama also got a lot of high information voters. In combination, Obama had millions more voters overall support him. Not sure what your point is.
Still, it points out which side was more informed. That is, which side had and absorbed more data.
The Democrats had 30 years of economic data to show that trickle-down economics does not work in today’s economy. The data showed that most of the prosperity trickled up. They had data that reasonably demonstrated that Romney was out of touch with the majority of Americans. They had data that Romney was condescending toward, and held contempt for the middle class. They had data showing that cutting revenue while at the same time massively increasing spending somehow increases the nation’s debt.
The voters who voted Republican had data that Obama was not born in the U.S., is a Muslim, an fascist, and a Socialist – all of which is false.
There are threads where this comment would be on topic, but this isn’t one of them. If you just want to dump on Obama voters, find a more appropriate thread.
The irony, to me, is that this is the sort of direction that advertising has been moving, consistently, for two decades. It is not surprising that the Obama campaign devoted considerable energy to these efforts, but it is very surprising that major business-involved players from the other side, (Koch brothers, Crow Family, etc.), were not pushing their organizations to follow a business model and do the same thing.
Sure, it is a game changer, just as it has been for every other advertiser who targets individuals for basic retail. Mass marketing will never go away, but targeted promotion is going to continue to grow in size and reach.
It is like an arms race. Once the directors see the results, they will demand the same actions from their subordinates. The Democrats were more successful on this occasion by getting their first with the most, but I suspect that the Republicans are going to follow very quickly. (The Republicans already have a very large “grass roots” base, so it is merely a matter of putting the technology into their hands for them to have the same level of volunteer commitment.)
Whoever’s running as the Democrat in 2016 will have access to the software that the Obama campaign has already built, and probably more consultants who worked on the data side of the campaign. That’s going to be a significant advantage over the Republican challenger, who presumably will have to build their own. This is going to matter for at least one more cycle.
Edit: It’ll be interesting to see how better analytics affects the primary seasons. There’s no time or money for every candidate to build their own system, so will access to the existing systems be a new method for the party establishment to throw their weight behind their favored candidates?
First, thanks for actually reading the link. However, I’m not sure if all advertisers actually do target individuals rather then demographic groups. The marketing classes I took 20 years ago certainly didn’t mention this, and I suspect a lot of advertisers are doing it the old fashioned way. Certainly Amazon is not. I can’t imagine that Georgia Pacific is on the cutting edge, and the the Kochs’ get that involved in that level anyhow.
True, but after 2008 when the Obama campaign did this quite well, they spent a lot of time evaluating what they could do better. What are the gong to come up with next time? I do data, and I can testify that a lot of very smart people don’t get it - and I suspect my colleagues are far more open to new ideas than your average pol. It will be interesting.
The TR article noted that the Romney campaign did a much better job than the others during the primary season. I suspect the Dems won’t release the tools to the primary challengers, so they are going to have to develop their own.
Which can be done. I helped a friend who was running for school board one year, and I got the database of all voters in our town, their registrations, and how often they had voted. I produced a mailing list for him based on the most useful voters, to safe money. Primitive, but it was surprising how much voter information is freely available.
This has nothing to do with the point. The issue is not what data the voters have about the issues, but what data the campaigns have about the voters. For instance, if the studies show that a significant set of voters have one misconception in a region, an ad can be tailored - just in that region - to attack the misconception. The articles didn’t say, but I bet there were lots of targeted ads in Ohio targeting Romney’s lie about Chrysler sending jobs overseas.
As a data analyst I don’t know how much of what was done even counts as analysis.
Sure they are slicing and dicing data to identify segments to which they can advertise, but the important part of having that work is the message being imparted.
Having some schtick to sell tyre workers is the main point. Identifying where they are to target advertising is a simple step. Hell you don’t even need a pivot table in Excel. Just look where the factories are. However a pivot table of jobs or industries by zipcode, or female tertiary students, or mexicans, or blacks would require little effort.
The trick is thinking of a message to sell to a particular group, devising a way to sell it and then identifying the whereabouts of the concentrations of the audience. But that last step is simply data matching, maybe data mining on a resume.
Romney only didn’t succeed at the data matching step because he failed at the prior steps. He had no message and no target audience.
But close your eye and think back. Sure it seems like things were close a couple of times but really, in all honesty, Romney’s whole campaign was a trainwreck. He was laughably inept, his own words and actions were his worst enemy. Come on, a presidential candidate who wanted to hide his tax returns. You’re fucking joking. So it’s no surprise that it ended like this - this was his standard performance level.
Just as an example of how unsurprising the Obama “data analysis” is, in the Australian 2007 election campaign the Labor party (in Opposition at the time) gathered demographic data for every electorate. They then targeted each electorate with a specific strategy. In the John Howard the PM’s electorate they discovered a high rate of Asian migration. So they engaged volunteer university students to go around in the evenings and on weekends to talk to the locals in their first language.
I was involved with an advertising focus group whose sole focus seemed to be creating unflattering portrayals of other parties’ members. My experience left me convinced that representative democracy is pretty much a fiction. I was constantly amazed at the BS that people would quote as fact that in reality was opinion or slogan and distressingly often the product of the ads we were critiqueing.
What they did went considerably beyond this, which is pretty standard. What if 80% of the tire workers are set for Obama anyway? Targeting them is useless and a waste of money. But if they could get a list of potential converts, and figure out the programs they watch most frequently, then the targeting becomes very precise. Cable is excellent for narrowcasting and much cheaper than the networks.
Like I said, targeting groups is standard marketing. They didn’t target groups - they targeted individuals. That is the innovation.
His message might have been one we disagree with, but it was a message. And the company he hired knows all about demographics. Perhaps Obama would have won anyway, but this gave him an electoral college landslide.
I assure you, Romney’s campaign was smart enough to run ads in Spanish. That is a gross demographic. Demographics of size one was the innovation. They had a database of all voters, and could guess who voted for them in 2008, identified those to win over, found those easiest to change, found what they watched, and bought time there. They also could target get out the vote using the same means.
Wrong. Obama underperformed among higher education and older age demographics compared to 2008 and outperformed his 2008 performance among the undereducated and young in 2012. Obama was able to turn out a startling number of first time voters, who tend to be politically detached and low information. Those are simple facts, whether you like them or not.