Where You Eat And How You Vote
The Atlantic has a fascinating story about all the technology available to the Romney and Obama campaigns this year, including the ability to micro-target voters based on their preferences. The article details some of the ways your preferences are identified by the campaign:
Both Obama and Romney's sites allow, if not encourage, visitors to login to their campaign websites with a Facebook account, thereby unveiling a wealth of information: email address, friend list, birthday, gender, and user ID. Obama's team, in accordance with the president's call for greater transparency, details his campaign's privacy policyin an exhaustive 2,600-word treatise. It begins like an online Miranda Rights: "Make sure that you understand how any personal information you provide will be used." Then things get a little weird.
Among other points, the policy says the campaign can monitor users' messages and emails between members, share their personal information with any like-minded organization it chooses, and follow up by sending them news it deems they'd find worthwhile. In other words, target anger points. Then there's something called "passive collection," which means cookies -- lots and lots of cookies. Obama's campaign, as well as third-party vendors working with, spray trackers so other websites can flash personalized ads based on knowledge of the trip to barackobama.com. And finally, near the end of the policy, comes one more caveat: "Nothing herein restricts the sharing of aggregated or anonymized information, which may be shared with third parties without your consent."
Romney's site apparently wants even more from its visitors, asking users who login with Facebook to "post on (their) behalf" and "access (their) data any time" they're not using the application. You can deny both functions.
Romney has been using the technology to target different voters on the web. We're living in the future!
Did you know that Republican voters drink Diet Dr. Pepper while apathetic Democrats drink 7up? Politically active Republicans visit FoxNews.com (you don't say) while disinterested Democrats visit MTV.com and dating websites (not going to make any jokes).
Anyway, if you click on the image above, you'll see that election technologists have figured out how restaurant preference correlates with voting patterns. As the parents of young children, we rarely get out to eat, much less to those chain restaurants we'd have to drive to the suburbs to get.
I'd say that Dunkin Donuts, Panera Bread and McDonalds are the three I've gone to in the last year. Which means that election technologists would peg me as either Democratic skewed with a lower chance for turning out or Republican skewed with a higher chance for turning out.
They'll never figure me out!
- Comment (30)
- · Quote
- · UnfollowFollow (2)
- Pages:
- 1
- 2
Comments:
Re: Where You Eat And How You Vote
By the way, Dunkin Donuts has created interactive bus ads that spray the aroma of coffee. Annoying? Or Inspired?
Mar '11
Re: Where You Eat And How You Vote
Does the aroma of coffee overcome the aroma of bus?
Dec '11
Re: Where You Eat And How You Vote
What about the voter who wants them both to bugger off?
Where am I supposed to eat?
Edited on April 11, 2012 at 4:32pmFeb '11
Re: Where You Eat And How You Vote
*Really* apathetic people, of course, aren't going to be visiting a campaign website at all, so I don't know how this approach can tell them that apathetic dems drink 7-up.
"We're living in the future!"...in 1964, Eugene Burdick published a novel called "The 480," about a political campaign's use of information technology (the might IBM 7094!) to categorize the American electorate into 480 groups and optimize the targeting of each one. It was based on actual work done for the Kennedy campaign by a company called Simulmatics.
Re: Where You Eat And How You Vote
Well, apparently we're living the past, too.
Jan '11
Re: Where You Eat And How You Vote
Again ... I know I keep harping on it ... the ability to accumulate data is not the same as knowledge.
Data mining is an exercise in taking one attribute (eating at Dunkin' Donuts, for example) and seeing whether it occurs more frequently in results, or whether it happens to occur in tandem with other attributes. But that only gets so far. One needs to investigate how and why the attributes occur more (or less) frequently ... and that inevitably triggers overconfidence in the analyst. Why? Because if the analyst knew (before the data mining exercise) why a particular attribute affects the result set ... he wouldn't have needed to do the exercise in the first place. In real life, correlations only start the investigation; they don't end it.
If you collect data and discover a surprising correlation between Democrats and Dunkin' Donuts ... beware the analyst who immediately says he knows why that is. He doesn't.
Feb '11
Re: Where You Eat And How You Vote
Clear out your history and clear out cookies before and after visiting the site on the web and you should provide them with no data. Where they absolutely go data rich is if you do plug in via Facebook or LinkedIn or some other such site. The thing with sites like Facebook is that you'll have a means of cookies communicating through cross-referencing with your friends to provide a wider profile. There are lots of Facebook marketing books out there, such as "The Like Economy". It's easy enough to infer what it is that's going on if you have any type of programming background.
Edited on April 11, 2012 at 4:58pmFeb '11
Re: Where You Eat And How You Vote
KC Mulville: Again ... I know I keep harping on it ... the ability to accumulate data is not the same as knowledge.
Data mining is an exercise in taking one attribute (eating at Dunkin' Donuts, for example) and seeing whether it occurs more frequently in results, or whether it happens to occur in tandem with other attributes. But that only gets so far. One needs to investigate how and why the attributes occur more (or less) frequently ... and that inevitably triggers overconfidence in the analyst. Why? Because if the analyst knew (beforethe data mining exercise) why a particular attribute affects the result set ... he wouldn't have needed to do the exercise in the first place. In real life, correlations onlystartthe investigation; they don't end it.
· 3 minutes ago
But if you're marketing, you really don't need to know cause and effect, just that there is a correlation and how to focus your resources. That's what this is all about -- focusing resources, figuring out how to manipulate (making people mad), and knowing who knows who.
Apr '11
Re: Where You Eat And How You Vote
KC Mulville: Again ... I know I keep harping on it ... the ability to accumulate data is not the same as knowledge.
........In real life, correlations onlystartthe investigation; they don't end it.
If you collect data and discover a surprising correlation between Democrats and Dunkin' Donuts ... beware the analyst who immediately says he knows why that is. He doesn't. · 1 minute ago
It's true that a lot of those correlations have obvious unhelpful origins. Church's Chicken is most frequently eaten at by Democrats? Surely not! It's not as if these campaigns aren't good at the subsequent analysis, though. The modern technique was pioneered by Mitt in his 2002 gubernatorial race, and the 2008 Obama campaign was its finest hour to date.
The 2nd most upsetting data (after the King of Bain approval plummeting), to me, from the whole campaign so far has been that Mitt's felt pushed to spend on primaries to the point where he hasn't been spending on this stuff. Mitt's clocking in at about $1m, while Obama's spent $12m on it. We desperately need to be doing better.
Mar '11
Re: Where You Eat And How You Vote
It doesn't surprise me that the Un-Cola is the choice of the apathetic Dem.
However, I am encouraged by the preference of Republican voters. After all, if I'm a Pepper and you're a Pepper and he's a Pepper and she's a Pepper then we should surely have enough votes to win.
Feb '11
Re: Where You Eat And How You Vote
david foster: *Really* apathetic people, of course, aren't going to be visiting a campaign website at all, so I don't know how this approach can tell them that apathetic dems drink 7-up.
"We're living in the future!"...in 1964, Eugene Burdick published a novel called "The 480," about a political campaign's use of information technology (the might IBM 7094!) to categorize the American electorate into 480 groups and optimize the targeting of each one. It was based on actual work done for the Kennedy campaign by a company called Simulmatics. · 23 minutes ago
If I go to Obama's website and log into Facebook, then they can find out about apathetic Dems. It doesn't need to be the apathetic Dem who went to the website. If I went to the Romney site, they'd see a lot of active Dems I grew up with and would think they were being spied upon.
Jul '10
Re: Where You Eat And How You Vote
Okay, so on the Monday before election day we all go to Cici's and Sonic and put up signs that say "Vote Tomorrow."
Then on Tuesday we swap 'em out for "Vote Today" signs, and take the first set over to the Cheesecake Factory and Boston Market.
Sep '11
Re: Where You Eat And How You Vote
I find this intriguing--but I'd bet that either campaign doesn't need to do a whole lot to figure out exactly how the Hemmingway household will vote.
You're registered to vote, right?
Your county registrar keeps records of voters--including which elections they have voted in. In a four-year election cycle there are eight elections--practically everybody will vote in the presidential general election (i.e. November 2012); practically nobody will vote in the primary election in an "off" year. (This will vary from state to state, depending upon when state, county, and municipal elections are held. In Pennsylvania, for instance, state races, including the legislature, are all held in even-numbered years. So the primary election in April of 2013 will only include municipal races--very low turnout.)
You can easily identify voters who only vote in the presidential general--and voters who vote in every single election. Those "super voters" are crucial to a primary campaign--but can safely be assumed to show up for the presidential general prepared to vote a straight ticket.
They don't need to know where you buy over-priced free-range bagels--they already have your voting history.
May '10
Re: Where You Eat And How You Vote
I just knew that waiter looked familiar!
Apr '11
Re: Where You Eat And How You Vote
The media consumption charts are pretty flattering. Apparently, we like business, documentaries and sports (the documentaries are the flattering bit, to my mind), while they like teenage boy stuff, lifetime movies and soaps.... Oh, I guess this goes back to your gender gap post.
The other media consumption chart says that we consume more radio and internet, while they consume more TV and Newspapers. I think I prefer our side of that divide, too, but think that this is more subjective ground for superiority claims.
I think it's interesting that PGA Golf is more Republican than Nascar, and College Football gives a bigger vote lead than either.
I'm entertained that Adult Contemporary is absolutely middle of the road by both axes, but disappointed by the strength of the Democratic tilt of the Spanish language stuff.
I agree that the specific websites reflect well on us, but I'm confused by the claim that Democrats watch more UFC (go us!), but we go to UFC.com more (Oh...). I also don't get why apparently car ownership as a whole trends so strongly GOP.
Thank you for this story, Mollie! Lots of fun stuff there.
Apr '11
Re: Where You Eat And How You Vote
Palaeologus
Okay, so on the Monday before election day we all go to Cici's and Sonic and put up signs that say "Vote Tomorrow."
Then on Tuesday we swap 'em out for "Vote Today" signs, and take the first set over to the Cheesecake Factory and Boston Market. · 19 minutes ago
How much would it cost to sponsor an extra "Free Ice Cream Day" at Ben and Jerry's? Those things always seem to take most of the non-working day to queue through. No way more than half of those who have done it have the willpower to go vote afterwards. Anyone know anyone at Unilever?
Jul '10
Re: Where You Eat And How You Vote
Of course, if you eat at Kilroy's of DC soirée fame, you may not be Captain America, but he has asked you to call him Steve and let you hold the shield.
Jan '11
Re: Where You Eat And How You Vote
Hang On
But if you're marketing, you really don't need to know cause and effect, just that there is a correlation and how to focus your resources.
After a deep pause, I have to admit that you're right. The indigestion comes from accepting the idea that campaigns are much more like product marketing than any collective political conversation. As soon as I read your post, I said to myself, KC, get your head out of the clouds (or whatever other useless place that springs to mind).
Dec '10
Re: Where You Eat And How You Vote
I guess you are ok as long as you dot eat at (shudder) Olive Garden.
Apr '11
Re: Where You Eat And How You Vote
Guruforhire: What about the voter who wants them both to bugger off?
Where am I supposed to eat? · 1 hour ago
Edited 1 hour ago
Aren't you in northern Scotland? I'd have thought "in northern Scotland" was a pretty good answer to that question. ;-) Probably not enough to avoid the stuff online, but I think being confronted with political information online is probably an unavoidable cost of frequenting online political sites.