Data Crunching for Obama

This article buried in the Saturday’s New York Times reports that the Obama campaign has invested heavily in microtargeting.

Microtargeting uses computers and mathematical models to take disparate bits of information about voters — the cars they own, the groups they belong to, the magazines they read — and analyze it in a way to predict how likely a person is to vote and what issues and values are most important to him. Often these analyses turn up surprising results; for instance, Democrats have taken advantage of the fact that many evangelical Christians are open to hearing a pro-environmental message.

Though this is technique has long been favored by the Republican party, especially during the 2000 Bush campaign, even Republicans agree that he “Obama campaign has appropriated it and taken it to a new level.”vijay.jpg

One of the largest data banks is Catalist, a for-profit company that specializes in providing data for the Obama campaign. Turns out its chief technology officer is 34 year old Vijay Ravindran, former director of the ordering-services group at Amazon.com, where he led a team of about 130 engineers who built and maintained the site’s “shopping cart.

From the Washington Post:

The work being done in Catalist’s McPherson Square offices—which, with its multiscreen computer terminals, resembles a Silicon Valley start-up—is helping revolutionize the fields known as data mining and microtargeting. … Catalist was founded in August 2005 by Harold Ickes, the longtime Clinton deputy White House chief of staff, after the 2004 campaign to address the Democrats’ inability to harness data. One of the first hires was a young engineer, Vijay Ravindran. … “With my hiring, he made a decision that this was going to be a real company,” Ravindran says. As the chief data-architecture guy at Catalist, he’s part of a new trend in political technology: As data become more important in campaigning, candidates are increasingly turning to the tech industry for business-level expertise.

In a feature on political strategists and microtargeting, from fastcompany.com [via the newstab, thanks brijo1], Ravindran says:

“In the political space, I felt it was very important to build a computing architecture that would take in real-time data, get them into a standardized format, and then load them into a place where they could be snapshotted out for particular purposes. That didn’t exist before. Now we have an architecture that scales more than 15 terabytes of data while providing an interface for users to work with. We expect to leave this election cycle with a piece of permanent infrastructure that enables groups to do microtargeting more efficiently than ever before. It all boils down to one principle: Leave no data behind.”

Below the fold is a video where Ravindran talks a little bit about what he does. Fascinating stuff.

21 thoughts on “Data Crunching for Obama

  1. to take disparate bits of information about voters — the cars they own, the groups they belong to, the magazines they read —

    Violation of privacy? I smell a lawsuit not too much into the future…

    M. Nam

  2. Violation of privacy? I smell a lawsuit not too much into the future…

    Why? Firstly, both parties do it. Secondly, this sort of consumer micro-targetting information is already compiled and available for marketing firms. That’s actually where it comes from.

  3. Ennis,

    First, just because both parties do it does not make this right. Second, I did not mean a lawsuit against the Democrats or Catalist – it should be against the companies that provide them the data. When you buy a car or a TV or or subscribe to Time or join a local association, there are strict privacy guidelines that they must adhere to. If it turns out that they have shared your information with anyone, they will be in deep do-do.

    M. Nam

    PS: The only exception is data that’s public. When you buy a house and it’s registered in your local municipality, anyone has the right to see when you bought it at what price etc etc.

  4. This explains the daily pile of mailers my spouse and I got from various candidates. We have voted in every election since we became eligible and must have landed on some lists. Of course I would have loved to be part of the pre-polling/surveys but Calif being solid Dems territory all I got were 2 uber short phone surveys for propositions 7 and 8.

  5. Violation of privacy? I smell a lawsuit not too much into the future…

    in canada at least political parties are above the privacy legislation as are non prfits. look up gramm-leach-bailey for your equivalent.

  6. When you buy a car or a TV or or subscribe to Time or join a local association, there are strict privacy guidelines that they must adhere to. If it turns out that they have shared your information with anyone, they will be in deep do-do.

    They always sell your information. Try this – use a fake initial in one of your purchases and watch it percolate.

  7. They always sell your information

    True. But “they” are unethical. What do we make of the folks who use this stolen information?

    “They” bootleg movies too, only to satisfy the needs of the market. IMO, Catalist, the political parties and Ravindran have the same dubious ethical standards as those who watch bootleged movies, listen to pirated songs etc. Worse, they do this for a profit.

    M. Nam

  8. How is this stolen information? I’m quite confused.

    MN – read this wiki article on Choicepoint. It gives you a good idea of the sort of data mining that goes on routinely. Then we can talk in specifics.

  9. Ennis,

    Sorry – stolen was a poor choice of word. “Unethically obtained” is better.

    You car dealership has a privacy agreement – however, they can share it with the DMV, who then are supposed to keep it private, but then, it’s the Government. What do you expect – they leak informaiton to insurance companies, who bombard you with fliers.

    I know the lines: “Everybody does it…that’s the way things work”. But that does not make it right.

    Let’s see if the master of “Change” can change this. There’s always hope!

    M. Nam

  10. Fascinating, reading the FTC rule regarding privacy rules that apply to car dealerships, I don’t think the car model information is private information. I believe the privacy rules apply how the vehicle was financed and any personal information that may have been obtained during completing the financing paperwork. More details here.

  11. I haven’t read the post but I can just tell you intuitively that micro-targeting makes no sense. As if people vote based on size. And for the Obama camp to focus on desis just reinforces the worst stereotype out there about us.

  12. 1 · MoorNam said

    Violation of privacy? I smell a lawsuit not too much into the future

    Always wondered about that! I have worked on such data myself (marketing industry). Aware of any lawsuits on it?

  13. 12 · Manju said

    I haven’t read the post but I can just tell you intuitively that micro-targeting makes no sense.

    If it doesn’t then there is another bubble ready to burst in the marketing industry ;)

  14. Extrapolating information about attitudes and political leanings from profiles of consumers doesn’t strike me as being inherently sinister. Of course, the technology could conceivably be turned to other kinds of Big Brother-ish use. This election happens to be taking place in a moment when better educated and therefore likely more liberal people are more inclined and better equipped to network online.

    Traditional scandalmongering techniques are alive and well, though, and are technologically up to date in a very broad or crude sense, as in the Auntie Zeituni scandal. While The Times of London broke the story on Thursday with the clear intention of damaging Obama on these shores at the eleventh hour, and, I suspect, without being bound by U.S. confidentiality and privacy laws, the story was addressed quite rapidly by several segments of the American press on the Left, Left-ish and Right, including a selection of random, scandal-happy blogs.

  15. Ok, I have to jump in into this whole “privacy” argument.

    A lot of the demographic information in the voter information is available as part of public records. The states file and disclose this information. Pollsters (and pollster tech companies like Catalist) are operating within the ambit of the current laws.

    If you feel scared about it, I suggest you take it up with your neighbourhood Federal circuit judge.

  16. A lot of the demographic information in the voter information is available as part of public records. The states file and disclose this information. Pollsters (and pollster tech companies like Catalist) are operating within the ambit of the current laws.

    Difficult to believe. Catalist is targetting individuals/family based on data they got – so it’s got to be more granular than that. They probably got their hands on voter lists. You know, someone else could get their hands on voter lists and put them to…a different use.

    M. Nam

  17. This trend started with behaviourial economics…the business world caught the craze with neuromarketing [ Buying in ]. Actually Clinton’s campiagn manager came up with an interesting book [ Microtrends ] on profiling various groups of people. All these is advanced data mining [ Supercruncher, Numerati ] applied to society/economics/business etc.

  18. You know what’s ugly, unnecessary and against our commenting policy? Personal attacks.

    I mean, if you really have something “important” to say about someone who isn’t even related to the topic at hand, have the guts to do it under your own name, so that the person you are flaming knows whom to spill a drink on, at the next party. Thanks.