An interesting roundtable took place last week at Cambridge University with the provocative title “Are we being sold online?” While this is a shocking way to frame how companies like Facebook and Twitter monetize data, it is of course accurate. Privacy settings are oftentimes reassuring, but are merely protections from the public space. They don’t prevent the platforms from collecting and using that data for various purposes.
The most interesting takeaways for me are the extent that data can be mined and the argument against social anonymity.
What Facebook knows about you
One of the most fascinating talks was given by Michal Kosinski, the Operations Director of Cambridge University Psychometrics Centre. Kosinski is doing some fascinating research looking at personality traits derived from social profiles and crowd IQ.
The first thing that he points out is that in addition to fairly straightforward information that he can predict from a person’s social profile such as gender, age, and race, he can also parse data to determine sexual orientation, personality, intelligence quotient (IQ), political views, religious affiliation, and whether your parents are divorced (off of the top of his head). He further points out how accurately his predictions hold.
Kosinski asserts that single data points are fairly insignificant, but an amalgam of information once compiled “has a completely new quality.” He should know. Through Cambridge, he runs the myPersonality database that has 4 million profiles and 6 million tests on file. If you’re a married thirty-something father of two kids who lives in Cincinnati by way of Seattle and likes Diet Pepsi on Facebook, Kosinski probably has a good sense of what your preferences are. But beyond that he know your friends better than you know them, too. As it turns out, the company you keep says a lot about you.
Bluefin Labs recently ran an analysis of the Facebook preferences of Romney fans versus Obama fans, a simple analysis of just two convergent data points. The information was notably a bit weak, but consider if they ran that data in a more sophisticated model. The fact that Obama supporters like Heineken beer is pretty worthless, if Heineken knew with more specificity what sub-groups of Obama supporters liked their beer they could develop a pretty cost-effective marketing campaign. Kosinski makes the point that corporations, ISPs, governments, web browsers, social channels, and others can run the same analyses that he does. All of this with your complete consent.
With the bluntness of a scientist he says: “You don’t pay for it, which suggests that you are a product. And you are sold.”
What if it’s not a bad thing?
Kosinski gave the example of the software he uses daily that collects data about him:
Facebook understands more about your friends than you do, you use Bing to search (he is a Microsoft consultant, so I assume he has to say that), Gmail reads your email and your friend’s emails, you use Google Docs, you buy books on Amazon, your cell phone can locate you even without GPS, search engines know what you search for, the movies you watch and books you read.
But then he makes a poignant point: “If you want to be a black sheep and opt-out of those technologies, it is bad for you, your friends, and for the economy.” In other words, you are less connected, your friends are less connected and you have less power to participate in the economy, effectively cutting yourself off from the consumer long tail.
Would you rather have residual income or a one-time buy-out?
I had a disagreement with a fellow a few months back when I suggested that Facebook wouldn’t think twice to sell personal data to third-parties if it served its interest. Facebook’s statements and people’s sentiment may disagree, but there are many examples of near-monopolies that parse words in order to push the envelope of consumer privacy. So long as our personal data is entrusted to others online, this will always be a justifiable concern.
But by selling aggregated, highly segmented data that continues to improve – it seems that social networks enjoy a residual income stream that they’d rather not compromise. In other words, the most important consumer protection may be the fact that our data is being collected, analyzed, and sold to advertisers in segments.
What do you think? Do you agree that you are being sold? Do you object? Is there a benefit to the 21st century Luddite life?