(Feednet 3): let’s make a bracket for the Cambridge Analytica Facebook scandal




(Because someone — the Man? Zuckerburg himself? flagged my TinyLetter for abuse, and I refuse to be silenced or investigate further while the hottness of my takes are at stake, this email is a blog post for now. The future of Feednet’s distribution is TBD…)


Facebook is still in. People writing scandalized thinkpieces about ‘the American electorate’ are still in. The American electorate is definitively out. Christopher Wylie’s hair is still in. The March edition of Feednet is all Cambridge Analytica, all the time.


A quick summary for people who don’t have a weird metrics fetish, from the Verge“Over the weekend, reports from The New York Times and The Observer confirmed that voter-profiling company Cambridge Analytica had amassed data on over 50 million Facebook users. This information had been collected legitimately by an academic researcher, Aleksandr Kogan, who passed it on to the profiling firm…Cambridge Analytica gathered this information to develop “psychographic” profiling tools, which it claimed could tailor political ads to users’ personality traits.”


I freaked out about Cambridge Analytica when this article came out in early 2017. I stopped having stress dreams about them in November, when everyone started writing articles abt how their hip dystopia marketing tools are just Myers Brigg w/ no real ROI, and Ted Cruz, who used these tools the most of all the presidential candidates, lost. Most of what Wylie describes (as the Verge points out) is mostly just expensive, high volume microtargeting. Microtargeting IS creepy, but it’s also standard practice for large entities marketing on social media channels (Obama did it).


Standard practice or no, 50 million people’s data is uh, still a lot of data to use in strategizing around a political campaign without getting consent. Some of the very obvious and true conclusions coming out now are as follows: 1. Facebook should have better data oversight practices. 2. Cambridge Analytica is really vile — both because of their their data practices and their weird supervillain promises to clients. 3. Tech companies collect a terrifying amount of unregulated, unacknowledged data, and the manipulation that happens with that data promotes US imperialism, capitalism, and the profit of the elite. All of these are fairly true conclusions!

But most of the media hype surrounding the Cambridge Analytica scandal focuses on psychometrics. That makes sense: the idea of creating psychological profiles for people, guessing intimate parts of their personality based on their social media presence, and then using that information to manipulate them on a scale so large that it can actually change election results, is a freaky idea. But it’s only freaky if it actually works.


And when we ignore Cambridge Analytica’s own hype, does we have any proof it actually does? Is targeting “introverted teens in Bloomington zip code” really that much more valuable than “teens in Bloomington zip code,” when it comes to taking over the world and our brains? And if so — should we be using these tools?

To help suss out the hype from the more preposterous hype, I started a fight with organizing communications expert and noted mean twink Zenen Jaimes Perez. And as Zenen said, Ted Cruz WON Iowa…

Zenen: Cambridge Analytica’s work for Trump tapped into a very powerful strain of anger and fear, regardless of the medium. Ultimately, we can’t see CA’s metrics, right? But intuitively, it makes sense: humans aren’t that complicated. It seems like you could easily figure out what motivates people based on personality. Also… using personality predictors to write messaging sounds like a good way to create dogwhistles.


H: But there’s also this, from one of the many many data focused takedowns about this scandal: “A lot of this data doesn’t give us anything that we don’t already know. Hersh uses a simple case: who owns a boat? Someone who accesses that data will learn that boat owners are likely to be Republican. “That’s a totally useless data point,” he says. “If I have the demographic data — if I know that there is a white man in a Republican town near Virginia Beach who’s rich — I already know they’re Republican regardless of the boat. Boat ownership doesn’t provide any more information.” Is there a difference between a white man Republican who has or does not have a boat?


Z: Yes — duh. It tells you what kind of thing compels that white man Republican! Boats!


H: But I’m skeptical that you could write a better ad swaying a Republican with a boat than for a Republican you didn’t know was a boat owner or not. Well, maybe. But that isn’t psychometrics, it’s boats! I don’t think that self reported extroversion/conscientiousness/etc is more useful for microtargeting than demographic data, like this boat data expert is saying.


Z: Yeah, and again, we don’t have access to the evidence that would prove one way or the other. But I say the personality thing matters: if you are trying to make an ad for someone you know has a depressive personality, you would make that ad different than for someone who doesn’t (barring all the other variables you might be controlling for) — doesn’t that intuitively make sense? They had over 100,000 tests running at one point…


H: My counterpoint to this is astrology.

Z: Please no.

H: Everyone is like i’m an Aries, I’m bold and outgoing, or whatever Aries’ are. But it’s just based on a feedback loop of what your personality ‘should’ be (sorry Chani fans).

Z: Ugh. But Cambridge Analytica’s quizzes aren’t completely based on arbitrary things — like astrology…

H: But people self reporting their answers to any quiz seems useless! Most personality quizzes are so arbitrary, like, “are you shy sometimes?” I also don’t have any evidence but this seems intuitively useless to me. Which, I guess, not having conversations like the one we’re having now is the whole point of microtargeting…


Z: But still, with personality quizzes like the one Cambridge Analytica used, you’re self reporting characteristics that point to personality traits. And take Harry Potter House Sorting quizzes — don’t tell me those don’t pull from personality traits! Because then you’re saying all personality tests, Meyers Briggs, all of them, are fake.


H: Yes that’s what I’m saying! They’re all fake! And as a result, not scary!


Z: I DISAGREE. Plus, there are situational tests that are better markers, like the DISC assessment.


H: Is DISC the BDSM one? From the dude who made wonder woman?


Z: Yes.

H: Well then let’s put the DISC test on Facebook, get 500K people to take it and get rich.


Z: Sounds goods. In terms of what you’re saying about left groups using microtargeting, I definitely hear from progressive groups saying ‘oh we’re doing microtargeting but we’re better cuz we don’t play into people’s fears in our messaging’… maybe. But the $$ and capacity of progressive groups is different… and it’s possible we’re just not as good at it.


By the way, another personality quiz that is 10000% accurate is the the gay cliques assessment test.


Also, all the Harry Potter Sorting quizzes are absolutely correct and objective and i will hear no arguments against them. And I’m a Ravenpuff … so Cambridge Analytica is right.

TL;DR: The ideas above are what an ideal Facebook ad campaign looks like. You may not like it, but this is peak performance. If you can’t microtarget Otters, Ravenclaws, and people with their Venus in Aquarius in your Facebook ads, you are not competitive in the new digital age.

PS — There won’t be an April Feednet because of my real job. Send me “well, actually” responses to this edition until then. Or if you liked this, tell your friends to sign up. See you in May, nerds!