Thread by @davekarpf

I'm watching #TheGreatHack on Netflix.

15 minutes in so far and my initial reaction is "oh come on, can we just not with the Cambridge Analytica villain-hagiography?" (1/probably a lot)

Minute 12: the film credits Cambridge Analytica for Ted Cruz's victory in the Iowa caucuses, showing a clip of CA's CEO bragging about the win. IIRC, the Cruz campaign fired CA soon after, because they found the data and modeling to be basically garbage.

Minute 15: Another clip of Alexander Nix, giving a staged sales pitch about how powerful his product is.

...Folks, just because the sales guy tells you his beans are magic doesn't mean you have to believe him.

Minute 20: Now we have whistleblower Christopher Wylie.

Wylie explains how they used a Facebook app to build psychological profiles of every voter.

Left unsaid: whether these profiles were actually any good. (How much more can you predict from them than from the voter file?)

Minute 23: Wylie calls CA a "grossly unethical experiment." "You are playing with the psychology of an entire country without their consent or awareness..."

Sure, okay, sounds bad. But several BILLION dollars were spent on ads in the 2016 election. This is different *how?*

Minute 25: footage of activist/Marketing Prof David Carroll giving a keynote address about how "propaganda works."

I respect the activist tactics Carroll used against CA. Pretty cool, very creative.

But Carroll begins from the *premise* that propaganda works. (...)

(...) He doesn't provide *evidence* that it works.

And his tactics were designed to reveal that CA (1) had a lot of data and (2) that's creepy.

Both can be true without helping to promote CA's own marketing hype.

Minute 29: Brittany Kaiser. "I have evidence that the Brexit campaigns and the Trump campaign could have been conducted illegally."

Hard yes.

There is an important difference between "they broke laws" and "they were devastatingly effective."

Former yes, latter probably not.

Minute 35: the filmmaker asks about Kaiser whether "the idea of a company conducting large scale analysis of a population and then [marketing to them] challenges the individual sense of autonomy and freedom and to the idea of Democracy."

...Dude, that's just online marketing.

These filmmakers are leaning really heavily into the notion that there's some great violation of the public trust if people "have been experimented on" without their consent.

That's a dramatic but empty phrase. All websites run A/B tests. All of them.

(which, *ahem*, you can learn about in my 2016 book, Analytic Activism.)

Minute 41, Kaiser explains that the Trump campaign and CA didn't focus on every voter, but on *persuadable* voters in *swing states.*

Folks, this is standard practice in every political campaign, dating back several decades. I was taught this stuff in 1997-1998. Commonplace.

Minute 52: Mark Zuckerberg testifies before congress, states that CA told FB they'd deleted the data, and FB only later found out that this wasn't true.

...Look, there's a ton to criticize FB for, but this shouldn't be on the list.

FB sent a cease and desist letter. CA replied saying they were complying. What more was FB supposed to do? Send the Facebook-cops to CA HQ?

There is a ton wrong with Facebook's business model, and with Facebook's business practices. This shouldn't be part of the grievance list.

1 hr mark: now we're into a long sequence of CA marketing materials.

CA made some bold claims about its efficacy to prospective clients.

It's a strong rhetorical device, but it's also hollow.

1:07-1:10: okay, can we talk about the whole "weapons-grade" microtargeting line?

DoD and other military entities have shown an interest in advances in propaganda techniques. No surprise.

That doesn't make your FB data the new bullets, folks.

1:15: now they're using that (awesome) undercover video that Britain's channel 4 did on Alexander Nix.

The video catches him overclaiming the company's role in the Trump campaign, and then boasting about illegal activity.
(...)

(...) you could read this video as proof that CA was far more central to the Trump campaign than others claim. Or you could read it as proof that Alexander Nix is a quite a bullshit artist when he's trying to attract more clients.

The movie leaves the former impression.

Okay, I've reached the end of the movie. I found it less frustrating overall than I had feared.

Bottom-line: if you want to prop up Cambridge Analytica as an object lesson in datafication, this movie is a good vessel for doing so.

But (...)

(...) the movie uncritically adopts the premises of CA's biggest critics (Carroll, Cadwalladr, Wylie) -- namely that all of CA's marketing claims should be treated as prima facie evidence of CA's effectiveness.

And that's not how we should ever treat marketing materials. (...)

(...) Left completely unexplored is how today's digital microtargeting compares to yesterday's digital microtargeting or the political marketing practices from decades past.

And the answer there is that it isn't particularly different.

Did the LeaveEU campaign target persuadable voters in the Brexit campaign? Yeah, it would be weird if it didn't.

Did the Remain campaign target persuadable voters as well? They sure better have!

Did the Trump campaign spend way more on Facebook ads than the Hillary campaign? Yes.

The Hillary campaign had a massive data operation. They just didn't prioritize Facebook.

(self-plug: I've written about this elsewhere civichall.org/civicist/digit… )

We HAVE to get better at talking critically about digital campaign practices without amplifying the boldest marketing claims the vendors make.

The direct effects of digital propaganda are small and short-term. Particularly in elections that are already flooded with advertising.

Final thought: it seems like usually the debate boils down to "Cambridge Analytica is bad and the data markets should be regulated" versus "CA is ineffective and everything is fine."

There needs to be room for a 3rd option. CA was ineffective. Data markets need to be regulated.