Thread by @aetherlev

Okay I want to talk about the TikTok/K-pop stan let’s-troll-Trump operation and specifically about the data gathering aspect of it.

If you’ve been keeping up with the whole MO of the Trump campaign, Cambridge Analytica, micro-targeting of demographics on social media shit etc etc, you’ll know that gathering people’s info in order to shill for donations is a HUGE revenue stream.

AFAIK that was the reason he ran. To campaign, get donations from the MAGAts, funnel vast lakes of money into his businesses.

Like, it was just that simple. Doesn’t take much to feed racists the messaging they want to hear and suck money out of them.

The Russians/GOP/various other factors caused him to win, which was not in the plan, but anyway Trump at least knows that firing up the base = MONEY.

Yes yes, donations from big GOP donors, but you have to understand - there was a vast lake of untapped racist small donations that Trump plugged a hose into back in 2016.

That hose is powered by data.

(For reference: I’m a senior web dev specializing in data processing and analysis so I sort of know what I’m talking about.)

Have you seen those spammy emails that Trump supporters get, asking for money? Yeah, they’re carefully tuned to get cash out of them.

(Side note, I used to work in online advertising.)

Your regular ad campaign that’s pitching, say, workout leggings uses a huge amount of data and A/B testing to determine what is effective. It’s advertising by statistics.

Part of the reason that FB is so goddamn big isn’t because it’s any good, it’s solely because you can advertise to users based on their bio details and their Likes. Having those signals to determine how likely a person is to buy a product is the stuff of an advertiser’s dreams.

Google works the same way TBH but at least Google’s core functionality is useful.

You think having personalized search results is creepy and invasive until you’re like me, a web dev looking for obscure code errors...

Like, Google is aware of the context of my searches and it always returns good results. If I’m looking for info on bash, for example, it’s not going to show me videos of street fights, I’ll get StackOverflow pages on the bash Linux shell.

Anyway: the thing about this stuff is that it runs on ACCURATE data.

The worst fucking thing about data analysis is when your data is corrupted or inaccurate, and you have no way of filtering it out. It means insights derived from the dataset are 100% junk.

Now, other thing to know is that zoomers and millennials are THE tech savvy people and I would hazard a guess that the vast majority know how internet advertising works. They know about the data tracking.

Probably not in any detail but the use of fake social media profiles and fake emails specifically to make advertising data worthless is a thing going back years.

So the troll operation in this case, by all accounts, used fake emails and temp Google Voice numbers.

And listen, people: you don’t know how fucking genius this is.

I will bet a month’s salary that some poor database admin is now staring at a dataset of over a million ticket reservations that they were going to use to shill for donations and feeling sick.

Not even just donations: I’d bet they were planning on using this data to cross reference against other sources and do additional micro-targeting.

Let’s say they were going to use it to ID moderates to advertise to them.

They might have planned to ID any Democrat trolls and just slide that particular subset on to someone who can fuck with voter registrations.

Or cross-reference signals on location, and determine how to direct their campaign spending based on that.

Anyway - point being, they were probably expecting some level of trolling but not this.

That dataset is probably complete junk. Any registration that doesn’t correlate somehow to known good information is now suspect.

Like, for voters who they have a profile on already, like committed MAGAts, they can probably select them out and add this data point to the profile for future use, but novel signups? The thing that Parscale was bragging about IIRC?

Junk.

You can’t tell the difference between a legit email/phone number and a fake email/Google Voice number.

By automated means, is what I’m saying. Obviously if the email is “[e-mail]” that’s a pretty safe bet for being fake.

So at the very least, this means a TON of extra work for the marketing team at Trump HQ just to clean up the data (somehow) and make it usable.

At worst, it means they toss it.

And trust me on this: Trump may stiff contractors, but he ain’t stiffing them.

That team keeps the money hose working. They are absolutely getting paid.

So imagine if you spent a ton of money setting up a rally and getting the devops in place to collect marketing data (least you think those systems are cheap, they are not) and then...

A few hundred thousand tech-savvy teenage trolls showed up to ratfuck you.

I can’t think of any better example of the power of the people in the modern world than this.

So in conclusion, all the photos of the half-empty arena are nice, but the thing really making me giggle is the thought of Parscale staring at that database and wondering when he’ll be fired.

Quick PS to this thread: part of the reason that this is so genius is because Trump HQ likes to shill for donations by email.

Allow me to enlighten you to how much time and money you have to spend on good, large scale email campaigns in order to get results.

Answer: A FUCKTON.

And you live an die on the accuracy of your mailing list. Fake emails are the bane of your life.

Trump HQ now have to face whether they can drop $insert-ridiculous-money-here to run an email campaign on this dataset that is almost certainly 90% worthless and for which they may not be able to filter out the real registrations.

Again: without known good data to use as a filter, they can’t tell which is a fake registration and which is one from a voter they haven’t got on their lists yet.

And Trumpikins desperately needs to get to those voters.

Like I can’t even tell you how much this whole thing makes me both smile and just internally clench because I know what it feels like to stare down the barrel of a billion records you spent excessive time/money compiling and it’s all fucking wasted.

Other devops/data analysts will know that feeling. :P

Anyway I might be going viral and I don’t have a soundcloud, so get out and vote, Nazis fuck off, support #BlackLivesMatter, and give all the money and power to women of color and trans people

CLAIRE OUT