GUEST POST: Some implications of the Cambridge Analytica testimony

2018-03-29T07:47:53+01:00Thu, 29th Mar '18, 07:47|0 Comments

The following guest post was sent in by Ryan Murdock following my post with an excerpt from Christopher Wylie’s testimony to the House of Commons culture committee two days ago. In the meantime a link to the full testimony is provided below the post.

I’m not sure how much your readers know about online social media advertising, but it’s highly relevant to this story.

I watched the video you linked here, and I got thinking about Christopher Wylie’s comments with regard to potential action of this sort in Malta. I think it’d be incredibly easy to influence an election there.

You need to understand some basic things about online advertising to grasp this. Christopher Wylie started talking about conversion rates but he was interrupted before he could expand on it.

The relevant leverage points are targeting and testing.

In terms of targeting, Facebook offers powerful tools to even a very small publisher with a tiny ad budget. For example, say you publish a book about yoga. You can upload an email list of all the customers who bought your yoga book, and Facebook will analyze those customers by matching the email address with a Facebook profile, and then will generate a “lookalike audience” of people who resemble your existing customer base for you to advertise to.

Parameters can include things like age, gender, location, relationship status, hobbies, favourite entertainment (…SuperOne…?), purchasing behaviour, level of education, and more.

For a publisher, that’s like saying, “Please exclude everyone who’s NOT interested in yoga.” This targeting ability gets your message in front of exactly the right readers so you don’t waste your limited ad budget. But this is just the beginning.

The most powerful thing about Facebook marketing is that you can test ads in real time.

If, say, you put an advert in a newspaper, you have no idea if it boosted sales, who saw it, or what traffic you got from it. At best, you could offer a coupon code, and mark down those customers who quote it as those who saw the ad. It’s hit or miss.

With social media advertising, a publisher can see exactly what sort of click through rate an ad is getting, the conversion rate of the sales pages (ie. what percentage of those “click throughs” either bought your book or gave you their email address in exchange for some free content), and what your “cost per click” is for those ads. So even a small publisher can run an ad set for a week, get enough data, and tweak whichever variable isn’t lining up, and then test again until they hit on the optimum advertising funnel.

In the case of a publishing company, the optimum sales funnel is one that speaks to the most pressing need of a customer so that customer decides to buy the book. It’s an incredibly effective way of reaching people who want to hear your message, when previously you had to shout your message to the world and hope you could cut through the noise.

But this process would be so easy to twist to nefarious goals.

It doesn’t require a conspiracy theory. Something that was intended to be benign can be misused simply because a startup like Facebook had no idea what they were building or what the ramifications would be of this interconnected world. We’re figuring it out as we go along.

If that type of precision is available to even a small publisher with a tiny budget, I can only imagine the sort of sophistication a group like Cambridge Analytica could wield, especially when paired with manipulative psychology, crowd psychology, and an understanding of the core motivations of their target audience. That last factor requires a willing (and preferably local) partner.

Back to this UK Parliamentary Committee hearing, and Christopher Wylie’s testimony.

I think it would be incredibly easy for this sort of influence to be exerted in Malta.

The biggest factor would be their ad budget. If they had solid backing and could spend, say, even $25K per week, they could drown out all other messages. It’s particularly easy here because Malta is already so tribal.

Just reinforce the party line that “Red” hears, drown out or provide even the most illogical rebuttals of “Blue,” and the followers of “Red” will parrot it. That’s even easier when your audience turns to a controlled news source (say, a government-owned channel) for all their information. And easier still if they only speak Maltese, because they’re cut off from unfiltered information from the outside world.

There’s 50% of the population sorted.

JosephMuscat knows full well what appeals to the rest: crumbs from the table, the illusion that “the economy’s doing great, there’s more to come, and you have money in your pocket.” Amplify that message and make sure the swing voter sees it to the exclusion of almost everything else. The illusion of massive economic success and the promise of more on the horizon is enough leverage to sway a Maltese swing voter, and to sway anyone whose self-interest (ie their pocket) outweighs familial or tribal loyalty to the Nationalist Party — which seems like just about everyone outside readers of this blog.

How many times have you seen stories about some random agency giving Malta an A-1 rating, or reports from bodies like the IMF citing cherry-picked quotes of praise? It’d be interesting to see data on the timing of these articles, and to know if they correlate with stories of corruption. Don’t like inconvenient news? Drown it out, or make your biased message seem just as “true.”

These core messages were amplified: “Corruption stories are just a plot by the Nationalist Party — besides, what about X?” and “Don’t wurr-ry! Our economy is great and you have money in your pocket thanks to us!” The 2017 election was fought over the issue of corruption. People saw this all-pervasive corruption and voted in favour of it anyway because they valued money more highly. That is not a victim story, or a conspiracy of global capitalism. It takes a Muscat who understands the motivation of the people to make it work. And that should be cause for serious self-reflection.

But back to the main point.

Given how simple it is to target, test and tweak social media advertising, what Christopher Wylie’s talking about would be really easy to do in the case of Malta.

You could see exactly which messages were resonating with the Maltese electorate within days, and amplify those or retest until you hit the winning combination. The bigger your ad budget, the more data you can gather at a faster rate.

And keep in mind that they’re just targeting an island of 400K people which is already highly polarized, and highly receptive to a skewed message. That’s nothing compared to the challenges or cost of advertising a book in the US market. It’s even narrower if you’re targeting Maltese-language.

Of course this is all speculation on my part. But even someone with a basic understanding of how social media advertising works can grasp the level of sophistication that could be achieved.

I hope this helps fill in some of the context in what Christopher Wylie is saying.

Finally, a slight aside:

Christopher Wylie also mentioned Ghana, where the corrupt Minister of Health was able to pay for Cambridge Analytica’s services by hiding it in the health budget. Sort of like “selling” the hospital concession to a nonexistent company which produces no results, but is paid millions per year out of taxpayers money…? So the corrupt government leadership in question wouldn’t even pay for such services from their own share of the spoils. They could flip that off on the taxpayer, too.

Christopher Wylie’s testimony is here.