Cambridge Analytica: Ninjas of Neuromarketing (Part II)

Cambridge Analytica: Ninjas of Neuromarketing (Part II)

In Part I of this blog, we introduced Cambridge Analytica (CA), and touched on neuromarketing techniques being pioneered by its parent company, Strategic Communications Laboratories (SCL). To read it first, click here.

A Closer Look at Surveillance Marketing 

We realize that your company doesn’t have the same resources Cambridge Analytica and SCL enjoy, such as “routine access” to top-secret data from the UK’s Ministry of Defense.

But we do have access to data and analytics—lots of it. Chances are, we aren’t optimizing that data. If we started thinking more like Nigel Oakes and Alexander Nix, the CEOs of CA and SLC respectively, we’d probably both be selling more of our services and products. Which is why we’re taking a look at some of their marketing strategies.

Not just because blogs addressing current events test better on our social media analytics (which they do), but also to help our readers reposition their own, hopefully ethical, social media campaigns.

The Medium Is the Message

How I learned to stop worrying and love the surveillance....

Now Google knows that you’re the type of person who would consider deleting Facebook because of a popular news story. Sigh.

Many marketing techniques used by CA and SCL are similar to those you use yourself, such as targeting Facebook ads to specific audiences. For example, we’ll promote this blog to people who care about social media marketing, growthhacking, psychology, and politics. That’s part of cluster marketing, segmenting your market into different groups. It was first described in 1986 by Richard D. Czerniawski.

Until Czerniawski “discovered” soccer moms, for example, Walmart thought the same women buying kids’ soccer equipment were also purchasing yoga products. It’s all athletic equipment being sold to women in their 30s, right? So, they would stock those items in the same aisle. When they segmented the market and began collecting and crunching the data, however, they realized that two very different groups of women—with very little overlap—were buying kids’ soccer gear and yoga mats. They redesigned the store accordingly, and sales improved across the board.

“If you’re collecting data on people and you’re profiling them, that gives you more insight that you can use to know how to segment the population,” explained CA’s chief data officer Alexander Tayler. “[You] give them messaging about issues that they care about, and language and imagery that they’re likely to engage with. We used that in America and we used that in Africa. That’s what we do as a company.”

“Persuading somebody to vote in a certain way,” continues Nix, “is really very similar to persuading 14- to 25-year-old boys in Indonesia to not join Al Qaeda.” Or, convincing soccer moms that it’s worth paying $5 more for applesauce packaged in squeezable single servings, rather than wasting their time with yoga mats that they don’t have time to use anyway.

Do you know your market? How well? Can you name three different segments of your market, and identify some of their motivations and desires? If not, it’s time to get your SCL on, and collect the data you need to get that information.

Psychometric Profiles

Of course I got Rainbow Dash, she's the BEST My Little Pony.

These quizzes make me feel like someone really cares.

You’ve heard a lot about the Cambridge Analytica-funded quizzes that convinced Facebook users to reveal their deepest, darkest secrets to the Trump and Brexit campaigns, just to learn which Lord of the Rings character they are. (I’m Galadriel.) As useful as those probably were, such quizzes are just the tip of the data iceberg.

Your psychometric profile is also aggregated with your credit records, phone patterns, travel habits, Google searches, shopping history, and all the books you’ve bought, songs you’ve listened to, and television shows you’ve watched, using traceable, digital methods. Some of that data is owned by Google, Facebook, and your government, but massive amounts of public data collected over the past decade or so is available to anyone who wants it, thanks to companies like Acxiom.

These hundreds of data points can be parsed out to determine your sex, race, age, nationality, marital status, career, lifestyle, and most importantly to Cambridge Analytica, “OCEAN,” the classic “Big Five” personality traits of Openness, Conscientiousness, Extraversion, Agreeableness, Neuroticism. You can take an online quiz (heh) if you want to know how you rate.

Whether you take (another) quiz or not, several people out there already know how you score. If you’ve “liked” more than 70 items on Facebook, astute algorithms can parse out your scores on all these traits with about 85% accuracy. If you like Nicky Minaj, you’re probably an extrovert; if you prefer Marcus Aurelius, an introvert. This, in theory anyway, allows advertisers tailor propaganda to you. The more items you “like,” buy, read, or watch, the more accurate your psychometric profile will be.

Until 2013, the default setting for your Facebook “likes” was public. Depending on how you’ve configured your security, they’re still available to your contacts. Do you have any Facebook friends who you’ve never met in real life? I do too. Which means that our data is still, for all intents and purposes, public.

Now, scientists don’t put much stock in psychometric profiles. They say that no matter how accurate the data, there’s little evidence that targeted ads could convince you to, say, vote for a certain politician. But have you ever purchased something that you didn’t really need? Or even want? If so, you already know you’re susceptible to marketing. Successful billionaires like Robert Mercer and Peter Thiel, as well as corporations and militaries around the world, are investing in this stuff because it works.

“Robert Mercer did not invest in this firm until it ran a bunch of pilots—controlled trials,” says “David,” an anonymous CA leaker. “This is one of the smartest computer scientists in the world. He is not going to splash $15 million on bullsh**.”

The takeaway is that, despite what some experts think, psychometric profiles probably are an effective way to better target your audience and close the sale. Depending on your company, it may be worth buying some data and preparing profiles for your own clientele.

Growthhacking

The term “growthhacking” was coined in 2010 by Sean Ellis, because he wanted to hire marketing professionals who measured analytics in small increments, and reacted accordingly, rather than designing huge, intuitive ad campaigns. The game had changed, and he was the first to call it.

We, as a species, were suddenly swimming in information, collected by our credit card companies, computer dating sites, and social media addictions, and it was being provided as free, real-time analytics or cheap data aggregation packages to anyone who wanted it. Computers could crunch this data in millions of ways, and your company’s resident growthhacker could test tiny tweaks to an ad campaign, aimed at certain subsets of their target market, in real time.

For example, let’s say you wanted to sell a product (beer!) to men between the ages of 24 and 49. You notice that your Facebook ads with an orange background do better in Tennessee, while blue ads get more clicks in Michigan. You might not even realize that those are the colors of popular regional sports teams, you just know that it works. A growthhacker would change all ads displayed to people linked with Tennessee to orange, with Michigan to blue, and smile as conversion rates improve by 1.4%. #winning

While Cambridge Analytica’s variables were probably different, the process was similar. “Build the wall” might test better in Tennessee, while “bring back manufacturing jobs” might get more clicks in Michigan. If that were the case, CA could immediately begin running the different messages in those markets. They might not even realize why it works, or care, because 1.4% growth means a lot in a swing state.

You could also, say, match people who clicked “build the wall” with their personal psychometric profile. Those who test as “neurotic” could then be targeted with crime-themed messaging on their Facebook feed. You might even find that conflating the two issues, feeding target voters messages that (incorrectly) say undocumented immigrants commit crimes at higher rates than the general population, get an even higher conversion rate, perhaps 1.6%. So you start running those memes more.

In most countries, that’s all totally legal, but is it ethical? Either way, there have been consequences. Take SCL’s 2013 campaign for Uhuru Kenyatta, a Kenyan politician who was under indictment for crimes against humanity in 2012, and polling a distant second in Kenya’s presidential election. SCL crunched data from 47,000 Kenyans and discovered that fear of “tribal violence” was a key data point among potential swing voters. So, they targeted those voters on Facebook and WhatsApp with ads reflecting those concerns. It worked; Kenyatta won with 50.1% of the vote. The downside was that more than 100 people were killed in ethnic violence, and one of their own employees, Dan Mureşan, was found dead in a hotel room.

The moral of this story for your company is that growthhacking works—perhaps better than any of us think. Keep measuring every analytic, crunching every data point, and tweaking every ad. If anything, do it more, because it’s incredibly effective, even when we don’t know why. But at least consider why, especially if your social media campaign might be controversial, negative, or disrupting something you enjoy, such as democracy.

Using Your Friends

Even though I've never met them IRL, we're all such good friends. Right, Ivan?

Cruising the information superhighway with all of your “friends.”

The most effective sales tool in any marketer’s arsenal is a trusted, personal recommendation. It’s also the gift that keeps on giving: Customers acquired through a friend’s recommendation spend 200% more than the average customer, and make twice as many referrals themselves.

More than 90% of consumers trust recommendations from friends and family, while only between 43% and 47% of consumers believe online advertisements. But, 66% of us believe an anonymous online review, even in forums where a verified purchase is not required.

It seems that our brains react to online friends in many of the same ways that they do people in our physical environment. These relationships affect us emotionally, like “real” friendships, and almost certainly affect our purchasing and voting decisions. These online acquaintances can mimic the single most important factor in determining your political beliefs: friendship.

There are perhaps 48 million bot accounts on Twitter; 50,000 of them have been definitively linked to Russia’s attempt to influence the 2016 US presidential election. Facebook thinks it may have around 270 million fake accounts. All of those were created by someone, for a reason. Sometimes, the reason is you.

As a consumer or voter, the lesson is clear: Keep a jar of cynicism close to your computer, and take a swig before reading online reviews or believing anything that anyone on social media tells you.

As a social media marketer, the data is just as clear. You need to consider budgeting for employees to “troll” spaces like Yelp!, Facebook, Twitter, Amazon, Reddit—really anywhere that your clients go for personal recommendations—with positive reviews of your product or service. That’s legal, and there are few effective safeguards in place to stop you.

Disruption Is More than Just a Buzzword

It’s not that we don’t trust social media billionaires to make responsible decisions… OK we kind of don’t.

A lot of the handwringing over the Cambridge Analytica–Facebook scandal is focused on “doing something” that will magically stop the propaganda altogether.

Good luck with that.

Sure, better regulation of our personal data would be great. For instance, the EU now requires people to opt in to share data, while US users must still navigate deliberately byzantine UX nightmares to opt out of sharing theirs. That single regulation would frustrate many annoying data collectors.

But regulation won’t stop some of these people. Regulations didn’t keep Alexander Nix from asking Wikileaks for data, and they couldn’t keep the Mercer family from investing $15 million in a company that had “routine access” to classified data from the UK’s Ministry of Defense. Or stop Putin from doing whatever he’s (allegedly) been doing. These propagandists are some of the wealthiest, smartest, most powerful people in the world. They’re going to be able to work around regulations, and it would be naïve to think otherwise.

So, what we want is for our readers to stay educated, even as you use SCL’s neuromarketing techniques for your own (hopefully more ethical) ends. And to remember that as social media marketers and growthhackers, you’re more familiar with these technologies and strategies, and their potential for abuse, than almost anyone else.

Stay on top of it. Keep up with cutting-edge digital marketing. Tweak your message slightly for a .2% growth rate among older, male “My Little Pony” aficionados. Build psychometric profiles for them. Target them on their favorite subreddits and Facebook pages with faux social media accounts claiming Derpy Hooves is the best little pony. Disrupt major players in the fedora market with your online marketing strategy. Pay attention, learn every lesson, and do your thing.

But also remember that in today’s political struggle—World War WWW, if you will—you just happen to be on the front lines. Sure, you just want to sell veggie burgers or a dog-walking app, you’re sick of hearing about politics. But if you’re marketing anything online, you’re operating in the single most hotly contested battleground of the age. So, stay alert, and to paraphrase Homeland Security, if you see something, tweet something.