“I hate advertising!!” You hear people say this all the time. But they don’t really hate advertising. What people hate are irrelevant marketing messages.
Consider, for instance, how crudely TV ads are targeted. Tune into CNN for an hour, and you’ll swear that the only people watching are 70 years old and suffer from a lengthy list of unfortunate medical afflictions, with ad after ad pushing cancer drugs, class-action mesothelioma lawsuits, erectile dysfunction aids, and walk-in bathtubs for the elderly.
Turn on a football game, and you will see ads targeting one demographic segment: young men in their 20s. Ads for beer, macho trucks, Hooters, and horror movies dominate every commercial break. Which can make it tough for families wanting to spend Sundays together enjoying an NFL game. Try explaining the blood-curdling screams of a hapless teenage girl being chased by an axe-wielding psychopath to your six-year old boy… “Opens this Friday at theaters everywhere.” And never mind that women make up an estimated 45% of the NFL’s 150 million American fans. There are plenty of women who get together for football parties on Sundays, drinking wine — not beer — while rooting for their favorite team. Are they really interested in seeing ads touting the towing capacity of a crew cab V8? (Netflix already knows how that time can be better used.)
And what about guys watching a romantic flick on the Lifetime Movie Network? Do they really need to see a feminine hygiene products fly across their 60″ HD screen while hearing about “absorbancy” and “freedom?”
Clearly targeting sucks in traditional media channels like TV, print, radio and outdoor, but things are changing. In the not-too-distant future, consumers won’t have to see another irrelevant ad ever again. Thanks to big data, predictive analytics, artificial intelligence, machine learning and psychometric modeling, marketers will be able to target consumers with pinpoint precision.
In fact, it’s already happening. Trump wouldn’t be sitting in the Oval Office today had it not been for pyschographic profiling. And Facebook.
An OCEAN of Data
Psychometrics (also called psychographics) uses psychological traits to build marketing segmentation models. OCEAN is one such model. Based on the following five dimensions, psychologists have figured out how to make accurate assessments of people — their needs, fears and behavioral propensities:
- Openness – How open you are to new experiences?
- Conscientiousness – How much of a perfectionist are you?
- Extroversion – How sociable are you?
- Agreeableness – How considerate and cooperative are you?
- Neuroticism – How obsessive and/or anxious are you?
The creators of the model could prove how effective and reliable it was, the only problem was getting a hold of the data. Marketers could only gather such information using printed paper surveys — cumbersome and time-consuming. And people living in a world prior to social media had a different sense of privacy and propriety, so it wasn’t easy getting people to participate.
But then came the internet. And Facebook.
All of a sudden, consumers were surrendering huge amounts of personal data, creating a vast digital goldmine with every Google search and Facebook “Like”.
By 2012, one researcher named Michal Kosinski figured out he could predict a user’s skin color with 95% accuracy, and he could do it just based on 68 Facebook “Likes”. He then figured out how to determine sexual orientation with 88% accuracy, and their political affiliation with 85% accuracy — again, using fewer than 100 “Likes”. But it didn’t stop there. Kosinski could tell what religion you were, whether you smoked or drank alcohol, and even whether your parents were still married or divorced.
Kosinski continued refining his models to the point where he was able to understand a person better than their average work colleague using no more than ten Facebook “Likes”. Seventy “Likes” were enough to decipher more about a user than what their friends knew, 150 what their parents knew, and 300 what their life partner knew. At around 600 “Likes”, Kosinski could even tell more about a person than what that person thought they knew about themselves.
According to Kosinski, the day he released his findings, he received two phone calls: one threating him with a lawsuit, and another offering him a job. Both calls were from Facebook.
Then only a few weeks later, Facebook “Likes” became private by default. About a week later, Facebook launched its IPO. Today the company has a market cap over $500 billion, entirely predicated on its ability to deliver highly targeted ads thanks to the world’s most powerful — and now proprietary — warehouse of psychographic data.
Read More: What Is Your Bank’s Big Data IQ?
Facebook + Data For The Win
Cambridge Analytica is one firm that subscribes to the power of the OCEAN framework. This data mining and analytics company has been using Facebook as a tool to build psychological profiles that represent some 230 million adult Americans. And they claim to have as many as 3,000 to 5,000 data points on every one of them, gleaned from the digital footprints of users as they pop around the web — from Facebook to Google to ESPN to Fox News.
You may have heard of Cambridge Analytica. Why? Because they claim to be the masters of dark, data-driven digital marketing arts — yes, the folks who say they got Trump elected. Brad Parscale, the digital director for Trump and the rest of the Trump campaign apparently deny Cambridge Analytica’s claim, but it really doesn’t matter. Whether Trump used Cambridge Analytica’s OCEAN models or just let Facebook do the segmentation legwork automatically, the underlying premise is the same: data is all that counts.
While other political campaigns took the more traditional route with the predictable TV attack ads, Parscale poured almost all of his $90 million budget into Facebook. People anywhere could be targeted with the messages they cared about, like infrastructure, so Parscale would make ads that showed a bridge crumbling.
“I could find 1,500 people in one town that cared about infrastructure,” Parscale told Leslie Stahl on 60 Minutes. “Facebook lets you get to 15 people in the Florida Panhandle that I would never buy a TV commercial for. That’s micro targeting.”
Parscale said the campaign would run 50,000 to 60,000 different ad versions every day, some days peaking at 100,000 separate variations — changing design, colors, backgrounds and words — all in an effort to improve response rates and engage users.
The very fact that Trump is sitting in the White House today should serve as testament to the marketing potency of data, particularly when coupled with digital marketing and machine learning, Parscale contends.
“Twitter may be how Trump talks to the people, but Facebook is how he won,” Parscale says.
Get On The Digital Data Bus, Or Get Left In The Dust
Marketing used to be about leveraging focus groups and the Four Ps to carefully craft a singular message that would be broadcast to thousands (if not millions) of people. But such an approach almost seems quaint in today’s world. Any marketer that doesn’t consider themselves a data scientist today might be a dinosaur tomorrow.
“The first people who will lose their job because of AI will be marketing managers,” warns Dave Snyder, Executive Creative Director of new media company Firstborn, in an interview with AdWeek. “If your job is really to move numbers around a spreadsheet and optimize your media budget based on what’s performing, the computer is going to be way better than you, and faster.”
As the publication MarTech explains, artificial intelligence now allows marketers to drive campaign performance with unimaginable efficiency, through relevant and personalized advertisements that leverage enhanced targeting algorithms and creative optimization. And MarTech columnist columnist Carl Erik Kjærsgaard says AI learns differently than humans do, eliminating biases and human errors, especially when it’s applied to the media-buying process.
“AI learns from people’s past what media combinations affect them, and by how much,” he explains. “It uses algorithms to distribute money toward the most effective combination of media buys. And through machine learning, it knows from past campaigns what combinations work best.”
Digital media firm AKQA is built a tool using IBM Watson that scours online platforms to identify unexpected potential consumer segments for brands. Ad agency JWT in Canada is working with “programmatic creative” to change digital ad creative on the fly, particularly regional and pricing information. And Saatchi & Saatchi LA has been running AI campaigns for brands, including a Facebook campaign for Toyota using IBM Watson to yield a thousand different ads based on people’s interests.
Marketing now boils down to one word — data — and the weaponry used in modern marketing warfare is changing quickly. It’s only a matter of time before bank and credit union marketers figure out how to exploit their financial institution’s data using AI tools.
And then just imagine what could happen if consumers’ financial data could be coupled with Facebook data. Oh, the possibilities…