Data pricacy is a hot issue, especially with the new European GDPR affecting the entire world. (See: GDPR is not a compliance but a marketing issue. Companies like hotjar who track your mouse movements on websites are already in hot water. But the real question is what the behemoth of social media, Facebook, will do.
In several countries parliamentary inquiries are under way to find out how Russia and other dodgy players affect local elections. At the core of this issue is the ability to target Facebook users with precision.
Precise targeting depends on what Facebook knows about its users. Facebook is notoriously secretive about what it knows and how it uses the information, not least because it really does know a lot; too much, some data privacy activists claim.
Facebook's business model
Facebook has a very clear business model: it wants to know every about you, so it can sell this information for the purpose of precision-targeted advertising. That means it has to keep you on the platform as long as possible and gather as much information as possible.
For many people, Facebook is already equivalent to "the Internet". In countries like Taiwan, Facebook has a penetration rate of close to 90% of the population, with an average of 3.5 hours a day spent on the platform. Which means that a large percentage of the population get their news from Facebook. Sounds scary?
Facebook is doing everything to make itself relevant. It shows you the news and the weather, prioritizes information sources you want to see, lets you play games, interact with friends, shop for products, and find a new job. It is continuously launching new products with the sole goal of gathering even more information about you.
But what does Facebook really know about you?
Some of the information Facebook has on you is pretty obvious. It knows your email address and your phone number, and, through GPS and Wifi, your location. If you filled in your Facebook profile truthfully, it has data on your education and your profession.
That basic data set alone is often enough to target ads and even identify you. If you live in a small town and your profession is "baker", you are probably identifiable even if you chose a pseudonym on your Facebook account. (How many bakeries are there in your village?)
However, it's really once you start using Facebook that the platform gathers the most valuable data. Let me stress this: the biggest value comes from data you don't turn over to Facebook consciously.
Because it knows your location, Facebook knows your movements. I go from place A (my home) to place B (my office) almost every day at the same time. That means Facebook knows the absolute best times to show me those ads for breakfast joints or dinner restaurants.
By the same token, Facebook knows when and where you go on holiday. It knows the class of hotel you stay in and thus probably your income or at least your position in a company. Try staying at a 5-star hotel a few times and you'll soon see Jaguar and Gucci ads.
According to Facebook terms, the information is gathered from the "family of Facebook apps". That means even if you avoid using Facebook on vacation, it can correlate that data from Whatsapp (owned by Facebook), and of course Messenger, Instagram, and whatever else it will buy or create in the future.
Because it owns different apps and knows your usage patterns, it can target ads even more precisely. Many people don't turn on Facebook during the day but use Whatsapp for business communication. That means Facebook knows your working hours, and thus very likely your profession.
Someone using Whatsapp exactly from 9-5 every day is hardly going to be the CEO of a company, regardless of what you put in your profile. If you use Facebook at exactly 5-7 pm every day, with changing GPS data, machine learning algorithms are inferring that you can't be reached during the day but have a long commute.
If you stop at the same pub every day at 6, the same hospital every two days, or a Church three times a week, Facebook can tell that you are either an alcoholic, a diabetic, or a religious fanatic.
Facebook even knows if you are a smoker or not: If you move outside your building and spend 5 minutes on Facebook every 2 hours, you probably left the office for a fag. And if you move between home and the local pool every morning, it's time for Speedo ads.
Check-ins help Facebook confirm your intentions
Of course location data isn't always accurate: my pool is part of a hospital complex, so I could equally be going for medical treatment, or work at the hospital. This is where the "check-in" feature comes in.
Check in is like saying to Facebook "yup, I confirm that". The app already knows you are at a restaurant (GPS data), but if you actively check in, that data point gets a lot more weight. If GPS shows me at the aforementioned hospital-slash-gym, checking in at the pool tells Facebook that I'm not visiting my doctor but going for a swim.
The same is true for every group you join and every page you like. Once you share a post, Facebook is even happier. Sharing is more of a commitment than Liking or reacting to a post in another way. You are not alone.
Of course, Facebook also considers the people you are connected to. If most of your connections are gay men, it will fairly assume that you are gay too. If two-thirds of them like gyms, the algorithm will rightly assume that you are either a stalker or a gym rat. If you spend 3 hours at the gym, and regularly post pics of your buff torso, Facebook (GPS data + image recognition) will show you different ads than if you are 60+ and only spend 30 minutes in that location.
Likewise, if all of your friends continuously share posts about anti-abortion rallies and check in at a church on Sundays, Facebook can infer that you are a conservative redneck even if your profile is empty and you've never shared or liked a post in your life.
All that data
All in all, Facebook takes about one hundred data points to put you into specific categories. You can check your own by opening Facebook on your computer and clicking on the triangle in the tool bar. Select "Settings" and the "Ads" on the left. Now scroll down and click on "Your information". Under "Your categories" you will now find your full "advertisers profile".
Be prepared to be shocked. What Facebook shows you here is the bare minimum of what they have to disclose by law. It does not include all the inferences the algorithm made by correlating various data points, image analysis, and machine learning results.
If you find yourself in the category "parents" and "frequent international traveler", they will assume that you will be traveling with children. If you are connected to your spouse and children on Facebook, the platform will know whether this time you are traveling alone or not.
Although I have never actively told Facebook my political views, they have me down as a "Liberal". Must be all those Anti-Trump posts I like to share. There is a lot more to explore on those pages, for example, "Interests" and "People". My list includes Barack Obama, Ray Kurtzweil and Amy Schumer, and a whole list of people I honestly have never heard of. But overall, the data is pretty accurate.
And there is more...
Facebook Pixel is another way Facebook collects data about you: outside the Facebook app. But I am sure by now you get the picture: most of the data Facebook has about you is not data you actively provided or can control.
The smarter machine learning and artificial intelligence will become, the less your privacy. It's nice that Facebook lets you select which information can and cannot be used to target ads, but so far no one has forced Facebook to delete private data, not to mention inferred metadata.
More worrying is the use of Messenger or WhatsApp data for targeting. Users have alleged that after chatting with someone about a specific product or topic, they have started to see relevant ads. While most people will accept that Location, Likes and Shares can be used for targeting, I don't want Facebook to listen to my private chats. If I am telling my significant other that I am "going to buy you some underwear" and subsequently see Calvin Klein ads, I will be seriously pissed.
The more the merrier
Now imagine if Facebook did admit to buying external data, like credit reports, loyalty programs from stores or airlines (they deny that). Facebook would then be in a position (which, again, they fervently deny) to show you ads for milk when you are about to run out (based on your purchase data).
The more data they have, the more correlation is possible, and the more accurate the ad experience, and the more worrisome the privacy issues.
According to its terms of usage, Facebook can do whatever it wants with your data. This is exactly the point that brings it in conflict with laws like GDPR and privacy activists. I don't want to see an ad for bread after I send my spouse "don't forget to buy bread" on WhatsApp.
The more Facebook is involved in political processes like elections, the more it will feature in discussions about digital citizenship and digital democracy, meaning that sooner or later, its domineering approach to data ownership and usage will cause its downfall.
Some governments will ban it, others will force it to limit the amount of information it gathers and how it uses that information. It would not surprise me if sooner or later Facebook will be treated as a public utility. Whatever the future holds, the challenges are real, both for users and operators of social networks.