How Facebook Is Listening To You

Inside the Facebook targeting machine

Jordan Rooney
Sep 10, 2019 · 10 min read
Visual credit

Originally published at on September 10, 2019.

You’ve heard the stories, maybe you’ve even experienced it yourself. You’re having a conversation, open Facebook, and bam! An ad for that product you just talked about is right there. Of course, it seems like Facebook is listening to your conversations. But is that actually the case?

Let’s get this out of the way, Facebook is not listening to your conversations. It doesn’t have to. It has something much better.

Well, they might be, but it’s highly unlikely. If you want to be extra cautious, go into the privacy settings of your phone and disable access to everything such as your microphone, location, camera, etc for the Facebook app (and Instagram, Messenger, etc.)

However, the reason ads pop up in your feed isn’t because Facebook is listening to your conversations.

The real reason it seems like Facebook is listening to you is that it has data on you. Heaps, mounds, disgusting amounts of data on you and everyone else on its platform. It’s impressive.

A printed version of the data Facebook has on you (I imagine)

We’ll get to exactly how Facebook already knows more than it ever would listening to your conversations, but first:

The Real Explanation

We see 5,000 ads per day. We’re spending more and more time looking at our phones. Out of every minute spent on our phones, 30 seconds of it is on social media.

Remember when Facebook had no ads? Those were the days. But for better or for worse, Facebook and other social platforms now command a huge portion of our attention. This is more valuable than gold for companies looking to sell their products or services.

The end result is that we’re being bombarded by countless ads every day. Most ads we don’t pay attention to and just scroll past. Some we may briefly glance at and then ignore. And then there’s that one that catches our attention. That thing we were just talking about and didn’t search for or look for and there’s no way Facebook would know about it. Except that it does. Facebook already knew. They know what you’re likely to buy better than you do. More about that in the next section.

What’s at play when we think Facebook is listening to our conversation is cognitive bias. We assume that they’re listening because that one, or a few ads, showed up right after we were talking about something. We forget about the thousands of other ads that showed up for other reasons. We pay attention to the one out of a thousand that fits a conspiracy theory when in reality it was the one in a thousand that happened to show up at the creepiest moment. It boils down to luck and the fact that if we’re shown a large number of ads, one will stick out, even if they aren’t targeted to the level that Facebook allows them to target to.

How It’s Done

It’s been shown that Facebook collects up to 52,000 data points on every user. What about if you don’t have a Facebook profile? In that case, only 1,500 things about you are being tracked.

Facebook collects data on you, your friends, their friends, the anti-vax group, tree-huggers, crazy gun-nuts, everyone.

The result is a full view of what humanity is like. All the different personalities, temperaments, interests, desires, etc. you get the picture.

A less nerdy, but more nerdy at the same time way to view this is by looking at the graph below.

Imagine all the different personalities of the world fall somewhere on this graph

Imagine this is all the different people out there in the world. They all fit somewhere on that graph.

On the top right you might have an innocent twelve-year-old who’s interested in pink hats and the new Toy Story movie. As far away from this person’s profile as you can get on the bottom left of the graph is a person probably interested in committing a mass shooting or some other unspeakable crime. In the top left, you might have a capitalistic banker scheming of ways to commit tax fraud so his company’s stock will go up. On the opposite end in the bottom right might be a hippie who spends 12 months a year recycling used water in order to save the planet.

On this graph is every person imaginable. You probably don’t fit into one of these categories exactly, but you do fall somewhere on the graph, and that’s the point. Facebook knows exactly what type of person each individual point on this graph is.

When you think about one billion people on Facebook, and the tens of thousands of data points Facebook has for each person, that’s a ton of data to try to be able to piece together what the world is like.

What this means is that of the billions of different people out there, Facebook creates a spectrum of what all those possible people are like.

Why Ads Get Creepy

The reason Facebook is so good at targeting you with ads though is not because it collects so much on you. It’s because it collects so much on everyone. It’s an exponential effect.

What Facebook does is not in analyzing the data. It’s in analyzing the metadata. Metadata is data about the data.

Think of it sort of like this:

You and Jeff create data, data creates metadata

How Terrorists Are Stopped

This is actually pretty similar to the way the NSA monitors internet traffic. When analyzing who the next bad guy will be, it’s not one google search that sets off the alarms. It’s things like: what other google searches have you been doing? Who have you been talking to? What type of searches have they been doing? Where have you been traveling? What type of things have you been buying? And thousands of other criteria — It’s giving context to everything that happens.

The NSA and CIA learned that it isn’t the content of a message that is all that valuable. It’s the data about the data, the metadata, that is much more valuable. In principle, when you have a lot of information, getting a more broad view of what’s going on, through the context metadata gives you, is much more valuable in predicting the next terrorist plot.

It turns out that this framework of looking at huge mounds of data isn’t just good at stopping terrorist plots, it’s pretty good at selling funny costumes for cats too.


The same way that the FBI or the NSA might look at your entire online life instead of one google search to see if your asking for trouble, Facebook also looks at you similarly. Sure you liked a funny cat video, but that doesn’t mean you own a cat, and it certainly doesn’t mean you’ll buy any cat products.

Imagine you own a cat though. Facebook knows because it can tell what’s in your pictures (this includes Instagram too). It has thousands of data points on you. It also has data on millions of other cat owners. This is the scenario Facebook ran into and what happened next is they came up with a $500 Billion use for it.

What happened was Mark Zuckerberg and the crew at Facebook went, “Hey! We have all this data on Sally the cat owner. We have all this data on a million other cat owners too. What if we pulled out a fancy computer and tried to see how all these cat owners are similar on that nerdy graph earlier in this post?”

The result is that across millions of different cat owners, Facebook is able to tell what type of person the typical cat owner is to a staggeringly accurate degree.

This is the power of metadata and exponentials. If Facebook knew things about just one cat owner, that doesn’t help much. But millions of cat owners? You’re able to know a cat owner better than their best friends know them.

This Is How Ads Get Creepy

If we only have one cat owner to draw data from, the creepiness (targeting ability), of ads goes up like the left chart (A), if we have what is the current situation, with millions of different cat owners, the creepiness levels go up more like the right chart (B).

Linear (A) vs Exponential (B)

Of course, this applies to every ad for every category on Facebook. It can apply to Viagra just as much as it can apply to a t-shirt.

What results is ridiculously effective targeting for almost any product under the sun. It’s also the reason we get that one-in-a-thousand chance of an ad popping up right after you talked about it. The targeting is just really, really good.

One More Layer

A lot of companies will practice what’s called uploading a customer file to Facebook. This lets them target you even better than the simple interests based targeting that you’re familiar with seeing when you’re browsing the web and an ad pops up on your Facebook feed.

Take the cat costumes example from above. Imagine you bought a cat costume and entered your details such as email address etc. The cat costume company will then turn around and upload these details from you and a thousand other customers, and Facebook is able to link this back to your individual profile, marking you as a cat costume buyer. Since they have a profile of you as well as a thousand other cat costume buyers, they’re able to get a very, very good picture of what exactly a cat costume buyer looks like by canceling out all the noise across thousands of different buyers.

This creates what’s called a “custom audience,” and includes all the aspects of that type of person common to the 1,000 people who have already proved they’re interested in cat costumes by buying one (or seven).

The company will then say to Facebook, “Okay Facebook, I told you which people have already bought cat costumes. Go find the 1% of the population that most closely fits these same people based on those 52,000 data points you have on everyone.”

This is what’s called a lookalike audience. It’s that weird feeling you get when you only think about something, and next thing you know there’s an ad for that product in your Facebook feed.

The Zuck’ Is Watching

But wait, there’s more! Not only is the data and tracking happening while you’re on Facebook or Instagram, but it’s also happening all over the web on millions of individual websites. We’ve all experienced browsing around the web, searching for whatever it is we’re searching for, flipping to Facebook, and there it is. An ad for the exact item we were just searching for.

This happens because Facebook has all types of sneaky trackers all over the web. Many are embedded right into websites and can even record where you mouse moves or what your scroll past, or spend a lot of time looking at.

Due to the trackers embedded on your phone and these websites, everything you do can then be linked back to your individual Facebook profile. So the tracking happens even when you’re not browsing Facebook.

Putting It All Together

By combining these things, Zuck has built a multi-billion dollar empire on collecting, analyzing, and brokering your data.

Add everything up, and what results is a surveillance system not unlike (in concept) what national security programs use to monitor the world and thwart terrorist or other attacks. The difference is they’re not trying to stop lives from being lost, they’re trying to sell more squeaky dog toys. It’s incredibly effective, and incredibly profitable (both for Facebook and for advertisers).

If this is a little much and you have some concerns, you’re not alone. Many people are wary of a private company having so much power over such a large portion of the population. If you fall into this category, there are a few things you can do.

What Can You Do?

This goes without saying, but if you’re concerned, either don’t use Facebook or use it less. Short of deleting your Facebook account there are a few things you can do:

1. Use a Privacy Conscious Browser

Firefox is a great mainstream browser that arguably has the best protection while not sacrificing any conveniences. Safari has made some measurable improvements as well but still has a ways to go. Brave is an excellent option for a light browser that blocks all ads and trackers. It also has a cool feature that shows you how much time you’ve saved by not having to load ads and trackers (I used it for a couple of months and saved 2.1 hours).

2. Install an Extension That Blocks Cookies and Trackers

Privacy Badger is great for blocking anything and everything that may be tracking you online with little to no impact on your usual browsing. There are others like Ghostery, but Privacy Badger seems to be the consensus pick.

3. Use a VPN & Delete Cookies

Cookies are the lifeblood of tracking you online. Deleting them might sign you out of websites, but it will speed up your browser as well. Using a VPN is something you should do anytime you’re not using your home WiFi anyway. VPN’s are very important to keeping things secure when you’re doing online banking, entering passwords, or really doing anything with an account online since nowadays, one point of weakness can mean disaster for your entire online life.

Wrapping Up

Look, Facebook isn’t listening to your conversations. If you’re still paranoid, there’s lots you can do. If you want to read up on other measures you can take the Electronic Frontier Foundation is a great resource. While Facebook isn’t listening to your conversations, it is watching you in a very intrusive way. It’s just more roundabout than you originally thought. Facebook is listening, and watching, all of us, whether it’s through the way we expected, or not.

If you’ve read this far, don’t forget to share this on Facebook.

Originally published at on September 10, 2019.

The Startup

Get smarter at building your thing. Join The Startup’s +799K followers.

Medium is an open platform where 170 million readers come to find insightful and dynamic thinking. Here, expert and undiscovered voices alike dive into the heart of any topic and bring new ideas to the surface. Learn more

Follow the writers, publications, and topics that matter to you, and you’ll see them on your homepage and in your inbox. Explore

If you have a story to tell, knowledge to share, or a perspective to offer — welcome home. It’s easy and free to post your thinking on any topic. Write on Medium

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store