- Okay, I know people say, "Oh, your phones aren't listening to you, but we were just talking about retro video games and here is an ad for home mini arcade machines, which looks kind of cool, but how did it know we were just talking about that?

Is my phone listening?

Is it a psychic?

What's happening?

- People always say, "I was talking about this product and then I saw an ad for it.

Matt, are they listening through my phone?"

And I'm like, "Well, they didn't hear you.

They didn't listen to anything.

The truth is actually more frightening."

You talked about that thing because they influenced you into having this conversation.

It's because the algorithm took you to this place, and that's the situation with our data.

- [Host] There have been some outlier examples of ad tech companies listening without consent, but that's against the law.

For the most part, advertising algorithms know you so well that they can predict what you would find interesting.

They're so good because they've studied all the data gathered about you from the data brokers.

- There's this kind of dossier that's being created about you as you're wandering around the internet and it's being used to decide what ad you see.

- [Host] Advertising fuels the economics of the internet, and advertising is fueled by you, me, us, our personal data.

So how do the algorithms actually work?

- You land on a webpage and that webpage gets your IP address, a unique identifier's return to a data broker.

That data broker looks up all the facts that were ever gleaned about you, a dossier that describes you.

- [Host] And then advertisers literally bid, they bid on you in a real-time auction based on the only lesson I remember from Econ 101: supply and demand.

- The demand-side platform that the advertiser is on says, "I have here an 18 to 34-year-old man child with an Xbox in Southern California.

Who wants to pay to cram something non-consensually into his eyeballs?"

And that takes place in a marketplace.

And you have advertisers on the other side who have standing orders.

They're like, "I want to advertise to 18 to 34-year-old man children in Southern California who own Xboxes."

And the marketplace conducts an automated high speed auction where it's just like, I'll pay so many fractions of a cent, I'll paid that plus 10%.

The high bidder gets to show you an ad.

- [Host] All of this happens nearly instantly, and it's possible because data brokers sort and cull personal data into super detailed consumer groups.

This is the process that delivers relevant, personalized ads for cute shoes or that new tea club, if you're into that sort of thing, which all the ads seem to know about you.

- [Haley] Here we have, for example, soccer mom.

- Okay, I might be able to identify with that.

A sporting good retailer could then go and check out a file on what a soccer mom's online activity is like.

- What are the things that they tend to spend money on and do they like to save?

Are there certain times of year where they might spend more or less?

- [Host] With all this data about soccer mom's consumer habits, retailers can send personalized ads with uncanny timing.

Some people like the results.

Soccer moms get deals, advertisers keep websites we love in business, but some consumer categories can be more concerning.

- Sometimes they'll have categories that are like, this person has erectile dysfunction, this person has a gambling problem.

This is a list of kind of people that are likely to fall for frauds.

So these can be really powerful and damaging lists.

- So this one, for example, diabetes interest.

- [Host] Oh, okay, healthcare demographic.

- The thing about that is like when you're thinking particularly about health data, that indicates that you're probably in a pretty sensitive situation.

- I'm now thinking about all the patients I take care of, and I shudder to think at the targeted ads that are deliberately targeting these individuals.

(text typing) - Data brokers have lists about sexual assault victims.

They have lists about dementia sufferers.

Those are some really concerning types of categories, and you could definitely see how an advertising company could take advantage of people on those lists.

(pensive music)