A man browsing digital advertisements on his phone
Stacker Studio

How AI predicts what you’ll buy

Written by:
April 29, 2022
Stanisic Vladimir // Shutterstock

How AI predicts what you’ll buy

It’s a jungle out there—few places so much so as the world of “smart” advertising.

There, marketing geniuses have developed increasingly sophisticated algorithms that take all the information gathered about you online or from your phone and piece together a customer profile that may include everything from your favorite pair of socks to your children’s names.

Analyzing current market practices, Wicked Reports explored how artificial intelligence, or AI, can be wielded to gather data and make sales predictions across the internet. Some techniques you may know, such as persistent cookies that turn your computer into a ping hub for the websites you visit. Others are much more sophisticated, compiling all of your characteristics by analyzing what you’ve bought in the past, what you’ve put in your cart and abandoned, and what you’ve searched for. From there, advertisers can even make a version of similar customers to market to them as well.

The digital advertising industry is expected to crest $20 billion in 2022. That’s far from enough to crack the top 10 biggest industries in the U.S., but it’s a substantial amount of money—particularly when compared to the big-ticket ad buys of the past in splashy magazine spreads. Companies today are more eager than ever to spend what it takes to bring in ideal customers.

Continue reading to discover some of the tactics AI uses to predict buying behaviors.

A person prompted on their mobile phone to accept cookies in a browser on a website
1 / 5
picture alliance // Getty Images

Compiling user movement across the web

You may know about cookies: tiny text files that websites deposit on your computer as a way to track online behavior.

When you visit websites from Europe, for example, a law there mandates that you click through a cookie agreement that’s much more transparent than in the U.S. There are session cookies lasting one browsing “session” (until you restart your computer or browser) and persistent cookies that stay until you delete them. Think of a cookie as a waving arm each time you visit the same website. Together, they form a heat map of how often and when you visit every website in your browsing history. They can even flag your presence to other websites as a way to combine your data.

A bulldog mix puppy being held by a young man in a plaid shirt
2 / 5
Anna Hoychuk // Shutterstock

Identifying user characteristics

User characteristics, and something called demographic segmentation, is a key way online advertising targets you. User characteristics are any of your qualities, from your gender and age to what car you drive and the pets you own. These user characteristics lead to the advertising concept of demographic segmentation, in which companies can buy lists of really specific people.

Are you a 25-year-old white man with one dog, a full-time job as an auto tech, and an apartment rental in a “transitional” neighborhood? We have just the plaid shirt for you.

A man driving a car and holding a smartphone open to a GPS application
3 / 5
mhong84 // Shutterstock

Mapping user location data

If you’ve used GPS in your smartphone or any of the hyperlocal dating apps, you’ve leveraged location data to your advantage—at least for now.

How does your phone know where you are? Cellphone towers ping your phone when you’re nearby.  In your home, your Wi-Fi network is likely hardcoded with your location. That’s also true of any Wi-Fi network you hop into or onto during your errands, at school, at work, and so forth. After that, GPS can pinpoint your phone to an alarmingly small area as you carry it around, so not just in your home but in one corner of one room. 

A man shopping for a jacket online
4 / 5
Gorodenkoff // Shutterstock

Matching new users to known customers who look and act in similar ways

Some items on this list are not very surprising, or we’re used to being told about them so they don’t seem as insidious and scary as they once did. But people are likely still surprised by the depths that companies will go to in order to better advertise to you. Your favorite clothing store, for example, might put together a complete data “picture” of you: what you’ve purchased from them, what size you shop for, where your address is, and more. Then they can reverse engineer someone just like you and buy a demographically matching list.

Anything can be filtered until just the exact desired customer base remains, and then they buy the ads.

A computer screen with a highlighted IP address
5 / 5
Habichtland // Shutterstock

IP address targeting by network connection

How much do you know about your IP address? Many of us are old enough to remember a time when connecting to the internet required knowing a specific IP address and typing it into our PC settings.

Today, the router you likely have in your home has a hard-coded IP address whose number values reflect where you are as well as which “node” you have on your local network. That information may be for sale to different companies because, with the right technology, they can use some IP addresses in order to infer the rest—and guess where you live. Apple is among the tech companies pushing back on IP targeting of this nature by masking IP addresses in its proprietary browser Safari.

This story originally appeared on Wicked Reports and was produced and distributed in partnership with Stacker Studio.

Trending Now