So, why do brands value this data? Well, simply put, it allows advertisers to reach the people they think are most likely to be interested in hearing from them. It's why you sometimes have that sense of déjà vu.
For example, advertisers can decide to show ads only to people who have shopped on a particular website before but left before clicking “buy”. In the media industry, we call this 'retargeting'.
Advertisers know the cookie IDs of users who have come to their website or can buy that information from another organisation and then advertise only to those users. More often than not, this is all done by an automated auction process. Again, in media agency terminology, we call this 'real-time bidding'.
The website that has an advertising space that needs to be filled will send information about the user (from the cookie ID) and the page where the ad would run to an online advertising exchange, where advertisers decide whether they want to bid on that particular space or slot, the advertiser, in general, would usually bid more if it is a user who has shown interest in their product in the past. The entire process happens in a fraction of a second, and that is how ads appear to read your mind and follow you around the internet.
Now, as people who work in media planning and buying, we're often told that this is a bit spooky and in general, consumers don't like to think of themselves being trading as a commodity in this way, but advertising technology is getting more sophisticated.
Not only can brands reach a certain type of user, but they can also serve ads that are more relevant. If you've looked at a pair of jeans or a style of jeans on a retailer's website, that retailer might put a photo of those jeans into the next ad that you see next.
In the Summer, a fashion retailer might show t-shirts and shorts to users in the UK but might show chunky jackets to users in New York. Advertisers also have a lot more control over where their ads appear, what time of day or week and what devices they want to serve ads to.
Advertisers can infer income, for example, from what sort of device or operating system a consumer has (people with Apple computers tend to be richer than those with PCs.).
Advertising is not an exact science, well, at least not yet, but it is definitely becoming more of one.