Why can targeted advertising be smart enough to show me ads at home for something I searched on my PC at work, but not smart enough to not show me the same ad 10x in a row or for services for which I’ve already subscribed?
In: 128
There are a few good answers here and a few slightly dodgy ones, so I’m going to do my best to try and answer this on a technical but also theoretical level. Source: I was a digital media planner at one of the big agency network for seven years.
First thing to get out of the way is that on a simple, technical level, it is entirely possible to stop serving ads to someone who has purchased a particular product. The challenge is that the conditions are rarely in place to allow this to happen, and I’ll explain why.
Firstly, how do advertisers know anything at all about you? Well, they use tracking. Tracking is a general term for anything that keeps a record of your online behaviour but the one you’ll all be familiar with is cookies. When you visit a website and accept cookies, there’s a likelihood that there will be an advertiser cookie in there. This is a file that sits on your computer and waits there to record information. It can only keep a record when it sees a bit of tracking code that it recognises (confusingly called a ‘pixel’). So if I have an ‘X’ cookie on my computer and it sees a ‘X’ pixel on a given website, it notes that down. The general point is that you have something that is tracking (in this case, the cookie) and something that is being tracked (in this case, pixels). Cookies are not the only tracking technology, but this will do for now.
Ad networks who have lots of cookies and pixels on lots of sites will be able to gather quite a lot of information about you. That’s why Google know an awful lot about what you go and what you are looking at; cookies can gather very detailed information about what specific pages you looked at, and from this they can start to build a picture of what interests and shopping categories you fall into.
Note the wording of that last sentence: “build a picture of what interests and shopping categories you fall into.” I phrase it that way because these networks are not building profile of you specifically as an individual. That’s quite a pervasive myth, but PII (personally-identifiable data) is neither useful or desirable to advertisers. Your cookie simply allows the advertising network to know which interest groups you fall into.
Anyway, you can do some pretty smart stuff with cookies and pixels, and one of them would be putting a pixel on the check-out page of a website that tells the cookie “they’ve bought X product” and that signals to the advertising network to stop serving ads for that product altogether to that person.
So let’s get back to the question; if this is possible, why doesn’t it happen all the time? Well, the first thing is basic fallibility. This stuff is quite technical to implement and quite simply there are still advertisers who believe that this stuff slows down their website (it doesn’t, at least not in any perceptible way). There are brands and websites who simply do not have the time or resource to implement this stuff correctly, so they don’t.
There’s also the matter that many brands use multiple advertising networks or partners to advertise their products, and it isn’t always possible to get these to integrate with each other. Facebook, for example, is a very popular advertising ecosystem, but is a bit of a “walled garden”, letting some data in but largely not letting data out, meaning managing your exposure to advertising across these networks in hard on a technical level. When I was a media planner, a big part of my job would be using good planning skills to mitigate this, in lieu of technical solutions for doing so. Nowadays some advertisers do use a centralised store of data (called a DMP) to manage this, but it’s still far from perfect.
Another problem is simply making all of the various advertising and commerce systems “speak to” each other, which they generally do not. Suppose you see 10x ads for a video game, then you go to the shop in real life and buy it. By what mechanism do you let the advertising networks know the product has been bought? There are actually ways to do this via CRM data (e.g. you scan a store card to get points / discount and they collect a bit of data in return). You can then signal to your networks (via the aforementioned DMP) that the person in question does get any more ads for that product. But it’s still uncommon and fallible.
Finally, there isn’t even much impetus to manage this stuff because there is little to no real evidence that over-exposure to an advertising campaign actually damages the brands or makes people not want to buy the product. Though this is a common sentiment in conversations, it has been extensively researched. The biggest risk with over-exposure is simply wasting money for excess ad impressions that don’t make the person more likely to buy the product.
**tl;dr targeted advertising is smart enough to stop showing you ads for a product you’ve purchased, but there are a number of technical and theoretical reasons why this rarely happens.**
Latest Answers