Originally posted on sciencefocus.
Advertising was just one business model that people considered at the start. Google originally thought maybe 15 per cent of their revenue was going to come from ads; most of their money was going to come from licensing their search engine to corporate clients. It’s funny; their search algorithm was the origin of programmatic ads. Now everything relies on ads.
In the old days, ads were sold the same way they’re sold by a newspaper: you got a call from someone who wanted to buy a banner ad on your website. Google invented AdWords in the early 2000s to give people the tools to use data about what people search for. This model produced a Willy Wonka’s Waterfall amount of cash for Google, and afterwards created the financial engine that is the core of what drives the modern internet experience.
You start from this intuitive place: if I have all this data about someone, shouldn’t I be able to get messages to them that will persuade them to behave how I want? There are two kinds of failure in this thinking: one is the failure of correlation versus causation. You often end up targeting people who would have bought the product anyway, and so the question of whether or not the ad gets you to change your behaviour is unclear.
Second, everybody assumes the data you get about people is extremely accurate. But it turns out that that’s flawed as well.
Advertising works, but its effect is so limited and small that in order to figure out whether or not it actually makes a difference, you have to conduct these enormously expensive experiments. A few years ago, Procter & Gamble, one of the biggest advertisers in the world, decided to cut $200m from their digital ad spending budget. The result was nothing changed at all – they sold the same amount – which asks what that $200m is being spent on.
Advertisers say to regulators, “don’t take my data away because if you do, the ads will become less effective and you’ll destroy our market.” But then we know nothing happens. Suddenly this begs the question, what have you been collecting all this data for this whole time? The problem is that programmatic advertising generates money quickly. It’s difficult to imagine an alternative monetisation model that has the same pattern. It’s like the scene in Indiana Jones when you have to replace the idol with another idol of the same shape and weight.
The bubble has to pop at some point. Look at the history of every other market bubble and it has exactly the same phenomenon: the underlying value of attention captured by advertising is declining. Advertising has allowed us to kick the can down the road on some difficult questions.
Are we okay with an internet that grows slower and is more expensive to access? In a world where everybody needs to subscribe to a search engine, should it be public or should the government subsidise access? What about a social network? A lot of alternatives are better from a social standpoint, but they may not grow in the way that Silicon Valley is used to growing.
An internet that doesn’t rely on advertising means we’ll see more innovative business models. Some of them will benefit content creators, others will benefit companies. Maybe some of them will benefit the public. It will change the nature of the internet, and that’ll take some getting used to.
Whether you’re a technology believer or a technology critic, you have to believe in the efficacy of the technology, and a lot of our critiques of the technology give this myth a power it doesn’t actually have.