An image of a green fedora hat, which serves as the logo for this site.Gordy's Discourse

Is the online advertising bubble about to burst? External link icon.

Wired reviews a book called Subprime Attention Crisis by Tim Hwang, a former Google employee. The book postulates that the whole online advertising infrastructure is a bubble waiting to burst in the same way the subprime mortgage bubble burst in the mid-late 2000s. This is at least partly because online advertising is largely based on nothing, like subprime mortgages were.

Wired's review says the book:

… lays out the case that the new ad business is built on a fiction. Microtargeting is far less accurate, and far less persuasive, than it’s made out to be, he says, and yet it remains the foundation of the modern internet: the source of wealth for some of the world’s biggest, most important companies.

Back in the old days, if a company wanted to advertise online they'd contact the publisher directly and rent space on their website to show an advert.

These days, though, many advertisers don't advertise on a site directly, they submit a target profile for their adverts and the money they're willing to spend to an advertising network — Google, Facebook or whatever — and then a bunch of algorithms decides who gets what advertising slot.

Advertisers rely on the network's algorithms to do their job properly and return the targeted audience the advertisers require. But do they? Hwang thinks most online advertising is worthless:

There are piles of research papers in support of this idea, showing that companies’ returns on investment in digital marketing are generally anemic and often negative. One recent study found that ad tech middlemen take as much as a 50 percent cut of all online ad spending. Brands pay that premium for the promise of automated microtargeting, but a study by Nico Neumann, Catherine E. Tucker, and Timothy Whitfield found that the accuracy of that targeting is often extremely poor. In one experiment, they used six different advertising platforms in an effort to reach Australian men between the ages of 25 and 44. Their targeting performed slightly worse than random guessing. Such research indicates that, despite the extent of surveillance tech, a lot of the data that fuels ad targeting is garbage.

Hwang goes on to explain how ad blocking software and click fraud further deplete the reliability of the online advertising infrastructure.

There will come a time when advertisers realise that much of the money they're throwing at the advertising networks is wasted, and thus the bubble will burst.

I have wondered about the longevity of the current system for a while now.

The whole idea of pay-per-click and pay-per-view is flawed, which is to say it's excellent for enriching the advertising networks but dreadful for the advertisers themselves. Clicks and views don't, of themselves, earn any money for advertisers; they want sales. This means the advertising network has different measures of success to the advertiser, and that pulls the two parties in different directions.

The networks have long told us to value click through rates and strive to raise them, but that measure is just for the networks. I'm well aware there's a connection between clicks and sales — a user has to click through in order to buy the product — but it would be far more useful if the networks' objective was to produce sales rather than clicks. What the networks earn should be a percentage of what the advertiser earns, but it's fairly obvious why the networks would want to avoid that sort of system.

Advertisers and their networks are also operating in an ecosystem that doesn't consider users much.

Advertisers pay money for targeted clicks based on dodgy targeting data. Users on the other hand want rid of the lot of it. Users are unwilling participants and there's no way you can get the best out of that sort of system. Users will always resist advertising to some degree of course, but the less friction advertisers create with users, the better it will be for them.

The online world is so laden with adverts these days that it changes user behaviour. Users install ad blockers to get rid of the ones they can, and they develop visual blockers in their brains to get rid of the ones they can't. Our brains simply filter them out.

Because we're overloaded with adverts, we take less notice of them. I sometimes do surveys about advertising and I've never seen the adverts they ask me about. It's only after the survey that I might look for the advert and note it has been there all along.

Surely it would be better for advertisers and networks to consider how they can get adverts in front of users in such as way that they won't be irritated, won't want to block them and might actually want to click on them.

The current advertising systems work on volume (because that enriches the advertising networks), whereas they'd have more success working on quality instead. Fewer adverts per site (I'd argue for just one per site) will enrage users less and thus make them less likely to want to block things. Linking the advertising networks' income to sales rather than clicks or views would work better for advertisers. And if algorithmic targeting doesn't work, go back to looking at a site and seeing what sort of visitor you think it'll attract. It's not that hard.

Better still, cut the advertising networks out altogether. Go back to a system where advertisers contact site owners directly and make an offer to buy some advertising space.

The upshot, though, is that I believe Tim Hwang is right. Online advertising in its current form will eventually explode.