We need to transition to clean advertisement

We need to move from dirty advertising to clean advertising. The best way to do so is to ditch personalized ads and move towards content-based ads.

Just like we need to transition the world off of dirty fossil energy and towards clean renewable energy, we also need to rid the world of “dirty” tracking advertising and move towards clean, privacy-respecting ads.

The downsides and unintended consequences of tracking ads (aka surveillance capitalism aka ad tech, pick whichever term you prefer) are well known and manifold. They range from surveillance to slow-loading pages to increased energy consumption to the weakening of democracy.

So that’s bad. What’s more is, they don’t appear to work well even in the very limited way they are intended.

In an internal memo that was published as part of a court filing, a senior project manager at Facebook stated that…

“interest precision in the US in only 41%—that means that more than half the time we’re showing ads to someone other than the advertisers’ intended audience’. And it is even worse internationally.” 

Let that sink in – Facebook’s ability to successfully target an ad as claimed is lower than a flip of a coin. Now that memo was written in 2016 and things have possibly changed, but I somewhat doubt that it got much better.

While it’s basically impossible to get to reliable “official” numbers, a more recent study (Neumann, Tucker, Whitfield 2019) indicates that the targeting solutions offered by data brokers isn’t reliably economically viable, meaning that the accuracy gains only are worth it (economically) on higher-priced ad placements, but not in the lower-priced segment. Other studies come to similar conclusions.

To me, it appears that all the tracking and the surveillance it entails fails to get us to reliably “better” ads or sales than traditional newspaper or TV ads, which also are equally notoriously hard to measure. I’m convinced we’ve come to live with a huge guaranteed downside for a potential upside that never materialized — except for the platforms and middlemen selling inventory. It doesn’t work that well for advertisers, nor for consumers.

I argue it doesn’t have to be that way. 

Let’s look at two established existing models, then consider a third: 

  1. Medium-based ads, aka traditional mass media ads.
  2. Personalized tracking ads.

Before today’s hyper-tracked and targeted digital ads, mass media sold ad inventory to advertisers who approximated target demographics based on the medium (or newspaper section, or radio/TV segment). This is, largely, what we think of as newspaper, radio or TV ads. 

This is a supremely blunt instrument, but as far as I’m aware we don’t actually know if it was less effective than targeted ads, since measuring the real impact of ads has always been notoriously tricky. 

In theory, targeted ads are more precise. And to be honest, despite the arguments and studies presented above, I think it is quite likely that targeted ads are somewhat more effective. However, ads following an individual around their digital lives just seems as bad an idea in 2021 as it did 10 years ago. The side effects are too grave to justify the means.

So we know that non-targeted mass media ads are a blunt and expensive instrument. We also know that targeted ads have massive negative externalities that are worth avoiding at just about all costs.

Types of ad targeting

But it’s not hard to imagine a third way: Ads that don’t follow the individual (targeted ads), and rather ad inventory that is tied to the content it lives next to.

Such content-based ads could approximate interest (rather than demographics), and do so entirely anonymously, without any tracking whatsoever. Everyone who’d look at a piece of content would see the same ads. This is fairly similar to the traditional mass media model, but with one major difference: Since content on the web is far more granular than mass media could ever be, it should allow for fairly good matching of ad inventory to reader interests. Only in this model, it would be untracked, anonymous, clean.

Now, if we don’t want the many, many billions of dollars spent on digital advertising to go towards toxic ads, we need to redirect them. They won’t simply be not spent — they need another place to go. Which means there’s an opportunity to build an inventory of clean advertising space worth billions, and make the web a better place doing so. If this model succeeds, we’d have restored an essential quality of the web: free, anonymous movement across websites and content.

We’d be back at a point where we read the internet, rather than the internet reading us.


Thanks to Johannes Klingebiel for pointing me to one of the studies above.

Leave a Reply