There seems to be a general argument, perhaps believed by folks at Google and Facebook in particular, that they need to suck up all this data about us to provide more and more targeted advertising. I’m still not at all convinced that’s true. Earlier this year, I suggested that Google and Facebook might be better off if they just admitted that targeted advertising didn’t work as well as people like to pretend it works. The fact is that it doesn’t work all that well, and comes with massive costs in terms of everyone thinking that all these companies want to do is suck up more and more data. And the “advantage” over other forms of advertising (contextual, brand, etc.) are really not that great. Earlier this month we highlighted a study that showed that, for publishers, targeted advertising didn’t show any real benefit, and that it was mainly being used to prop up the fees middlemen got, in being able to claim some magic sauce to better target ads.
Now, the NY Times has published an op-ed by DuckDuckGo CEO, Gabriel Weinberg making the exact same point: the internet doesn’t need creepy advertising to have a workable business model. Indeed, what made Google a success in the first place was the fact that its non-creepy, non-privacy instrusive contextual advertising was so freaking profitable because it worked amazingly well:
There is no reason to fear that sites cannot still make money with advertising. That’s because there are already two kinds of highly profitable online ads: contextual ads, based on the content being shown on screen, and behavioral ads
, based on personal data collected about the person viewing the ad. Behavioral ads work by tracking your online behavior and compiling a profile about you using your internet activities (and even your offline activities in some cases) to send you targeted ads.
Contextual advertising doesn’t need to know anything about you: Search for “car” and you get a car ad. Over the past decade, contextual ads have been displaced by behavioral ads, aided by the rise of real-time bidding technology that auctions off each ad on a site based on user profiling. These behavioral ads are the ones that leave a bad taste in your mouth. They follow you around from website to mobile app based on your private information and, intentionally or not, enable online discrimination, manipulation and the creation of filter bubbles.
I’d argue that Weinberg leaves out general brand advertising as well, which can work well. Part of the problem, though, is that behavioral and programmatic advertising gives the illusion of being “scientific” because you can show data (even if that data is meaningless or misleading). As soon as you can insist that you’ll be able to show data, then people get wowed by it, and think that they’ve magically solved the “I know that half of my ad spending is wasted, I just don’t know which half” problem. But the real problem is that even with all this behavioral targeting, most advertisers are still wasting way more than half their ad spend. It’s just that they can show pretty charts and spreadsheets to pretend they have data to back up that they’re doing the right thing.
This is a point we’ve raised before. We’ve been talking to companies for a few years now, trying to convince them to advertise on Techdirt in a non-creepy way with no tracking. And what has happened, multiple times, is that a marketing person gets excited and talks about how “this is great” and how they know that they can get a lot of people interested in what they’re offering if they were to support Techdirt just knowing that our audience would appreciate them being cool enough not to track them. And then it gets handed off to an ad team or a digital agency or an ad firm that they outsource this stuff to, and eventually someone has a spreadsheet. And doing a branding campaign without creepy tracking doesn’t fit into a spreadsheet. So they pass. And waste a bunch of money on someone who will give them data, no matter how meaningless.
Weinberg is right that there’s no reason it needs to be this way. It’s just that some people have become so enamored with “data” that they don’t bother to understand what actually works.
On top of that, he points out that focusing on creepy ads has lots of other costs as well — including pissing off your users who are much more open to alternatives.
What about compliance costs? Companies are quickly realizing that good privacy practices are a boon for business. People increasingly want to reduce their digital footprint and so choose companies that help them do so. Companies with good privacy practices in their DNA do not face significant compliance costs.
Much of Weinberg’s op-ed is in support of stronger privacy laws, and I’m less in agreement with him there. I’m not against all privacy rules — but I do worry about the unintended consequences of many of the approaches proposed. Not on the revenue prospects for Google or Facebook. As noted above, they don’t need to be so creepy. But, on how it might impact other aspects of the internet. We’ve already seen this with the GDPR and how it’s being used to stifle speech and actually entrench the power of Google and Facebook. I honestly just wish that companies — both the internet ones and the advertisers who buy the ads — would just start to realize that sucking up all this data may have sounded like a good idea, but has actually been a complete failure, and start to move away from that model. It’s not necessary.