What If Google And Facebook Admitted That All This Ad Targeting Really Doesn't Work That Well?
from the the-data-obsession dept
You may have heard the famous line from early department store magnate John Wanamaker that “half the money I spend on advertising is wasted; the trouble is I don’t know which half.” Over the past decade or so, various companies have argued that their ability to provide a ton of data, combined with whatever algorithmic magic they could throw at their platforms, could lead to a magical mythical world in which there were perfectly targeted advertisements. And, of course, in the past few years there have been literally just two places where advertisers believe they can get perfectly targeted advertisements that don’t waste half (or more) of their ad spend: Google and Facebook.
The end result of this thinking is that Google and Facebook need to engage in what people refer to as “surveillance capitalism,” collecting a ton of data on everyone, building a huge profile about every user, and snooping on basically everything everyone does all day. This is why people have been getting more and more annoyed about the privacy trade-offs over the past few years (though, not so annoyed that they’ve stopped using these platforms in any significant way — though, that could happen). It also has resulted in advertisers assuming that they must put the bulk of their ad dollars into those two platforms on the assumption that the money is better spent there. Indeed, the most recent IAB report on this noted that while the internet ad market continues to rise, 90% of the growth went to Facebook and Google (together the two companies represent about 58% of the total market share for online ads, but 90% of the growth in 2017).
Advertisers have been completely sucked into the belief that if you want to get results for your ads, you simply have to throw money at those two giants, and they’ll mix some magic pixie dust with all the data they’ve collected, and voila: perfectly targeted advertising. Everyone get so focused on magic words like “big data” and “artificial intelligence” and “machine learning” that they rarely ask the larger question: does any of it actually matter?
As more and more questions are raised about the data practices of Facebook and Google, it seems worth questioning whether or not they actually need to be collecting all this data, and how much of a loss it actually is if they don’t. Just recently, Facebook announced that — as part of a settlement with the ACLU — it was drastically changing how it handles certain ads: specifically that it would no longer allow such granular targeting for housing, employment, or credit ads — all three of which were seen in the past as leading to discriminatory outcomes.
If such targeting really was important and useful, you’d think that this would have resulted in Facebook’s stock price cratering. Instead, it went up.
The little secret behind all of this that very few people want to admit is that, in most cases super-targeted ads are crap. They don’t perform well. That’s because even if you’re putting the ad in front of the right demographic, most of the time they don’t care or don’t want to see whatever it is that you’re pushing. Or, it shows an ad for something you already have (or the ever popular laugher: something you just bought and don’t need to buy again).
Unfortunately, most advertisers don’t quite realize this yet, and Google and Facebook are in no rush to tell anyone (though, frankly, they should be more upfront about all of this). Some are realizing this through other means. It didn’t get that much attention, but back in January it was reported that, because of the GDPR, the NY Times stopped using behavioral targeting for ads… and found its revenue went up. The Times is doing much more basic targeting now: just contextual and geographical.
And, if anyone should know this, it should be Google. For much of Google’s existence, its big secret sauce was not deep knowledge about the people seeing the ads: it was just matching them against their search terms. That is, just a bit of simple contextual information, rather than tying it to a giant portfolio of data about you. It’s really just over the last decade that Google really focused hard on building data profiles on everyone and “customizing” everything. There may be some advantages to some of those customizations — and there are certain useful things that come with the data — but better targeted advertisements… don’t really seem to be among them.
Frankly, if Facebook and Google want to get regulators off their backs, they might start by coming to terms with this basic fact themselves and choosing to stop collecting so much data on everyone. Recognizing that they can still build incredibly powerful ad-driven businesses without so much data would be a big step forward. Right now, unfortunately, it seems that everyone remains bought into the myth that they need this data, that their business models are dependent on this data, and that this data is actually useful in the advertising context. Bursting that myth might mean that advertisers aren’t quite as enamored with Google and Facebook over the long haul (though, they’d still spend a ton of money with them), but it might lead to a better overall experience for users, and a hell of a lot less regulatory pressure.