Data Lords: The Real Story of Big Data, Facebook and the Future of News

Tobias Hase/picture-alliance/dpa/AP Images

This afternoon I saw a friend on Twitter say that he doesn’t buy the idea that if people just paid Facebook some sort of fee the data and privacy issue would go away. Because he subscribes to the Times, the Post and the WSJ and they each track his readership habits and sell that data to advertisers or make it available to them for targeting. This is at least partly true – I’ll discuss the ins and outs of that point in a moment. But this is a good opportunity to discuss the real relationship between publishers and big data. It’s actually very different than it looks.

First, what my friend says is true. These publications are all in the data collection and sale business. Indeed, TPM is too – not directly at all but because of the ad networks (like Google and others) we have no choice but to work with. The key on the main claim is that the issue is one of diversity of revenue streams. Each of those big publications mentioned has at least three big revenue sources that are relevant to this conversation. They have premium advertisers for which the kind of data we’re talking about has limited importance. They also have subscriptions. The final bucket is made up of advertising that is heavily reliant on data and targeting.

The difference is that Facebook is almost 100% reliant on advertising which is not only reliant on data and targeting but reliant on the most aggressive kinds of data collection, tracking and targeting. That is Facebook’s entire business. Anything that cuts deeply into that model and advantage represents an existential threat.

But here’s the really salient point. Almost every publication participates in the data economy. But the data economy is almost universally a bad thing for publications, especially ones that have real audiences.

Allow me to explain.

Different publications have different kinds of audiences. If you want to sell consumer goods to a mass audience, you might advertise in People magazine or Yahoo News. If you want to advertise to affluent investors you might advertise on WSJ or Barron’s. If you want to advertise to highly educated progressive news junkies who are what ad industry types call ‘opinion leaders’ you might advertise on TPM.

Different publications have different audiences and they have different relationships with those audiences. Some have deep bonds of trust, others don’t. Historically (and increasingly today), publications sell subscriptions. They also sell ads which can command different rates depending the value advertisers attach to talking to different groups of people. If you have a strong brand and a key audience, a publication can command profitable advertising rates because they become a kind of gatekeeper for a given audience.

The evolution of big data, tracking, targeting and advertising over the last half dozen years has transformed this dynamic.

Let’s take a hypothetical publication, Fishing Times. In the old days, paper or digital, if you’re advertising fishing equipment, Fishing Times is a must-buy. But tracking and targeting changes that. Now Google or Facebook or a number of other players whose names you haven’t heard of can target people interested in fishing or even Fishing Times readers wherever they go on the internet. Let’s call these “Fishing Times readers” collectively.

So Fishing Times’s ad department is selling access to the prime Fishing Times readership. But the Data Lords can say, ‘we can show your ad just to Fishing Times readers when they’re on Facebook, or on some meme site, on the Times or TPM or really anywhere.’ Because the Data Lords have the data and they can track and target you. The publication’s role as the gatekeeper to an audience is totally undercut because the folks who control the data and the targeting can follow those readers anywhere and purchase the ads at the lowest price.

That’s not all. The Data Lords can also create something called ‘look alike audiences’, a key part of what Facebook (but by no means only Facebook) does. This means that the data may show that people with brown eyes, Toyota Camrys, fans of Katie Perry and chocolate milk buy the most fishing equipment and are the most intensely loyal readers of Fishing Times. Now the Data Lords can show your ads not only to Fishing Times readers (wherever they go on the web) but to this demographic profile which may seem to have no connection to fishing but yet has a demonstrated propensity to purchase fishing gear in the same way Fishing Times readers do.

There’s one final part of the equation. Let’s go back to advertising at Fishing Times. We have an advertiser who sells hand crafted fly fishing gear. This is pricey stuff and it appeals to only one, fairly small part of the Fishing Times demographic. The advertiser may set up a private auction or marketplace with Google which will allow them to show their ads on Fishing Times only when a reader fits a very specific demographic profile.  So Fishing Times is getting those dollars but only a tiny fraction of them because the Data Lords have allowed them to pick and choose the small subset of the audience they want to show their ads to.

A few caveats: This transformation is not quite as total as I suggest. Some publications have sufficient brand cachet that appearing on their site retains a value of its own. Some have enough market power that they can keep the data/targeting world off their most prized publication real estate. But the dynamics I’ve noted are affecting all publications and most of them to a great degree. If you’re hearing about the advertising and monetization crisis so many publishers are facing, what I’m describing here is a huge part of what is going on.

There’s one more indeterminate asterisk that floats over all these calculations: that is, the Facebooks’, Googles’ and other Data Lords’ data pretty clearly isn’t as good as they claim. In recent weeks, for instance, Facebook has had to repeatedly reduce estimates of the “reach” of their ads (how many people actually see them). Since the value of ads is whether and how many people see them, having the reach numbers wrong is a pretty big deal.

In pure efficiency terms, most of this is great. As an advertiser I don’t have to pay high rates to reach New York Times readers on The New York Times. I can pay Google or Facebook to find them for me elsewhere for much cheaper prices. But as you can see, in this new world of data, tracking and targeting, it’s the Data Lords who have the power and get the money. It’s comparable to the way Apple’s iTunes and Spotify dramatically reduced the amount of money in music sales and took a big, big chunk of the reduced size pie for themselves. Amazon is comparable with books and a lot else.

There are significant public questions about whether these uses of data are acceptable in privacy terms and whether society is well served by having news publications bled dry by the social platforms. But capitalism, like life, isn’t fair. This is classic Schumpterian creative destruction. For the moment, my point is only to illustrate that the data and targeting world is almost a universal negative for publishers. They may participate in it – to some degree because they don’t grasp how much it undercuts their livelihood but far, far more because they simply don’t have any choice. But the pervasive use of personal data for tracking and targeting ads is bad news for almost all publishers, but particularly those who try to develop premium brands and produce things of value which require significant investment to create.