Off the Media

Chasing Ghosts: ‘Attention Minutes’ Are Voodoo, CPM is Still What Counts

offthemedia 11 Chasing Ghosts: Attention Minutes Are Voodoo, CPM is Still What CountsThough we often criticize sites like Huffington Post and Buzzfeed for their do-anything-for-clicks mentality, deep down we know it’s not really their fault. It’s the advertisers who also make this trade lucrative. The way publishers see it, they are just fulfilling a need.

I’ve been there myself. I know that pageviews are a terrible metric for measuring “quality”—I’ve compared it to the military measuring success by bodycount. But what else is there?

The CPM model, where advertisers pay for impressions by the thousand, has long been the dominant business model of online content. Publishers and readers have criticized this system—it’s easy to game, it’s simplistic, it might not even work. For all the rhetoric, deep down most of us admit that the reason this exists is because there really isn’t a scalable alternative and besides, the money is so good.

But now that publishers have abused this model to such a degree, it seems that the consensus on this is changing. They’re looking for a new way.

The Financial Times recently announced that they were testing a new way to sell ads to their advertisers, one based on how much time a user spends on a piece of their content. Jon Slade, the commercial director of digital advertising at the Financial Times, said in a recent interview that the hypothesis is the longer you show the reader an ad, the more it will resonate, and presumably, the more a publisher can charge the advertiser. Upworthy announced something similar in February, saying they had developed their own proprietary metric for selling ads, with what they are calling “attention minutes.”

The founder of Upworthy, Eli Pariser, explained how attention minutes work: “This method doesn’t store raw information—it follows a stream of events and every few seconds asks, ‘Is this person still paying attention?’ If they are, that sends a signal to a data warehouse that says, ‘Chalk up a few more attention seconds for this post.’”

They aren’t the only ones. YouTube now measures “Time Watched” and Medium is moving to what they call Total Time Reading. Chartbeat, which worked with the Financial Times to develop their new ad model, is now selling a data product for content producers called “engaged time.”

It’s a noble goal right? Instead of trying to trick people into clicks, now publishers are thinking about how to create great content that makes people stick. Or are they?

Upworthy’s recent traffic decline, which started earlier this year, could be seen as the impetus for this sea change. Although Mr. Pariser has denied it, it appeared Upworthy’s traffic took a dive when Facebook changed its news feed algorithm earlier this year. In an interview with Recode in May, Mr. Pariser oddly blamed the loss of traffic on a round of new hires,  saying, “We’ve spent the last two months aggressively growing our editorial staff—we’ll be announcing more about that soon, but it’s meant that internal resources were devoted to hiring and training, rather than curating. That’s a trade we’re happy to make as an investment in our long-term future.”

Could this be a case of Upworthy finding a scapegoat? How better to make money in the attention economy than coming up with a vague mish mash of signals and calling it “attention minutes”?

Because let’s be honest, this can still be gamed too. As I wrote in my book Trust Me, I’m Lying, what gets measured gets managed. And for the past 10+ years, blogs have used the most straightforward and cost-effective metrics—clicks and page views—squeezing all they couldfrom them until they lost their effectiveness. Why would this new metric be any different?

Jonah Peretti of Buzzfeed recently explained how these time-based metrics are just as easy to manipulate, with reality TV being a salient example. As he put it in his recent interview with Felix Salmon:

The reason that reality TV works well for time is that the classic reality TV formula, in the beginning, was the tribal council and somebody getting eliminated. So you could have 50 percent of the show being boring filler and you’re kind of wanting to change the channel but you’re like, ‘Oh, but I wonder if my favorite person’s going to get eliminated.’ So you have to watch to the end to see the elimination. In a way, that was a way of gaming time.

This comparison brings to mind Neil Postman’s great book of media criticism, Amusing Ourselves To Death, in that publishers are aligning themselves with the metrics that television is measured by—keeping you on your couch.

There is a history of web publishers moving the goal posts to charge advertisers extra, regardless of the consequences for their audience. Last year, the rage was for “sponsored posts,” which despite having very little ROI or data behind them, are incredibly expensive. That’s what I think this move is. An attempt to create a new form of advertising that costs a lot but is difficult to evaluate.

And I’ll tell you that as an ad buyer (I buy about $5M in online ads every year), I like CPM for one simple reason: it’s simple to convert into sales. I want to know that customers clicked and then bought my stuff, end of story. This new system doesn’t make my life any easier, and as the real customer for these websites, I have real trouble believing it’s going to take hold.

More importantly, this new way of measuring user engagement doesn’t solve the problems inherent in pageview journalism that these new metrics are supposedly addressing. There is still the one-off problem, where readers aren’t going to any one site because they trust its content, they’re going to whatever is being shared or going viral. There is still the problem of negative externalities being passed on to the user because the content is free, what is being exploited is their time. Stories are still going to be traded up the chain and publishers are still going to rely on sketchy sources to get the stories they want.

Instead of creating new ghosts to chase, publishers should look to make real, tangible improvements to their flawed ad models. Like I argued in my recent piece about the New York Times’ Innovation Report, the solutions to an entrenched industry’s business problems are often hidden in plain sight, a problem of chasing what’s new and sexy, instead of questioning old, outdated assumptions.

For this, publishers should look to Yahoo! (of all companies), who recently instituted their new Yahoo Prime View ads. They’re trying to solve a problem I’ve chronicled before that has tarnished the entire online ad industry—the fact that only a little more than half of display ads ever appear in front of a human being. This is the ultimate deception and deceit behind the CPM model. It’s not even real.

But Yahoo Prime View ads are based on vCPM, which are visible impressions, meaning Yahoo’s goal is that 100% of their ads will be seen by real humans, which would be a dramatic improvement to the online content game. This is what publishers should be emulating, not creating whole new paradigms by which to measure our attention. One brings us closer to reality, the other deludes us into thinking we’ve found a solution.

Ryan Holiday is the editor at large of Betabeat and the author of the recent book The Obstacle is the Way.

Follow Ryan Holiday via RSS.