The reigning absurdity in the digital news economy

Read on

Unlike its physical world ancestor, online publishing fails to reward investments in production and content quality. We urgently need to reconsider the error.

Today, the economic value of a journalism item stems from its popularity, i.e. the number of clicks (or views) it generates. A well-crafted listicle put together by a clever Millennial will generate more revenue that any public-interest piece, this in total disregard for who actually reads it, for how long, etc. That’s the absurdity in today’s system.

Every day, a large news organization produces hundreds of content items in all manners of form (text, image, graphics, videos) and levels of importance. Most fall is what I’ll call the Commodity News categorychunks ofinformation available everywhere at the same time. As a result, these lose their value in an instant. This makes the vast majority of what comes out of the news firehose.

By contrast, a small fraction of the output can be labeled as Value Added News; these items are supposed to be unique, more deeply reported than on average —they also cost much more to produce.

In the physical world, an item’s production cost can be passed on the consumer.

An iPhone 7 is reported to cost about $225 to produce (Bill of Materials plus manufacturing costs) and is sold for $650 —a cool 65% gross margin (before allocating marketing, distribution and R&D expenses). At the lower end of the spectrum, an ultra-cheap smartphone such as the Android One costs about $42 to produce and is sold for $50, a 19% markup. As often in the manufacturing sector, the high-end product commands the highest margin —by far. In the case of the iPhone vs. the Android One, not only Apple is grossing $425 per phone vs. $8 for the Android One, but Apple also wins on quantity: as we speak Apple’s Chinese contract manufacturing operation coughs up close to one million phones per day (and there is a three-week wait time to get one here in California.)

In terms of value for the money, the difference is clear: the iPhone has better performance, better components, a better design, and is better built than the Android One. This helps explain the gap (5x) in production costs. As for its sky-high gross margin, Apple taps into the highest possible end of the market, one that allows it to charge almost whatever it wants.

Now, let’s turn to the digital news business. Over the last 24 hours, the New York Times published about 180 articles for a total 128,000 words. How many of them can be labelled as unique in the sense explained above? 20%? 30%?

Actually, it doesn’t matter.

The price charged to advertisers is the same, regardless of the quality of the editorial hosting the ad module. To borrow from the phone manufacturing discussion, the cost of a story varies widely: a one-month investigation by a pair of reporters, reviewed by editors, verified by fact-checkers, will cost between $30,000 and $50,000; the number can actually shoot much higher if the piece originates from a hot-spot news bureau like Syria that carries a million dollars per year in fixed costs just to warrant a relatively safe environment for its correspondents. Symmetrically, a 500 words news piece that took half day to report, write and edit by a reporter and a sub-editor, will cost less than $2,000 (based on the estimated costs of the staffers involved). You can guess what’s coming: these two pieces of journalism will be sold to advertisers for exactly the same amount. Currently, the New York Times has no mechanism to pass on to its advertisers the higher production costs often associated with higher quality… Why?

The fault lies with the wrong industry metrics:

Instead of selling traffic to advertisers, publishers should sell audiences specifics.

For Buzzfeed, it makes sense to sell traffic for the explosive watermelonvideo that collected 11 millions views (even though a fraction of it actually translated into hard dollars for BF as the bulk of the audience took place on Facebook.) But for media that built their brand producing original quality contents such as the New York Times, advertising space should be marketed based on the specific audiences collected by strong editorial product. The same should also apply for native digital players such as Quartz or Politico — and also Buzzfeed for its hardcore journalistic production.

Now, imagine a new pitch where:

  • The NYT and others are able to label a significant portion of their output as “quality”, by using standardized, indisputable metrics.
  • This premium content is not sold on the base of the traffic it generates (usually x dollars per 1,000 impressions), but on demographics. This is actually easy to achieve when the reader is registered or, even better, is a subscriber; she can therefore be identified on a nominative basis with her entire socio-economics attributes (localization, income range, job sector, education, consumer profile.)
  • The premium content can also be sold on engagement, i.e. time spent. This notion can change the revenue equation. First, Engaged Time has a direct impact on advertising performance: According to a Chartbeat study (PDF here):
    — About 40% of visitors leave having spent fewer than 15 seconds engaged on the page.
    — But once a reader is engaged s/he is most likely to return:

According to the study (emphasis mine):

That correlation between Engaged Time and propensity to return is the seed of a foundational transformation taking place in the online publishing industry, from a digital content economy based on clicks and indiscriminate traffic to one that values quality and attention.
To put it another way: ads are much more viewable on pages that people actually want to read.
…[A]s Engaged Time increases from 15 seconds to one minute, viewability [of ads] goes up by over half, from 37% to 57%. Visitors who read for more than 75 seconds see more than 60% of advertisements.

This is the perfect demonstration for the correlation between editorial quality and commercial value.

As intuitively obvious as this assertion might be, it is largely ignored by the advertising community (on both sell- and buy- sides). Why?

First of all, early in the online game, publishers embraced traffic metrics where they should have defended audience specifics & engagement. Most are now are stuck in that losing model. In retrospect, a terrible mistake. (Easy to say now, I acknowledge.)

The second reason is the lack of a credible gauge to measure editorial quality. Aside of the rarely-used time spent to value content, there are very few metrics or signals available to assess journalistic value. This is at the very core of my work at the John S. Knight Fellowship at Stanford University.

In a future Monday Note, I will discuss a possible matrix of such signals and attempt to assess their reliability, relevance and feasibility.