Drewbot • "Content" Creep

“Content” Creep

Publishers have stopped referring to their products as journalism, writing, literature, photography, or art. Today, everything is simply “content”.

Those working in media, especially digital media, can attest to the word’s popularity. Quantitatively we can observe this trend in the chart above: over the last ten years, annual financial reports from The New York Times have leaned more and more heavily on the word “content” when describing their business. The share of “journalism” has remained relatively flat.

The NYT and other publishers rely on the word “content” to help them understand the breadth of their output. But by reducing their writing, photography, videos and more to a single, nondescript term they’re setting themselves up for failure.

The Rise of “Content”

“Content” emerged with the rise of the Internet, which detached pieces of work from their primary media. Before the web, we referred to works by the media format which delivered them: as newspapers, magazines, paintings, photographs, records, CDs, and so on. As digital representations grew in popularity these monikers became increasingly awkward. Is a newspaper still a “paper” when the majority of its readers view it on screens? To abate this awkwardness, we began to search for a more apt term. We landed on “content”, a bucket term which we ask to describe anything a publisher could publish, from the most revelatory art to the most hackneyed rags.

At this point, “content” was an innocent, sloppy fix. A stopgap until the Internet settled down and a proper term could be coined. Unfortunately, the pace of innovation quickened and today language is unable to keep pace with rapidly emerging new ideas, art¹, and businesses.

So we’ve stuck with “content”.

The Assumptions & Allure of Content

To achieve its representative breadth, the word “content” makes two assumptions:

Each piece of “content” is equal and is therefore interchangeable: As stated earlier, “content” is used to represent a wide breadth of works. A Pulitzer winning report and a Business Insider slideshow are both single instances of “content.” The word must remain formless, devoid of emotion, and of indefinite form and quality. Any characteristic which might differentiate two works must be ignored. This rhetoric categorization gives rise to the second assumption.
“Content” production is trivial: Since each bit of “content” is interchangeable, “content” is only as hard to create as the easiest instance.

Publishers buy into these two assumptions because “content” allows them to easily measure and analyze their output. Messy qualitative measures are hidden so output fits neatly within Excel cells. This is the allure of “content”: it allows comforting, structured data which simplifies the complexity of a large business and makes decisions less intimidating. Executives aren’t making qualitative picks regarding art or an artist, they’re merely signing off on whichever “content” produces more valuable metrics.

At The New York Times it’s conceivable that editors and executives have a handle on their output. But businesses with strategies dependent on massive levels of “content” production cannot know the quality of everything that ships. Think YouTube, where users upload more than an hour of video every hour. Or content farms like Demand Media, which claims to have created 2 million articles and 200,000 videos as of June 2010.

Demand Media’s Built in Blind Spots

Demand Media is the starkest example of a business built around “content” based metrics. If “content” is not trivial to produce, Demand’s costs become unsustainable. If each piece of “content” is not equal and interchangeable, Demand’s business metrics are fundamentally flawed.

Demand uses custom software to identify topics which are valuable to advertisers, based on advertising auction requests and web search search activity. Once a topic is identified, Demand employs cheap freelancers to churn out topical “content”, around which ad space may be sold in advertising auctions. According to their IPO filing, Demand Media “generated a daily average of over 5,700 text articles and videos during the quarter ended June 30, 2010”.

To manage this massive output, Demand Media embraces the metrics “content” allow. According to their IPO filing Demand employs two metrics to measure “content” value: page views and RPM (revenue per thousand page views). Demand believes these measures are “the primary indicators of [its] performance.” For long term analysis, Demand lumps a quarter’s worth of content into a unit called a “cohort”, which is then compared against past cohorts.

Clearly, Demand’s quarterly metrics assume instances of “content” are equal. RPM measures how well Demand is able to monetize² a given post and is dependent on page views. By the process of elimination, page views is the only metric listed in Demand’s filing which might address quality.

Unfortunately, even if we assume page views are capable of measuring quality Demand’s business model prevents them from doing so. Because Demand’s “approach is driven by consumers’ desire to search for and discover increasingly specific information across the Internet”, page views are only capable of reflecting how well Demand’s “content” has been optimized for search engines. If a piece appears in search results, is clicked by a user, and closed because the writing is shoddy, Demand is only able to measure everything before the click. At best the page views metric can measure the quality of the headline. At worst they reflect the SEO tricks employed by a site³.

Demand’s metrics cannot measure the quality of their writing or videos. This is at odds with the first third of their filing, which repeatedly champions their devotion to “quality content”. Later, they cite their “Valuable and Growing Content Library” as a competitive advantage: “Our wholly-owned content library, consisting of approximately 2 million articles and approximately 200,000 videos as of June 30, 2010, forms the foundation of our growing and recurring revenue base. We strive to create content with positive growth characteristics over a long useful life.” If Demand is unable to assess quality, an evergreen back catalogue is a pipe dream. (Especially when Google changes it’s algorithm.)

Because Demand’s models are so reliant on the word “content”, the assumptions listed above are more than a linguistic quirk. Quarterly reports and executive bonuses live and die by these metrics; yet these metrics can’t tell you if anyone wants to actually read this “content”, when it’s published or anytime in the future. Demand has created an environment which incentivizes SEO hacks more than good writing.

The “Content” Crunch

So what does all this lead to? One day, if you’re working at a business with a “content” based model, you’ll wake up and find out you’re not hitting your revenue numbers. Projected earnings from your amassed “content” trove aren’t being realized. Maybe Google changed their search algorithm or perhaps users are turning to apps to find articles. Maybe your articles just suck. Whatever the cause, you’re left trying to close the gap between the numbers you promised your boss or investors and the actual revenues your older “content” is delivering. Let’s call this “the content crunch”.

Unfortunately, the “content” assumptions which created your problem also compound it. Because you can’t directly measure or assess quality, and therefore cannot invest in it, your options are limited. In fact, you have 3 choices.

At first, you can make more “content”. Remember, you view “content” production as trivial. At first, this is what you do. You demand more posts from your writers or outsource to companies like Demand (Outlets like USA Today, The San Francisco Chronicle, and The Houston Chronicle have already done this). This is probably the lever you pull first. But this only treats the symptom, not the disease. You still can’t see quality. In another 3 months you’ll be back in the same position.

Next, you can increase the amount or value of advertising around your “content”. In a recent NYT article regarding Yahoo’s new CEO, media advisor Charlene Yi provides us with a business-speak example of this tactic that borders on parody. Yi advises Yahoo!, “It’s not about putting up pretty content, but about maximizing the value of the eyeballs in front of that content.”

If you find yourself in a “content crunch” increasing the value of your advertising is not a realistic option, as your “content” is low quality. Your only option is to increase the number of advertisements. So publishers caught in the crunch surround their underperforming “content” with an antagonizing amount of bottom-of-the-barrel advertising.

Last November, Brent Simmons aptly describes the unique webpage hell that results:

What I saw — at a professional publication, a site with the purpose of giving people something good to read — was just about the farthest thing from readable. The site has good writing. But the pages do everything possible to convince people not to try. “Don’t bother,” the pages say. “It’s hopeless. Oh — and good luck not having a seizure!”

They’re filled with ads and social-media sharing buttons — and more ads. And Google plus-onesies and Facebook likeys. And also more ads. Plus tweet-this-es. Plus ads. (And, under-the-hood, a whole cruise-ship-full of analytics. The page required well-more than 100 http calls.)

The site Simmons describes is all too familiar today. Sites are laden with “innovations” like fake hyperlinks that pop-up video advertisements and hover-over banners which take over the screen if your mouse lingers too long. Imagine trying to explain services like Instapaper, Readability, and Read It Later to people ten years ago, when publishers were’t actively hostile to their audiences.

Clearly, increasing the amount of advertisements around your “content” is not a real solution to the “content crunch”. You still can’t see or invest in quality and will continue to produce “content” of questionable value. Increasing the amount of advertisements around your decaying “content” may boost your numbers in the short term, but it helps drive away what little audience remains. Soon, your numbers will be worse and you’ll have to cram in even more ads.

All this leads you to the last option, which is a difficult pill to swallow.

If you’re still in business, you can attempt to adjust your business model to account for quality.

But how do you account for quality? The simplest way is to hire a sufficient number of talented people to either produce or edit all the “content” you output. Because this method limits your output to the number of skilled people you can afford, most digital media outlets look past this method. Media futurists hope social data or semantic technology (systems that understand meaning in data, like IBM’s Watson) will eventually be able to govern quality. But news apps built on social data like News.me and Summify are limited to small lists of output (since they’re dependent on the quantity of articles shared by friends) and tend to serve up thin, meme-friendly fare. Semantic systems are still in their infancy: experimental systems are making progress categorizing “content” by topic, but a objective metrics like quality still evade them.

A good example of a publishing company whose business model accounts for quality is Gawker Media. Gawker does cherish quantitive analytics, true, but they temper them with an editorial team of sufficient size and talent to access whether or not they’re publishing anything worth reading. In his recent ‘State of Gawker 2012’ email, publisher Nick Denton spells this strategy out clearly:

Our sites have assembled the strongest collection of journalistic talent on the web today. Sure, we know how to play the web game like Buzzfeed and Huffington Post. We measure. We hone headlines. We sell stories. Sometimes we oversell. But — and this marks us out — we believe that the best web content optimization strategy is something as old as journalism itself: the shocking truth and the authentic opinion.

Because these tactics have lead to success, Denton is shifting Gawker’s payment habits to invest in this differentiating staff. He writes, “We should not wait for a poaching expedition to pay someone what they deserve. I apologize if that has been the case and will do better in 2012.”

In the past, present, and future, Gawker will be shielded from the “content crunch” because they’re able to assess quality and invest in it. Despite the redesign hiccups of last year, Gawker’s sites are incredibly reader friendly and advertising is kept in check⁴.

For businesses like Demand Media, whose models depend on massive “content” production, hiring a talented editorial team to process millions of articles is economically unfeasible. As I see it, there’s no current way to measure quality as quickly as one can produce “content”. Until a balance is reached, “content” dependent businesses are fundamentally unsustainable.

“Content” and the metrics it enables are not unlike a drug. They offer a quick fix for publishers, but overuse leads to a brutal hangover. And while it’s easy to pick on Demand Media, their habits are hardly unique within the digital media industry. The New York Times leans heavily on “content”, as we see above, and smaller outlets are cutting their staff while embracing a “do more with less” strategy (to quote the 5th season of The Wire), upsetting the ratio of quality-measuring staff to “content” published. Add in the numerous Demand-like “content farms” (such as AOL’s Seed, Yahoo’s Associated Content, and the NYT’s own About.com) and it starts to feel like “content”-based metrics are the norm online.

It’s hard to believe a single word could slate an entire industry for failure. On its own, the word “content” is merely awkward. But as a unit of measurement, “content” affects business is real ways. Ignoring the variables audiences care about in order to populate Excel spreadsheets incentivizes weak writing short on substance and attention spans. All this would be tremendously depressing if it wasn’t creating an enormous opportunity for people with the courage to look beyond the numbers, where it’s too messy to measure, and invest in journalism, videos, photography, and art people might actually enjoy.

My favorite example: MySpace and the ease of digital production has produced so many bands that we’re running out of names. ↩
Do we have a new word for this yet? ↩
To be fair, some of Demand’s traffic does come from social sharing and from people visiting their sites directly (by typing the address in their browser). Demand’s CEO Richard Rosenblatt stated as much during an earnings call last February, as reported by the Wall Street Journal: “[Rosenblatt] said in December, more than 100,000 different eHow articles were shared on Facebook… Rosenblatt said the rate of people typing “eHow.com” directly into their Web browser is far outpacing the rate of growth from search visits.” But these figures are a façade: 100,000 shares is laughable when comScore counted 75.5 million monthly visitors the following January and Rosenblatt doesn’t reveal the actual volume of eHow’s direct traffic, just the rate of growth. Since Demand’s IPO filing states 41% of their page views in Q3 of 2010 came from search, the volume of search traffic is so large that virtually any direct growth would “outpace the rate of growth from search visits”. ↩
Gawker’s advertising strategy is unique online and a post unto itself. Here’s two smart things they do: they encourage advertisers to create advertisements unique to Gawker (and therefore more relevant) by choosing non-standard ad units and providing access to an in-house creative services team and instead of auctioning off unsold inventory they feature independent artists, preventing the presence of bottom-of-the-barrel, scammy ads. ↩