Chapter Two: Traffic Patterns

Why big audiences aren’t always profitable

gawker.png At first glance, the numbers don’t seem to add up: The New York Times has more than 30 million online readers and weekday circulation of less than 900,000 newspapers. Yet the print edition still accounts for more than 80 percent of the Times’s revenue. A broader recent study revealed the same phenomenon: It showed that the Internet occupied 28 percent of Americans’ time spent in media in 2009 but generated only 13 percent of total advertising spending.

To understand why, it’s important to realize that the prices advertisers pay digital news organizations depend on many factors. Some are tied to the overall market, especially the vast and growing amount of ad space (or inventory) that’s available online. Other factors have to do with a site’s own dynamics, including the size of the audience it reaches and that nature of that audience—its demographics, how much time its users spend with the site, and so on.

So the web offers a lot of advantages to publishers and advertisers. But its audiences are more wide than deep.

Journalists constantly feel the push and pull of these numbers. “What am I today?” asks Jeff Cohen, editor of the Houston Chronicle. “I’m an aggregator of eyeballs.… We’re doing around 79 million page views a month—almost a billion in a year.” Yet for all that web traffic, his newsroom’s 206 employees are about half the number employed in 2006.

Digital numbers are confusing when compared with traditional media metrics and are often inflated for all sorts of reasons. Users can be counted several times if they deploy multiple devices, such as a PC, laptop, and mobile phone, to access a site. Also, many people delete their computers’ “cookies”—small text files that allow them to be identified and tracked; because of that, they appear to be new visitors to sites rather than returning ones.

For all that, digital audiences usually far outnumber those from a traditional outlet. As a result, the industry faces a perplexing set of questions: Why do so many digital users generate so little advertising revenue? Is it simply that digital systems are more efficient than the previous oligopolies of the print and broadcast world? Or is there something more fundamentally askew about the way media companies make money off digital customers? And, more importantly, what should publishers be doing to make the most of the readers and viewers they have?

At its most basic level, advertising is a numbers game. A news organization needs a certain number of readers or viewers, and the more it gets, the more ads it can sell and the more it can charge those advertisers. Users also spend varying amounts of time with the magazine, newspaper, or broadcast, and the more time they spend, the more an advertiser values the audience.

Digital platforms, thanks to their ubiquity and ease of use, are terrific at the first part of the numbers game. News sites have demonstrated the ability to attract huge numbers of users. And in the first decade of the digital era, particularly as search engines became more powerful, publishers and broadcasters focused on building a mass audience. They poured resources into search engine optimization—the term used to describe a way to improve the odds that headlines will be picked up by Google or other search sites and that topics will be timely enough to appear prominently on a results page. News sites initially welcomed aggregators, such as Drudge Report and Huffington Post, that linked to their material and increased traffic.

As a consequence, audience sizes swelled, and publishers have proclaimed that to be a success. So when the Los Angeles Times in March 2011 logged a record 195 million page views, clicked on by 33 million users, the site’s managing editor took a moment to proudly announce those statistics on the site.

But within such numbers is another, less happy, story. The arithmetic shows that each user clicked on an average of six pages in March—or just one page every fifth day of the month. That statistic demonstrates how the other essential part of the advertising business—the amount of time and attention that users pay to a site—has been undermined by some of the tactics that publishers have used to attract large audiences.

The phenomenon is widespread. A 2010 Pew analysis of Nielsen media statistics depicts depressingly low levels of usage, even at outstanding national sites. “The average visitor spends only 3 minutes, 4 seconds per session on the typical news sites,” the study says. “No one keeps visitors very long.” And at top-trafficked news sites, ranging from Yahoo News to The Washington Post to Fox News, most people visit just a few times per month. Compare that to the media of past decades. A 2005 study showed that about half of U.S. newspaper readers spent more than thirty minutes reading their daily paper. And most of the less-devoted readers spent at least fifteen minutes with the paper.

There are many reasons for the problems news sites have in getting reader attention, also known as engagement. Consumers today have many digital options available. The experience of getting news on a computer or mobile device, thus far, is fundamentally different from the experience in TV or print; most users tend to flit from site to site, rarely alighting for more than a brief spell.

But there’s another way to look at these numbers, one that is more complex—and, in some ways, more encouraging—for the journalism business.

One person studying this issue closely is Matt Shanahan, an analyst who is relatively new to the media world. Shanahan is, by training, an electrical engineer, and he spent much of his career consulting for big businesses in the fields of aerospace and finance. In 2008, he joined a Seattle-area firm called Scout Analytics; the company was looking to serve industries that were failing to realize their potential in the digital world. “The e-commerce market was noisy and crowded,” he says. “Media and information, however, were mature industries facing severe revenue issues.”

Scout figured it could find new clients among media companies that were trying to learn more about the real nature of their digital audiences. The company signed up about seventy news and information sites—some geared to consumers, but more of them in the business-to-business category—and started looking at the activity of users. They were able to track consumers’ paths through sites and to figure out how often they visited and what they did on each visit. Shanahan was soon struck by an anomaly of the online news industry: the yawning maw that separates the size of audiences from the level of engagement those readers and viewers demonstrate.

Shanahan points to a website for a 90,000-circulation newspaper that serves a medium-sized city on the East Coast. (The name of the company is confidential because it’s a client.) This site gets around 450,000 unique visitors a month.

But those visitors differ widely, and Shanahan separates them into four types: The most loyal are the “fans,” who visit at least twice a week. Then there are the “regulars,” good for one or two visits a week. Sliding down the loyalty scale are “occasionals,” who stop by two or three times a month; and finally, the “fly-bys,” who come just once a month.

The most loyal visitors are a very small part of the overall audience: Fans make up just about 4 percent of the total number of visitors, and regulars 3 percent. Occasionals account for 17 percent and fly-bys for more than 75 percent of the total. In other words, more than three-fourths of the people who visit this news site do so just once a month.

Then Shanahan went deeper, to see how the different kinds of users behaved on the site. He knew the most loyal fans would generate more page views than the fly-bys, since fans visit the site more often. But the disparities in usage were far greater than one might expect.

Fans, despite their small numbers, were responsible for more than 55 percent of the site’s traffic. Fly-bys—those people most likely to come from a search engine or a blog—clicked on barely three pages a month. Overall, each fan generated about fifty times more traffic per person than a fly-by.

“When people talk about the size of an audience, that’s a sham,” Shanahan says. In his view, stated numbers don’t reflect how differently the varieties of users act in the way they navigate a site. Publishers mistakenly focus on “page views rather than length of time,” he writes on his blog, Digital Equilibrium. Referring to ad “impressions,” which are views (not clicks) of ads, Shanahan adds, “Using today’s standard, there is no difference between impressions that last one second, ten seconds, or two minutes.”

“The digital world has changed the revenue dynamics for publishers,” he adds in another post. “In the print world, a publisher’s shipment of physical media was the basis for generating revenue. In the digital world, consumption of media is the basis for revenue…. In other words, engagement is the unit of monetization.”

Shanahan says the benefit of more engagement isn’t just in higher ad rates, but in relationships that publishers need to build with their most loyal readers—something that has been lost in the drive to attract mass audiences. By chasing after large audiences rather than deeply engaged ones, he says, news organizations are sacrificing advertising revenue. Publishers who have a “direct relationship with fans can push better contextual advertising”—that is, ads that relate directly to a user’s habits and interests. “A publisher can know which fans saw which advertisements in which context on their site.” Sites can use that information to provide readers with targeted ads or offers.

And news organizations that hope to charge for online access need to locate—and cater to—fans, since engaged users are more likely to subscribe. That way of thinking is at the heart of The New York Times’s decision to charge for access to its digital editions. The Times built its pay scheme so light users of the site won’t have to pay for access, while heavier users—defined as those clicking on at least twenty stories a month—will be charged. But the Times also sees a connection between engagement and advertising. Michael Golden, vice chairman and president of The New York Times Co., discussed the different kinds of online audiences during a speech at a March 2011 advertising conference, shortly before the pay strategy went into effect. Speaking of less committed users, Golden said, “Their engagement is limited, but their numbers are impressive.… We have to balance engagement and reach. The higher the engagement, the higher the CPM.”

Two media outlets with very different editorial missions—Gawker Media and PBS—have tackled this issue of trying to differentiate more and less engaged segments of their audience when pricing advertising.

Gawker’s network of sites, started by British journalist Nick Denton in 2003, includes Gizmodo (for gadget lovers), Deadspin (for sports fans), and Jalopnik (for car buffs). In March 2010, Gawker began touting a metric it calls “branded traffic.” This was defined as people who have bookmarked the company’s sites or arrived at them by searching specifically for the site by name—by, say, typing “Deadspin” into a search engine.

Gawker found that roughly 40 percent of visits come via branded routes, as contrasted to links from search engines. And such visitors are more devoted and engaged, spending ninety-one seconds more per visit than others. That is a meaningful difference with financial impact, says Erin Pettigrew, Gawker Media’s marketing director.

First, the more engaged users are more likely to see highly profitable “road-block” advertisements—ads that take over the home page of the site for several seconds and then fade away to reveal the editorial content of the page. That revenue helps compensate when advertising rates fall at Gawker and other sites. As Reuters’s Felix Salmon has noted, the number of online ads has been growing so fast that advertisers can demand lower rates. As a result, wrote Salmon, “Denton says that since 2008 he has been getting only half the revenue per page that he used to get in 2004.”

One way to counteract that is to try to sell more “branding” advertising on the web—that is, ads designed to draw positive attention to a company’s name and image, rather than to trigger a direct response to an offer for a product or service. Brand advertising has long been profitable for traditional media, especially television and magazines. But because publishers have struggled to convince advertisers that the web is a good platform for branded ads, they’ve often missed out on this lucrative revenue stream. “We had been told ad frequency of more than two or three exposures is a bad thing,” said Pettigrew. “But for brand advertising, it’s different. Six to twelve exposures [of an ad] increased the persuasiveness. And this is the most useful segment to our advertisers.”

Because Gawker Media has an array of sites on different topics, it can sell engagement across its network. Thus, the same readers who click on gadget stories at Gizmodo can be served with branded ads at Lifehacker or Gawker. And if Gawker Media can demonstrate to advertisers that its readers are loyal, it can charge higher ad rates, Pettigrew said.

At the other end of cultural spectrum, Public Broadcasting System’s web strategists are also using engagement metrics to increase revenue.

Amy Sample, director of web analytics for PBS Interactive, says she and others at the site modified a formula created by web analysts Eric Peterson and Joseph Carrabis to get a better sense of which readers were most devoted. They came up with their own criteria to determine’s most loyal audience, based on the number of pages a reader views, the amount of time a reader spends on the site, and how often and how recently readers have come.

As it turned out, less than 5 percent of the visits on the site came from users who met all of PBS’s engagement standards. But those people are a critical group, says Sample. She found that they stay on for 13.5 minutes per visit (compared with a three-minute average for everyone else) and click on nine pages per visit (versus three for other users). PBS saw economic benefits from this audience. Such users were 38 percent more likely to donate money to PBS than less engaged users; they were also more prone to encourage others to use the site. And when PBS saw the usage patterns, executives decided that video, a favorite platform for frequent users, should be promoted more prominently. That translated into revenue, because the site’s video ads get healthy $30 CPMs, Sample said, or about three times as high as other ads on the site.

Engagement correlates with editorial content. To see how the relationship plays out at a large site, we can examine some numbers for, the main site for The Dallas Morning News. (These metrics are from the full year of 2010, a few months before the publisher began charging for access to much of the site.)

For the year, the site averaged around 40 million page views a month, driven by 5 million visitors who visit, on average, about twice a month and click on about four pages per visit. Those numbers are fairly typical for a site the size of Dallas’s and provided the publisher, James Moroney, some of the figures he used to calculate the rationale for instituting the pay-for-access plan (see Chapter Five).

But the broad numbers tell only part of the story. In fact,, like many big online organizations, is many sites rolled into one. To analyze its data, the company sorts its traffic statistics into various categories, including by content areas: news, entertainment, sports, weather, and blogs.

News gets the most traffic, in terms of total visitors and visits. News visitors average around two visits a month and click on an average of about 1.5 pages per visit. Their habits are typical of those found at many other news sites—not particularly engaged.

Sports does better in engagement. Users average about 2.3 visits a month, and about 3.4 pages per visit over the course of the year. During the fall of 2010, when the Texas Rangers were in the Major League Baseball playoffs and the Dallas Cowboys were on the football field, users clicked on four or more pages per visit.

And then there’s a feature on the site called High School GameTime, which includes rosters, schedules and results from the state where Friday Night Lights is based. Users clicked on nearly nine pages per visit in November 2010, during the height of the football season, and generated almost as many page views as the entire news section. Put another way, high school sports fans were seven times as engaged as the people coming to read news.

It’s easy to see why. The site offers a dizzying array of statistics, rosters, and standings for more than 200 high schools in the Dallas-Fort Worth area. Mark Francescutti, senior managing online editor for sports, says the site’s engagement demonstrates the power of “great local content… that is exclusive and is important to people.” And loyalty, not search engine optimization, is the key to maintaining the audience. “We might get lucky and get linked off Google, but we want people who will come back every single day.”

The site has a small but intense crew. The News’s four full-time high school sports reporters file frequently, and editors also rely on clerks who take scores and statistics over the phone from stringers around Dallas. On Friday nights, scores are updated during games, not just reported when the games are over. There’s also a live chat where reporters update games—“controlled chaos,” in the words of Kyle Whitfield, the site’s editor. High School GameTime aggregates heavily from other sources. “Our writers are not robots,” says Francescutti. “We don’t have that old journalistic ego that says, ‘If we didn’t write it, it’s not important.’”

The News used High School GameTime as part of a package deal with Time Warner Cable that also included print ads, a radio show, a player-of-the-week contest, and a banquet at the end of the football season. That brought in up to $700,000, says Richard Alfano, a general manager. For the next season, he says, the News will sell a $1.99 mobile app for High School GameTime that will include play-by-play from at least 100 games a week.

The key, says Whitfield, is focusing on something that readers care about deeply and that no other news provider could do as well. “It’s more difficult to sell Cowboys coverage, because Cowboys fans are everywhere around the country,” Whitfield said. “We were able to organize our resources and monetize it, which is oh-so-rare online.”

Rapid audience growth is often accompanied by thin engagement. Such has been the case at, a freelance-driven site that has built an audience of more than 22 million unique users in three years.

Examiner is owned by Clarity Digital Group, which is controlled by Philip Anschutz, a Denver entrepreneur who has made billions in energy, railroads, entertainment, and sports franchises. The site, which was started in April 2008, has brought aboard more than 72,000 freelancers who have written on topics ranging from roses in Rhode Island to parenting in Portland. There’s not a great deal of supervision: Writers must pass a criminal background check, and they get some quick training. Their first story goes through an editor, but after that, the writers usually post directly to the site.

Page views are a key factor in determining writers’ pay, which amounts to between $1 and $7.50 per thousand views, according to AdAge, or a few dollars per article. According to Mike Noe, senior director of recruiting, fewer than a third of the writers are currently “active,” which, in Examiner parlance, means they’ve posted something to the site within the last ninety days.

The content that produces mimics much of what has traditionally appeared in the back of newspapers or at the end of broadcasts—subjects like sports, weather, hobbies, or opinion. Writers are hired in large part based on their zeal for a topic. “In a traditional newspaper, the reporter might not be passionate about the [Denver] Broncos,” says Jen Nestel, Examiner’s director of community. “We do the reverse. We take someone who is already passionate and we teach them how to write.” The site doesn’t claim to replace the newsgathering functions of traditional media: “Finding out how the school board works is hard,” says Rick Blair, chief executive officer of “It takes a special kind of digger. I could see other folks using platforms like ours to do that. But we don’t have the tools or the accredited manpower.”

For all its success in building an audience, Examiner has quite low engagement: Its readers see about 65 million page views a month, or only about three pages per visitor. That is likely tied to the site’s dependence on search engine optimization, or SEO.

“The problem with SEO is, the visitors are snackers,” says Blair. “If people come in through the front door [the home page], they read seven to eight pages. If they come in the side door [such as a search engine], they read maybe two,” he says. Suzie Austin, senior vice president for content and marketing, adds, “From the very beginning, we did search engine optimization right. The benefit is obvious—you get a lot of eyeballs. The downside is, there’s not a lot of engagement. Page views per user is growing, but at a low rate.” And as sites use SEO to boost traffic, advertisers take advantage of the flood of page views around the web to “name their price,” says Tom Woerner, Examiner’s senior vice president for national sales. There are two ways for publishers to deal with that, he says: “Play the price game, or add value to what you give the advertiser.”

So Examiner is shifting from simply selling display ads to selling the value of its ability to project stories beyond the confines of its own site. Examiner coaches its writers on deploying social media to broaden the influence of their stories. In the marketing business, using social networks is now considered a form of “earned media”—that is, it’s more akin to publicity, like an appearance in a news article, than to an advertisement or paid product placement. “Thirty years ago, if you got a story into Sports Illustrated about your product, that was ‘earned’ media because you didn’t pay for it,” says Woerner. Today, earned media includes messages that go out via Facebook, Twitter, and blogs. “Marketers have to be willing to give up a little control.” Woerner also said,“The key for traditional media is how they’re engaging with their audience. They got used to the role of the gatekeeper. They need to invite the audience in.”

To attract ads from Iams, Procter & Gamble’s pet-food company, for example, Examiner invited (but didn’t require) its writers who focus on animals to write about pet adoption and shelters; just as importantly, editors encouraged writers to distribute their stories via social networks. Iams didn’t control the content, but given how innocuous the Examiner’s coverage of animals tends to be, the company was unlikely to be troubled by photo galleries of adorable homeless puppies and feature articles about courageous German shepherds. About 840 writers responded with more than 5,200 articles and additional posts on social-media sites linking back to the stories. Those extra links from Facebook and Twitter to Examiner stories helped drive up advertising rates. Site executives say that ads sold in this effort get CPMs of more than $11, as contrasted to their usual display ads that get CPMs of $3 to $5.

Most other online news organizations are also establishing fan pages on Facebook, setting up Twitter feeds, and encouraging readers to share links. They are doing this not just because the networks are where the audiences are, but because they think social media will bring readers who are more engaged than those who come through search engines. At Gawker, Google-driven traffic “is waning,” says Pettigrew, the marketing director. Facebook is now the top referrer, and Twitter is gaining. But it wasn’t easy for Gawker management to come to terms with social media. “We didn’t want to join in the ‘fan-page game,’” she said, lest readers become more accustomed to accessing its stories from Facebook than from Gawker’s home pages. “You want to own the distribution.” But eventually, Facebook’s power as a traffic-driver won out. “You can’t ignore the way people want to access content.”

Vadim Lavrusik, former community manager at social media site Mashable, says that “readers who come through social are far different in their behaviors. They tend to view more articles on average and stick around the site longer.” Facebook and Twitter visitors spent 29 percent more time on, he said, and viewed 20 percent more pages than visitors arriving via search engines.

Similarly, at The Atlantic’s website, “The percentage of referrals from social nets is coming in at about 15 percent. And it’s growing,” says Scott Havens, vice president of digital strategy and operations. There’s a wide array of social sharing tools on, including Facebook, Twitter, Digg, and Reddit. The Atlantic has also started using Tumblr, a microblogging platform that allows anyone—from individuals to media companies—to post text, photos, and videos. It has a distinctive visual format and is another way to drive engaged traffic. also uses Tumblr, including links to a wide variety of sources. By doing that, the magazine can “introduce people to Newsweek who would never read it” on its site or in print, says Mark Coatney, who worked at Newsweek before joining Tumblr in 2010. And he says that while Newsweek’s Tumblr audience is smaller than the audience it gets through Twitter or Facebook, its readers are more engaged.

The argument about whether it’s more important to build large audiences or engaged audiences has not been settled. Two news organizations that haven’t jumped on the engagement bandwagon are New York magazine and Newser. “The notion of engagement has been touted for a number of years,” says Michael Silberman, general manager of “This is not important in driving our business. We want to grow uniques”—that is, the number of users—“so we’re really thinking about the scale. Secondarily, we want to drive page views.” He might change his mind if decided to start charging for online access, but that isn’t on the table for now. “Engagement only makes sense in a subscription model,” he says.

At Newser, an aggregator with about 2.5 million unique visitors a month, the audience breaks down in ways similar to mainstream news organizations. Executive Chairman Patrick Spain says about 12,000 users are “addicted” and come to the site many times a day; 225,000 are “avid” users who visit Newser many times a week; and more than two million people pass by, with just a click or two. But Spain argues that the passers-by are useful, because they are more likely than addicted users to click on ads, though whether clicks on ads are a good indicator of engagement is an open question.

For decades, many news organizations enjoyed higher profits dependent largely on bigger audiences. Magazines and newspapers priced their wares artificially low to boost circulation, even though that brought a group of lightly affiliated readers who had to be lured again and again with cheap come-ons and giveaways. When this worked economically, it was because advertisers could be persuaded to buy access to a big audience they didn’t know much about. Today, advertisers have far more choices and far more information. Moreover, many of the firms competing for ad dollars never would have been defined as “media companies” years ago. Facebook now delivers almost a quarter of all digital advertising views in the U.S. Search advertising, dominated by Google, soaks up almost half the dollars spent on online ads.

So it is much harder for media companies, new or old, to compete purely on audience size. They will never grow fast enough, or increase ad rates high enough, to counter the massive numbers accumulated by giants like Google and Facebook. News organizations have to offer something more.

When The New York Times announced details of its digital subscription plan in March 2011, Andrew Swinand, president of global operations for the Starcom MediaVest Group, a media-buying agency, said it wouldn’t hurt—and might help—the site’s advertising revenue. “I’m paying for an engaged audience, and if that audience is willing to pay, that demonstrates just how engaged they are,” he said. In a later interview, he added that editors need to start thinking about engagement in broader terms, not just the amount of time people spend on a site or the number of pages they click. “I want to be able to look and say, ‘Who are these people [using a site], and what are they spending their time doing on it?’” By doing that, Swinand says, news organizations can help companies feel more confident that their ad dollars are being spent wisely. The Times’s Golden adds that “if we all go the way of outdoor advertising [e.g., billboards] where it depends on who passes by, it’ll be hard to build value. Engagement is the proxy by which people value content.”

Audience size is still vitally important. A site with 10 million unique users will get more attention from advertisers and agencies than one with a fifth that many. Large companies want to make mass purchases of ads; they won’t deal individually with a host of small sites. But the chase for traffic has put news organizations on a sugar high of fat audiences and thin revenue. It has also devalued their journalism, as they have resorted to such tactics as celebrity photo slideshows to boost search-driven traffic. In diminishing their brands and commoditizing their content, they have fallen short in the crucial goal of attracting engaged, loyal users. This needs to change. By producing relevant journalism, deploying data intelligently, and relying on social media—not just search engines—to drive traffic, they can gather a more devoted and involved readership, one that advertisers will also prefer.

To continue to Chapter Three, click here.

To download this chapter as a PDF, click here.

Has America ever needed a media watchdog more than now? Help us by joining CJR today.

Bill Grueskin, Ava Seave, and Lucas Graves are the co-authors of "The Story so Far: What We Know About the Business of Digital Journalism." Grueskin is dean of academic affairs at the Columbia University Graduate School of Journalism. Seave is a principal of Quantum Media, a NYC-based consulting firm. Graves is a PhD candidate in communications at Columbia University. For further biographical details, click here.