Skip to Content View Previous Reports



News Web sites still defy hard classification. No formula or set of models has set in. That, indeed, constitutes one of the findings of this study.

The universe is changing so rapidly that of the 38 Web sites examined, at least a quarter were either thoroughly redesigned or made noticeable changes between September 2006 and February 2007, usually to make them more user-centric.

The field is marked by experimentation, and in some cases noisy crowding.

A few sites even now are still largely “shovel ware,” an online morgue for the content their owners produced in another medium.

Other sites are made up of a few packages for top stories and then largely wire copy after that.

Even so, we did not find sites that scored poorly across the board.

On the other hand, we found no sites that excelled at everything. That may reflect the fact that the Web is so rich in possibilities that sites need to make choices. The greater the focus on speed and immediacy, the harder it is to take the time to build depth into coverage — multiple links in story packages to background material, documents, full text of interviews, archives and more.

Indeed, every site studied except two scored in the lowest tier for at least one of the five areas of content we examined. And no site scored in one of the lowest two tiers for everything.

What qualities of the Web’s potential are being exploited most?

1. User Customization

The Web allows for a nearly infinite array of style and content, a level of choice that can overwhelm. Hence a growing premium is now placed on the degree to which users can customize content to their interests, pre-select the stories that come their way or the form they come in. We called this User Customization.

In general, there are two types of customization. People can tailor the design of the page itself (Web site customization). Or they can choose to have different kinds of content delivered to them from the page, including RSS (Real Simple Syndication), podcasts, mobile phone delivery and more (delivery customization). We examined sites for both.

Allowing visitors the ability to pick and choose what they were interested in or tailor its delivery appears to be an effort the news Web sites in our sample have focused intently on. After branding or editorial control, a high degree of customization of material was the second-most-developed potential we found in online journalism. Twelve of the 38 sites were highly customizable. (To be so designated, they possessed at least five of the six elements we examined).

There was little pattern about what kinds of sites fit into this grouping. Their creators ranged from online-only entities like Global Voices and OhmyNews International to the weekly magazine the Economist to NPR to the local station King 5 TV, in Seattle.

Some kinds of customizability, moreover, were more popular than others. The move now appears to be toward making content come to the user. The features the sites were most likely to offer were multiple RSS feeds, usually prominently displayed and podcast options (though sometimes not as prominently displayed). And many of the site upgrades in early 2007 had to do with adding some kind of mobile phone delivery.

Sites also tended to emphasize advanced methods for finding a specific news story.

Interestingly, one feature more likely to be absent, even in these sites that scored well in customizing, was the option to customize the homepage story layout. About half of the highly customizable sites (and half of the sites overall) did not offer any kind of flexibility here.

Apparently, for now, the ability to have content sent to you, or to find what you want, is taking precedence over letting people make a page theirs.

On the other side of the spectrum, just three sites fell into the lowest tier of customizability — offering nothing more than a simple keyword search. Visitors to (the Web-based local “newspaper” in Northern California), to (the Web site of the latest weekly news magazine phenomenon), and to, the pugnacious alternative weekly of San Francisco Bay, had to accept it as it came to them.

No other content category had so many sites scoring so well.

2. User Participation

One of the chief appeals of the Web early on was the notion that online media would become a dialogue, not a lecture, in which the user could speak for himself or herself. That is one reason Time magazine made everyone in 2006 the Person of the Year (a feat it accomplished by showing a small mirror on the cover).

Potentially, the Web offers many ways to accommodate participation — everything from a simple e-mail link to having a story’s author post user content as a part of the story mix.

What we found in the sites studied is that the participatory nature of the Web is more theoretical than a virtue in full bloom.

We examined 10 different features that broke participation into two different types. One was the extent to which people could express their Individual Voice. That included offering e-mail, writing blogs, commenting on stories, rating them or entering a live discussion, or even taking an online vote on a question. The other was the extent to which users were heard from in a site through a Group Voice, such as by tracking of the most e-mailed or most viewed stories and then featuring those lists on a site.

Just three sites, the blog Daily Kos, the citizen-based site called Digg and AOL News, possessed enough of those features to earn top marks for participation.

On the other hand, a dozen sites studied earned the lowest marks, with no user content, no live discussions, rating of news stories, or compilation of the most viewed or e-mailed stories of the day. On the bulk of those sites, visitors could not even e-mail the author of a news story to comment or raise questions. Another 10 sites earned the second-lowest marks in participation.

Most news sites, whether stemming from traditional media outlets or not, place a high premium on reported news stories and keep control over their selection (and sometimes creation).

Visitors are sometimes invited to express themselves by responding to the stories through user comments or e-mails to the author of a bylined story. But those features were not standard, even among sites that scored at the higher levels for participation.

What the higher-scoring sites were more uniform on was tracking the Group Voice — a list of most viewed, most e-mailed or most linked stories. All the sites in the top two tiers for participation had at least one such list that users could access, and some offered all three. In the lowest tier for participation, however, this option was completely absent.

3. Use of MultiMedia

The third major area of Internet potential is the fact that the platform works, at least theoretically, for all media formats — video, audio and text. How much are sites exploiting that? Are most sites still expressions of their own roots, with TV sites more video-oriented but not as rich with text, and print sites the opposite? Is there any pattern to what kinds of sites are doing more here than others?

To get answers, we catalogued all the content on a homepage (as text there or linked to items) for 11 different media options. We then noted the percent of the content devoted to each of these media forms to get a sense of which type the site was emphasizing.

What we found was that the multimedia potential of the Web is also not as developed on many sites as people might imagine. Only six of the 38 sites earned top marks for offering a rich range of media formats.

And nearly half the sites (17) earned the lowest marks. For those, more than 75% of their content was just narrative text — “Still Reading the News” sites.

The ones in that last group were not all from traditional print outlets. They ranged from the news pages of the aggregators Google and Yahoo, to the blog Michelle Malkin, to the citizen-based sites Digg and OhmyNews International to newspaper sites like the San Francisco Bay Guardian and the New York Times. Only one — PBS’s Online NewsHour — offered less than half the content as narrative.

For those sites with at least a quarter of the content something other than text, what kind of media form were they using? For most, the next-biggest medium used was an older one, still photos. Nearly a quarter of the sites filled at least 20% of their homepages with pictures.

And how many sites were really multi media, or used at least five different media in addition to text and still photo? Just six.

Most in that group were TV-based sites — ABC, CBS, BBC, Fox News — but also included and the site of a local Washington radio station, WTOP. The media used tended to be slide shows, interactive graphics, and live streaming video. But none of those accounted for more than 4% of the overall content on a site.

In short, the Web, for now, is still largely dominated by the content that fills newspapers — text and still images.

4. Site Depth

Another potential of the Web is its infinite depth. It can to link to past reports, biographies or referenced documents, graphically display certain elements, offer analysis and bring in outside insights.

Depth is also in some ways the hardest potential to measure. A related link may add important information or insight to the main report or it may mostly repeat what was already said. To get a sense, at least, of the extent to which news Web sites try to broaden their coverage, we looked at four different features: how frequency a site was updated, the number of related story links it offered with its lead story, the use of archive material, and the use of links inside news stories.

As a rule, sites scored lowest in depth than any other area studied. Nearly half (18) earned the lowest marks for depth, another 16 fell in the next-lowest tier, and only three earned top marks.

And, as noted above, one site was unique in this category: Google. No other even came near Google’s average of 900+ related links attached to the lead news story. And every headline down the page gets this treatment.1 In a sense, Google defines an extreme, but a powerful potential of the Web.

Visitors could spend the good part of a day just following the links for a single news story. If someone were to actually do that, though, the value might be disappointing. With no editing process, related stories automatically pop up from all different outlets. In some cases the reports are nearly identical wire stories carried in different outlets. In other cases, a later link is to a report that was written before the main story and thus has old or incomplete information. All the stories are in narrative form with occasional photos attached. (As for links inside the stories, Google was not scored here since that content is not their own.)

After Google, there was quite a drop-off. The other two sites at the high end were Global Voices and, both of which had more than 10 related links, as well as links to archive information and frequent updating. Global Voices also embedded links into the news stories themselves, while did not. Many sites (10 out of 38) still treat even the lead stories as stand-alone reports, without even one related link as normal practice.

Inside the content itself, news sites were even less likely to offer consumers links to additional information, either on their own site or from another place. More than half (17) contained no links inside their top 4 stories.

5. Editorial Branding

More than any other quality, sites built themselves around the idea that their organization’s standards, judgment, and professionalism are the core of the site’s brand.

In other words, the notion that the Web has no standards, no professional rules of conduct or editing, is not true when it comes to sites connected to traditional news organizations, to many blog sites studied, or to many citizen-media sites.

Critical to the notion of branding in our study was whether a site was promoting its organization’s particular content, had discernible editorial standards and promoted its staff with the use of bylines.

We looked at three distinct elements:

  • The range of sources and originality of the content
  • The level of staff control over the editorial process
  • The use of bylines in top stories

We found that those values still dominate, and that is true even of some of the most innovative and user-driven sites we studied.

Across the three different measures nearly two-thirds of the sites studied (24) earned top marks for emphasizing their own brand and standards. That is more than any other content area that sites emphasized.

And all but five of the sites studied had some in-house editorial process they exercised to select stories. That was sometimes combined with user input like story ratings or a list of the most-linked-to stories, but staff people made daily decisions about what to post.

Even the fairly sophisticated citizen news site OhmyNews International, with 100% “user” content, has a heavy editing process of the content that comes in from approved “contributors” from around the world. The same was true of Global Voices, another citizen-media site.

Some of the sites have no editorial branding of their own and instead rely on the established brand identity of other outlets that they present second hand. At Google, AOL News or, for example, most of the content comes from establishment news organizations.

The majority of sites studied did offer some content through their own brand name. The more traditionally rooted sites usually also made some use of wire reports, though the mix of original to wire varied greatly. Some sites, like those of MSNBC and USA Today, relied much more on wire than their own work. Others, like the New York Times and BBC News, primarily featured staff reports. And, as mentioned above, some of the newer news options contained the greatest degree of original work, along with stringent editorial practices.

6. Revenue Streams

Increasingly, a fundamental question is whether the Web can subsidize journalism, and at what level? Sites have struggled with getting people to pay for content. There is growing concern about how ads work online. Are too many ads counterproductive? There are different kinds of advertising, not to mention premium areas of paid content and registration which is free but often sends consumer information to the site and to advertisers.

To understand all this, we looked at three potential revenue streams. First, did the sites include ads, what kind of ads, and how numerous were they? Next, what did the site demand of the user: payment for certain content or registration, or could visitors roam free, other than leaving their digital fingerprint?

On the days we analyzed, the number of ads that greeted a visitor varied widely, though nearly all the sites studied had some ads on the homepage. We also found that the number of ads on a site was the element that varied the most from September 2006 to February 2007, sometimes higher and sometimes lower. Perhaps that speaks to the still experimental nature of economic models online. (Our scores for the sites reflect the February download, except when the variety seemed to be simple day-to-day variance rather than policy changes in which case we took an average of the figures.)

The only site that was completely ad-free was the news page of the revenue giant Google. The aggregator’s main search page — and the company as a whole — is largely structured around advertising, but for now anyway it has kept the news pages ad-free. The other aggregator studied, Yahoo, began placing some ads on the main news page in the fall of 2006. Even the two government-funded sites — and — contained ads from their corporate sponsors. As logo links to the corporate Web sites, though, those ads were much less intrusive than those found on many other sites.

All in all, the more traditionally rooted sites were at the top of the pack for total ads on the homepage. The Des Moines Register led with 25 (nearly all of which were external as opposed to self-promotional), followed by the New York Times, Fox News and the Washington Post. Two blogs, Little Green Footballs and Daily Kos, had mid-to-high levels of ads, as did, the Web site of the local Washington radio station. The site with the greatest display of self-promotional ads was, with an average of 14 against just 3 from outside companies.

There are still some places where users can get the news without first giving away their own personal diary. In fact, there seem to be quite a few places. Close to half the sites studied had no registration process (not even a voluntary one) and offered all content on the site free, including all archive content. None of the sites required registration at the outset, though many prompted you to on a voluntary basis.

Premium content, the kind requiring payment for specific areas, is also rare, with just four sites featuring some of it on the homepage. Even a bit more surprisingly, the practice of charging for content that is more than a week or two old is also not widespread. On 32 of the 38 sites, users could search and access more than a month’s worth of old content at no charge. One site,, charges for all archive material, while the others offer the first week or two for free and then impose a fee.

Yet, most sites are limited to ads, user registration, or some combination of both.

The least common economic group was sites with no user requirements and less than five ads on the homepage. Six sites fit this bill and ranged in character from a publicly funded site to an aggregator to a local newspaper.

Consumers have more choice if they are willing to click through a few more ads in order to escape registration or fees. Eleven sites had no user requirements but between 6 and 10 ads on the homepage. But alas, as organizations seek to figure out how to succeed financially online, revenue streams will be an area likely to change, and will be worth watching closely.

News Web Site Groupings

Our study led us to conclude that it is probably too early in the history of news Web sites to develop a firm typology, or set of classifications, for them. Also, the study included 38 sites, a number that is hardly definitive. Still, we offer five tentative groupings.

High Achievers

Only a few of the sites studied excelled across more than two of the content areas we studied. They might be called High Achievers, sites that scored in the highest possible tier for at least three of the five content areas.

Only four of the sites qualified, and they had little in common beyond the breadth of what they offered. They were a network TV site (CBS), a newspaper (Washington Post), a British television and radio operation (BBC) and an international citizen media site (Global Voices).

And what did these sites emphasize? All of them scored highly for the originality of their content. All of them also scored highly for the extent to which they allowed users to customize the content, to make the sites their own or make the content mobile. None of them, interestingly, scored particularly well at allowing users to participate. Only two, CBS News and the Washington Post, involved a lot of multimedia components.

The Original Brand Crowd

Another grouping would be those sites that promote their own original content above all. Call them The Original Brand Crowd. In every case, those sites scored in the highest range for the degree to which they controlled and promoted their content or editorial judgment. What that content was varied widely in style. The sites ranged from a number of daily newspapers, a public television station (the NewsHour on PBS), a news service, (Reuters), and in several cases blogs. The editorial judgment and standards here may vary widely. So may, in the judgment of some, the quality. Yet it was their judgment, their approach, they emphasized most.

Some of those sites were offering little more than what they had published in their newspaper, so-called shovel ware (such as The Week). Others, such as the New York Times, offered a good deal of content that was updated often and that had not yet appeared anywhere else. For all of them, though, their appeal, in the end, is what their writers have to say, and their standards, their practices, their content. This was also the largest category of sites.

In all, 16 of the 38 sites studied fell into this group. The 16 were the Web sites of the New York Times, the Chicago Sun Times, the New York Post, the Des Moines Register, the Economist, NewsHour on PBS, the Boston Phoenix, Reuters, Salon, the San Francisco Bay Guardian, Little Green Footballs, the blog site Michelle Malkin, The Week, the online magazine Salon, Crooks and Liars, a blog that features video, and the citizen-media site ohmyNews International.

Us and You

A third grouping of sites involves those that earned their highest scores (and perhaps were building their appeal) around a combination of two categories: the branding of their content and the ability of users to interact with it: the Us and You sites.

Many of them were just as strong as The Original Content Crowd in producing and promoting their own brand standards for the news. But these sites have also put a major emphasis on allowing users to do more. In most cases, that meant offering users the ability to customize the material.

Some venerable journalism names fit this grouping. What places them here is their willingness to give up agenda-setting and let users decide what they consider important. These include Time magazine and National Public Radio, the online-only site Slate, a local TV station (King5 TV in Seattle) and Daily Kos, a liberal political blog.

Jacks Of All Trades

The second-largest number of sites of those studied form a group that does not excel at one thing but tried to manage most or all of the categories. They may produce some original content, but don’t stand out for doing so. They received the lowest possible grade, evincing little or no effort, in no more than one category. They are, in other words, demonstrating skills across the range of Web potential.

Six of our three dozen sites fell into in this grouping: Yahoo, USA Today, CBS 11 TV, in Dallas, MSNBC, CNN, and Crooks and Liars. Interestingly, these included three of the four top Web sites in overall traffic (Yahoo, MSNBC and CNN).


A fifth grouping of the sites studied includes those that earned their higher marks or put most emphasis on letting the user control the material, and thus might be called User-Centric. That could mean either letting the user customize the material or interact with it directly by producing material or commenting on it. The sites scored higher in those areas than in creating material. Six sites fit here. Three scored well in both participation and customization: Digg, a site where users submit content and ranks stories; Topix, which aggregates local and world news stories on one site; and AOL News. Three others scored their highest marks for offering multi-media content and then allowing people to customize that: Fox News, WTOP, and Benicia, a local news site in Benicia Calif., which relies heavily on bloggers.