Skip to Content View Previous Reports



The Big Picture
From a distance, people may think 2004 must have been a year of ascendancy for cable news. In September, Fox News earned enormous publicity for attracting more viewers during the last two nights of the Republican convention than any other source, including the broadcast networks. With the exception of the first nights of the Gulf War in 1991, no one could recall another moment when a cable news channel had bettered a broadcast network news program in live head-to-head coverage of a breaking news event. Was it a watershed? Perhaps.

Yet the impression that cable’s audience is ever-growing, or that 2004 was cable’s greatest year, is mistaken. Indeed, in assessing what is going on in cable audiences, four much more complicated trends stand out.

Overall Viewership

Between 1996 and 2002, cable viewership climbed fairly steadily. The usual pattern involved crises engendering growth. During major events, more people than before would gravitate to cable, and afterwards, some portion of them would continue as regular viewers. The peak and valley effect was particularly the case after the 2000 election and then after the terrorist attacks of September 11, 2001. Both events boosted what might be called cable’s core audience, the group that tended to watch day in and day out.

That growth pattern, as we reported last year, seemed to cease in 2003, even with the war in Iraq. Cable saw a huge audience spike — one of its biggest ever — but over time it lost all of it.

What happened in 2004?

For most of the year, cable news viewership fluctuated between roughly two and three million viewers. CNN, Fox News, and MSNBC were unable to hold onto the viewers they gained during the Iraq war in 2003 and subsequent major news events such as the capture of Saddam Hussein in December 2003 and the death of Ronald Reagan in June 2004.

That ceiling on the cable audience held until the party conventions and the election, which bumped viewership to roughly 4 million. The cable channels managed to keep about half of that expanded audience in November, but by December the audience was back at 2.55 million — slightly higher than the 2.47 million viewers in January 2004, but less than the 2.59 million viewers watching in December 2003, when the capture of Saddam Hussein helped spike viewership.

Prime Time Cable News Viewership
1997 to 2004
Design Your Own Chart
Source: Nielsen Media Research unpublished data

Year-To-Year Growth

There are a number of ways to calculate year-to-year ratings. Typically, the cable networks take each month of ratings and calculate an arithmetic average. That number tends to take maximum advantage of momentary audience spikes.

Given the volatility of cable news audiences, however, that may not be the best way to measure cable’s core audience, that is, the audience it gets most of the time. It also tends to punish cable in the years that there are no major news events, and reward it unduly when there are.

A more accurate way to assess cable’s core audience is to use the median — a measurement that captures the midpoint between a channel’s highest and lowest viewership.

Comparing 2004 and 2003, overall median primetime cable news viewership rose by 6%, from 2.45 million to 2.60 million. The daytime median audience rose 5%, from 1.48 million to 1.56 million.

Thus the cable news audience grew in 2004, but it is now growing at a much slower rate than just two years ago. The 6% growth for primetime audience in 2004 follows a very modest 3% median audience growth rate in 2003 over the year before. Compare that with 2002, when the median audience grew by 41%, or 2001, when it grew by 32%.

Prime Time Cable News Median Audience
All networks, 1998 to 2004
Design Your Own Chart
Source: Nielsen Media Research unpublished data

What’s more, had it not been for three months covering the conventions and the election, overall cable viewership would have been even flatter. To get a sense of the ratings boost provided by the final months of the election campaign, consider what the median audience looked like for the 12 months ending in August 2004. During that time, the total number of viewers of cable news fell by 4% from the same twelve-month period the year before.

Prime Time Cable News Audience Growth
Annual growth rate, all networks, 1999 to 2004
Design Your Own Chart
Source: Nielsen Media Research unpublished data

How would things look if audience data were calculated the way the cable channels do it, using the arithmetic mean average? Calculating the data this way, the cable news audience fell by 12% in 2004, from 3.22 million viewers in 2003 to 2.84 million. The daytime audience dropped even more, by 21%, to 1.61 million. Yet as we said before, we believe this drop overstates things because it tends to exaggerate the impact of the audience spike during the war in Iraq in the spring of 2003.

Prime Time Cable News Average Audience
All networks, 1998 to 2004
Design Your Own Chart
Source: Nielsen Media Research unpublished data

Fox’s Leadership

How are the three major cable news networks faring individually?

In 2004 Fox remained cable news’s undisputed leader in ratings, or the number of people watching at any given time.

By one way of measuring, looking at total viewership throughout the day, Rupert Murdoch’s news network generally commanded around 55% of cable news audience during the year, according to Nielsen Media Research’s measurement of “total day” viewing. This figure measures the average number of viewers over the course of an entire day — 24 hours. CNN is second in total-day viewership, with roughly 30%; MSNBC captures the remaining 15 percent.

More commonly, networks and advertisers look at ratings for different dayparts.1

Measuring ratings this way, Fox News is still the leader, and is still growing. Its median audience in prime time rose by 10% in 2004 over the year before, to 1.47 million viewers from 1.34 million. In daytime, its median also increased, by 11%, to 856,000 in 2004, up from 770,000 in 2003.

CNN remained in second place, with 815,000 median prime-time viewers, a drop of 2% from 832,000 in 2003. Its median daytime viewership, meanwhile, was 482,000, a drop of 4% from 2003.

MSNBC, still in third place, had median viewership of 341,000 in prime time, though that represented a healthy increase of 19% over the 287,000 viewers it had in 2003. Its median daytime viewership, meanwhile, barely changed, from 222,000 in 2003 to 224,000 in 2004.

Will Fox be able to sustain its growth past the election season?

From now on, to grow, Fox News must focus more on winning over viewers who already had access to it but heretofore chose not to watch, or not to watch as much. In that context, it is striking that the rate of growth in median audience has been declining since 2001. In that year, Fox’s median audience grew 113% compared with 2000. In 2002, the year it overtook CNN, the median Fox News audience grew 73%.

Since then, median audience growth has been much smaller: in 2003, the year of the war in Iraq, the median audience grew 18%. In 2004, as noted, the increase was just 10%.

Nonetheless, Fox’s audience trends are better than those of its rivals.

Prime Time Cable News Median Audience Growth
By individual network, 1998 to 2004
Design Your Own Chart
Source: Nielsen Media Research unpublished data

Audience Take Two: How Many People Watch Cable

The other issue when it comes to understanding the cable audience is the question of total audience, or how many different people watch cable through the course of the day. Ratings measure how many people are watching at any given moment, which is a fine measurement for advertisers who want to know how many eyeballs may watch a particular spot. And it makes sense for broadcast television, where each program is a unique offering.

For cable, however, in which the network provides a similar product for many hours of the day, ratings are incomplete. In trying to assess where people get their news, it is useful also to know how many different people are going there. In the TV business, this measurement is known as “cume” (as in cumulative); it’s analogous to the online industry’s measurement of “unique visitors.”

CNN has long argued that despite its lag in ratings- the number of sets tuned to it nationwide at a given time — more people over all watch it. The Project obtained monthly “cume” ratings from CNN for the fall of 2003 and all of 2004. (Viewers are counted as part of a channel’s “cume” measurement if they watch for six minutes or longer.) The data show that indeed, CNN consistently gets more “unique viewers” — the total number of viewers who tune to the channel at some point during a given month -than Fox. The pattern is as consistent as the ratings pattern: CNN holds a sizable lead in this measurement, which is called cumulative audience. Fox News is in second place, and MSNBC last. In the typical month some 64 million different people tune to CNN at some point. Fox generally attracts 56 million individual viewers. MSNBC lags behind with 48 million.

The pattern of cumulative viewership showed CNN leading Fox News by a margin of 5 million viewers or more for most of 2004. The gap narrowed slightly in September and October but widened again in November. CNN reached its highest cume in recent months in December 2003, the month of Saddam Hussein’s capture; in all, 77.8 million different people watched CNN that month, according to Nielsen.

Cumulative Cable News Audience
October 2003 to December 2004
Design Your Own Chart
Source: Nielsen Media Research unpublished data

The numbers suggest that CNN is still the first choice for people trying to get a fix on the latest breaking news. For example, on Election Day, CNN beat out Fox in unique viewers by 6 million — 38 million for CNN, 32 million for Fox. MSNBC was far behind with 19 million. Given that CNN is the default option for so many people looking for cable news, the size of Fox News’s ratings margin over CNN is particularly striking.

CNN argues that its higher cume makes it a better choice for advertisers because its ads will reach more people over the course of a given day. Fox News argues that CNN’s “cume” figures don’t matter because they’re simply a measurement of “channel surfers.” Indeed, in advertisements in trade magazines that reach the media planners who make decisions about where to place ad budgets, Fox has taken that argument a step further by arguing, “If CNN’s advertising is misleading, why would you trust their journalism?”2

Last year, while we didn’t have CNN’s “cume” figures, we found evidence to support its argument in survey data. More people told pollsters that they watched CNN than Fox, which seemed to support the cumulative audience idea.

This year, however, the survey data support Fox’s hold on viewers.

The Pew Research Center’s 2004 poll on media consumption showed for the first time that Fox News had surpassed CNN as the preferred outlet for cable news. Fox News was cited as a “regular” source of news by 25% of people in Pew’s latest survey, up from 22% in 2002; CNN dropped three points, to 22%, from 25% in 2002.

Types of Cable News Viewers
By network, June 2004
Design Your Own Chart
Source: Pew Research Center for the People and the Press, “News Audiences Increasingly Politicized,” June 6, 2004

So which measure is most meaningful? Ratings, CNN’s “cume” data or survey research? Each yardstick has its own strengths and weaknesses. Research conducted by an academic team at Ball State University in 2004- involving surveys, consumers filling out “diaries,” and researchers observing media use first hand — also found that survey data appear to consistently undercount media use. People simply use media more than they tell or can recall when surveyed, by magnitudes of three and four times.3

Ratings are the closest thing advertisers have to determining how many eyeballs may see an ad at a given moment.

Taken together, however, the data suggest that the picture is more complicated than ratings alone suggest. If you want to know which network more people watch, Fox and CNN apparently are much closer than if you ask which network has higher ratings at any given moment.

This gap between CNN’s ability to attract an audience and its inability to keep viewers around for the long haul is the main challenge facing the network. But it is also a sign that the cable audience is not monolithic. The data suggest that it is worth looking at cable news in terms of two separate audiences: the day-to-day core audience and the occasional, “news on demand” audience. While Fox News has a larger core audience, CNN may be the winner when it comes to the “news on demand” people.

Cable Partisanship

In 2004, there was also growing evidence that the cable news audience was splintering along partisan lines. In particular, viewership of Fox News leans toward Republicans. To a lesser degree, CNN’s viewership tilts towards Democrats. Democrats and Republicans are equally likely to say they watch MSNBC.

The 2004 media consumption survey by the Pew Research Center for the People & the Press finds that 35% of Republicans say they are regular Fox News viewers, compared with just 21% of Democrats. Meanwhile, 28% of Democrats are regular CNN viewers, while only 19% of Republicans say they watch it regularly.

Percentage of Party Members Watching Specific Cable News Channels

Democrats Republicans Independents
CNN 28% 19% 22%
Fox News 21% 35% 22%
MSNBC 12% 10% 12%

Source: Pew Research Center for the People and the Press, “News Audiences Increasingly Politicized,” June 6, 2004

Much was made of those statistics in 2004, along with data showing more Republicans losing trust in the media generally. It was suggested that Americans were moving to their own ideological corners in their media consumption, that we were moving toward “Red and Blue truth,” in the words of Time Magazine, or a more European style of ideological media. Perhaps the American model of a nonideological independent press was dying.

But the data suggest to us something more nuanced and less spectacular. The polarization phenomenon tends to occur primarily within the cable news audience, and not necessarily across the entire television news spectrum. A broad look (see Overview) shows that this ideological splintering exists in cable as nowhere else, and exists more at Fox News than anywhere else. Indeed, MSNBC’s audience is evenly divided, and CNN, while it has lost Republicans to Fox, has almost as many independents as it does Democrats. In addition, Democrats are almost as likely to watch Fox as CNN.

Democrats simply don’t watch cable as often as Republicans. According to the Pew Research Center’s 2004 survey, while nearly half (46%) of Republicans are “regular” cable news viewers, only 36% of Democrats are “regulars.” Democrats are more likely to be “occasional” cable news viewers than Republicans (36% vs. 27%).

Republicans as Percent of Cable Channel Audience
1998 to 2004
Design Your Own Chart
Source: Pew Research Center for the People and the Press, “News Audiences Increasingly Politicized,” June 6, 2004

Within the specific cable channels, what appears to be happening over time is a migration of Republicans to Fox. In a breakdown of Fox’s audience by party affiliation, the percentage that is Republican has increased considerably over the past six years, from 24% in 1998 to 41% in 2004. But there has not been a similar migration of Democrats to CNN or anywhere else.4

The data tell us something fascinating about cable and how it has come to resemble talk radio not only in content but in appeal. But they suggest far less about some growing trends in the media over all.

The On-Demand Viewer

The Pew Research Center’s studies suggest that cable news has consolidated a core audience of viewers but that it is occasionally able to draw additional viewers who seek out news when significant events are happening. Those people could be described as “news on demand” viewers. Rather than routinely watching a TV news program at a given time of day, they tune in when events pique their interest.

The surveys suggest that this “news on demand” pattern is particularly true for younger viewers. Adults under age 30 who “regularly” watch television news are more likely to watch cable than the broadcast networks by 29% to 18%. This preference for cable among younger viewers is important. Rather than setting aside time to watch the network news at a specific hour — what’s known as “appointment television” — younger adults are more likely to go to cable, which is available any time they choose to tune in. And looking more closely at those cable viewing habits, there is evidence of the same news on demand behavior. Younger cable viewers are more likely than other groups to only watch on occasion, presumably when something is happening, rather than as a regular habit. The plurality of young cable viewers, 37%, describe themselves as “occasional” viewers, the highest of any age group in that category. Moreover, it’s been that way for quite some time.

The bulk of the cable news audience, however, is made up of older Americans, who in general consume more news than younger ones (with the exception of online news). Their responses are the mirror image of the younger group’s — they call themselves regular consumers of cable rather than occasional but ultimately prefer network. Among people over 65 years old, 46% are “regular” cable news viewers, but 57% are “regular” network news viewers.5

Regular TV News Viewers, Cable vs. Network
By age group, June 2004
Design Your Own Chart
Source: Pew Research Center for the People and the Press, “News Audiences Increasingly Politicized,” June 6, 2004

Does Cable News Have Room To Grow?

The question for the future is whether cable’s big growth years are over.

Media analyst Tom Wolzien of Sanford C. Bernstein, analyzing Nielsen data on 50 of the most-watched cable networks, noted that from 2000 to 2004 their cumulative share of the cable TV audience remained static. The growth in viewership has come in the acquisition of cable service by more and more households. Viewership remained flat among consumers who already had cable service throughout the period. Fox News in particular seems to have benefited from increased distribution.

In an interview with USA Today, Wolzien suggested that the cable networks are now “cannibalizing” from each other rather than winning viewers from broadcast. He predicted that for cable in general, barring better programming and more investment, the size of the audience would peak in 2009.6

Fox News’s growth to date seems to represent both phenomena — adding more cable systems and stealing viewers from its rivals. Consider that in 2002, the median prime-time cable audience was 2.37 million viewers, and 48% of it was tuned to Fox News. Two years later, the median primetime cable audience was only slightly more, 2.61 million viewers, but now 57% of it was watching Fox.

The overall audience did grow over those two years, slightly, but Fox’s share grew even more, a sign of cannibalizing its cable rivals.7

In the same period, Fox gained more audience than CNN lost, a sign that some of that audience also came from growing distribution. Indeed, between 2000 and 2004 MSNBC extended its potential reach of new cable systems by almost 30 million cable subscribers, while Fox News added closer to 40 million. CNN, meanwhile, which was already carried on most cable systems, gained only 10 million more potential subscribers over those four years.

Availability of Cable News Channels
1989 to 2004, at year end
Design Your Own Chart
Source: Kagan Research unpublished data,

The Election Effect

Cable news was one of the most-cited sources for election news throughout the presidential campaign. A January 2004 study by Pew Research Center asked people, “How often, if ever, do you learn something about the presidential campaign or the candidates” and offered a list of specific news sources. The survey found that cable news was one of the few news sources people were likely to rely on more than they did four years earlier — 38%, up from 34% in 2000.8

The late summer and early fall of 2004 showed growth in cable news viewership, but by October the audience had peaked. Indeed, Pew Research Center surveys show that by the final months of the presidential campaign the number of people turning to cable news as a source for election news was flat or had fallen.9 Fox News was the only network to hold steady, cited as a source by 20% of Americans in January and 21% in November. CNN, which had been the most popular cable source in January (22%), was the choice of only 15% of Americans surveyed in November; MSNBC, the choice of 7% in January, was named as a source by 6% in November.

The sharp, partisan divides between audiences for the three cable channels that became clear in 2004 -Fox News clearly the channel preferred by Republicans, CNN viewers more likely to be Democratic -showed up starkly in ratings for the political conventions, when the parties most plainly presented their case to the American public. Those ratings suggest that the division may be more one-sided than it first appears. Republicans clearly migrated to Fox, but audience figures suggest that Democrats were not as clearly aligned with CNN. A more thorough discussion of this phenomenon follows in Public Attitudes.

During the Democratic convention, viewership was distributed fairly evenly (in fact, more evenly than usual) among the three cable channels: CNN (plus Headline News) was the most-watched channel, with 43% of the cable audience over the course of the convention; Fox News was watched by roughly a third (35%) of the audience; and MSNBC, which is typically watched by 15% of the cable news audience, was watched by at least a fifth (22%).10 Cable viewership increased from 4.8 million on the convention’s first night to 6.8 million on the Thursday night of John Kerry’s acceptance speech.

Democratic Convention Viewership on Cable News
Design Your Own Chart
Source: Nielsen Media Research unpublished data
*Average prime time viewers, July 26 to July 29, 2004.

The pattern was remarkably different during the Republican convention a month later. From the first night, Fox News dominated the fight for viewers, attracting 65% of the cable audience that night and maintaining its lead over the next three. In addition, Fox News steadily picked up viewers, going from 3.9 million on the first night (out of 5.98 million cable viewers) to 5.3 million (out of 8.55 million cable viewers) on Thursday. On the final nights of the convention, indeed, it was the single most-watched network, beating the broadcast networks as well as its cable rivals.

Republican Convention Viewership on Cable News
Design Your Own Chart
Source: Nielsen Media Research unpublished data
* Average prime time viewers, August 30 to September 2, 2004.

While total viewership on both network and cable television was roughly the same between the Democratic and Republican conventions, Fox News’s surge in viewership during the Republican convention seemed due in part to partisan Republicans’ flocking to that network.

Convention Viewership on Cable News
Design Your Own Chart
Source: Nielsen Media Research unpublished data
* Average prime time viewers, July 26 to July 29, 2004 and August 30 to September 2, 2004.

(For a discussion of the substance of cable news convention coverage, see News Investment.)

Another factor was probably the general abdication from convention coverage by the broadcast networks. NBC, CBS and ABC each devoted only three prime-time hours to each convention, one hour on each of three nights, and skipped one night of both conventions altogether). That absence of broadcast network coverage contributed to the cable numbers.

Audience: A Conclusion

So what is the future for cable audiences? Most likely it will not be based on drawing dramatically new numbers of people freshly wired to cable boxes. Rather, it will have to be built around establishing viewer loyalty — by building its own dedicated audience, cannibalizing from the news competition, and winning more of the ‘information’ audience that drifts to channels like Discovery, History, or TLC.

That outlook has a number of implications.

One is that to grow further, cable will have to do more than the things Fox has done so far — attract an audience with talk radio-style programming and steal conservatives away from CNN. It will have to attract new viewers with new kinds of programming.

Another is that cable executives will increasingly focus not just on numbers of viewers but on something called “time viewed” — the length of time those viewers hang around — as well as demographics.

When it comes to time viewed, the implications for the future may not be good for news, or at least what traditionalists would consider news. The “traditional newscast” at best would attract viewers to a single program, not over time. Yet to date, the signature newscast (MSNBC originally had high hopes for “The News with Brian Williams”) has been a ratings loser in cable compared with the talk formats (like Larry King, O’Reilly, or Matthews). If that continues to be true, there will be even more emphasis on personality. That’s what’s working, that’s what’s winning and that’s what viewer loyalty seems to be built on. Here, Fox seems to have had the upper hand so far. It has more programs built around personalities, and it apparently has viewers sticking around longer. If personality is the path, then the Fox approach would seem to be the model others will copy.

Another question is how you define success. Any discussion of the future “growth” of the cable audience will likely break audience growth down into demographics. All ratings points are not created equal; viewers with higher incomes are more desirable to advertisers. That, industry insiders explain, is how Lou Dobbs’s Moneyline program was sold for so many years. It had relatively low ratings, but charged high ad rates because of the presumed desirability of the audience. If cable cannot increase the sheer numbers, can it increase its appeal to specific demographics? It would like to do so, but it has shown little boldness to date in going upscale, for fear of losing the overall ratings wars. Here, potentially, CNN and MSNBC may have an opening to do something Fox has not.

Indeed, the small percentage increases and declines in audience growth CNN and MSNBC have experienced illustrate the problem facing both channels as they compete against Fox News. Neither appears to have leadership willing to take a risk on an approach to the news that might prove either a breakout success or a ratings disaster. Instead, by attempting variations of what Fox has shown is successful, both channels in their own distinct ways have failed to create their own distinct formulas.

CNN has moved in the direction of talk shows, à la Fox, but has never gone for the ideological edge. MSNBC has also moved in that direction, but the political stances have varied — sometimes to the left (“Donahue”), sometimes to the right (“Scarborough Country”), and sometimes both. During the daytime, MSNBC has shifted toward a BBC or radio style “newswheel,” with regular previews of what is coming in the next 15 minutes, and regular news summaries.

If CNN and MSNBC do not change their content substantially, there would seem to be little reason to believe their ratings will change in any dramatic way. Yet if Fox News’s growth continues to slow, or even stop, it is possible that its formula of talk programs may no longer be the model the others try to emulate.


1. We focus on two dayparts: daytime, lasting from 6 a.m. to 6 p.m., and prime time, which for cable usually starts at 7 p.m. and goes until 11 p.m.

2. Advertisement, Mediaweek, November 22, 2004.

3. A comprehensive report on this project, known as “Middletown Media Studies,” can be found in the International Digital Media & Arts Association Journal, vol. 1, no. 1 (Spring 2004). For a summary of the project’s research on news media consumption, see Deborah Potter, “TV: The 800-pound gorilla,”, September 22, 2004.

4. Pew Research Center unpublished data.

5. Pew Research Center, “News audiences increasingly polarized,” June 8, 2004, Media Tables, p. 7.

6. David Lieberman, “Study: Cable losing steam,” USA Today, August 24, 2004.

7. Kagan Research provides data on each network’s clearances-the number of homes in which the channels are available.

8. Pew Research Center, “Cable and Internet loom large in fragmented political news universe” questionnaire, January 11, 2004, question 18.

9. During the presidential campaign, the question asked “Do you get most of your news about the presidential election campaign from [news outlet]?” The November version asked “Did you get most of your news about the presidential election campaign from [news outlet]?”

10. Nielsen Media Research unpublished data.