|By the Project for Excellence in Journalism
Though it remains America’s most popular choice, local TV news’s core broadcast audience is now decreasing sharply.
That may not distinguish the medium from other traditional news platforms. But in recent years we had seen some signals that the loss of audience might have stopped. The new data suggest that the hope, at least for 2006, evaporated. Even morning news, the medium’s growth area, lost viewership.
For this report, to analyze audience we examined ratings and share data from four different sweeps months — February, May, July and November, the periods stations use to set their rates for advertising.1 In previous years, we relied on only one sweeps period, a key one in May.
What we found was downward trends in every time of day, in every time of the year.
The key metrics for audience in television are ratings (the number of households watching a program at a given time among all households in the market) and share (the number of households watching a particular program among those households that have their TV sets on). Ratings give you a number for a program’s average audience. Share tells you the percentage of TV viewers at that moment who are watching that program. We found that both were dropping throughout the year and throughout the day.2
Local News: Declines in Ratings
Source: Nielsen Media Research, used under license
Local News: Declines in Share
Source: Nielsen Media Research, used under license
Early Evening News
Early evening news, the traditional dinner-time newscasts, saw a loss in ratings and share in every sweeps month. Those programs, indeed, seem to be bearing the brunt of changes in consumer lifestyles and viewing habits — people getting home too late to catch the news or not tuning in to the news even when they are at home.
Year to year, ratings fell for almost for every sweep month in 2006, especially earlier in the year. The number was down 8% in February, down 9% in May, flat in July and down 3% in November.
Share fell by even steeper rates, save for February sweeps (when it was unchanged), dropping 14% in July and 7% in May and November. That means that the problem is not simply people turning off the TV. People are choosing to watch other things. In past years, when other stations were adding news, such a decline might have meant that people were going to those places for news. There was less evidence in 2006, however, of new stations entering the field.
The problems are severe enough that some stations are even doing away with their earlier newscasts, shifting them to start later. For the Gannett group, that proved to be a successful strategy. In Miami, its ABC affiliate WPLG-TV shifted its 5 p.m. newscast to 6 p.m. and replaced it at 5 p.m. with more popular programming (Dr. Phil). The ratings for the time slot improved, and the stronger lead-in helped ratings for the new 6 p.m. newscast as well. What’s more, the change was accomplished with no loss of news programming. Two of Gannett’s other stations, in Cleveland and Atlanta, made similar switches.3
If early news was suffering enough to make programmers begin to re-evaluate their newscasts, late news, the period after prime time, had even more trouble. According to analysts, people are up earlier, home later or in bed earlier — boding poorly for the late-night newscasts.
Ratings were down in 2006 from the year before in every sweeps period. They fell a striking 14% in February, then 6% in May, 4% in July and 6% again in November.
And the story was even darker for share, suggesting that more people are choosing to watch something other than news.
Not only was every sweeps month down from a year earlier in share in late-night, but most months fell by double-digit percentages. November 2006 was down 20% from a year earlier, July 17%, May 7% and February 13%.
One newscast that seemed particularly hard hit was the 11 p.m. one, which airs on both the East and West Coasts (Mountain and Central late news starts at 10 p.m.). Examples of the slump could be found in some of the biggest television markets. In New York, two of the three main local stations, WABC and WNBC, saw their 11 p.m. news ratings drop 12%; the ratings of the third station, WCBS, rose less than one point. In Washington, D.C., all three local stations (WRC, WJLA and WUSA) saw late-night newscast ratings drop more than 10% in the 2006 November sweeps compared to the year before.4 Similar reports came from Philadelphia and from one Central time zone market, Chicago.5
Like the experiments in the evening news slot, several news directors seem to think late-night viewers might be found an hour earlier instead. Fox Television has been running its late-night newscast at 10 p.m. for years. Now some affiliates of the newly created CW Network (a merger of Warner Bros. and UPN) are offering it some competition. The CW network, like Fox, stops prime-time programming at 10 p.m.
In top markets like Chicago and Boston as well as smaller ones like Richmond, Va., and Yakima, Wash., local stations introduced or made plans for 10 p.m. newscasts, even despite stiff competition. General manager Peter Maroney of Richmond’s WTVR, a CBS affiliate, is partnering with the CW affiliate WUPV, providing the latter with resources for a10 p.m. newscast planned for March 2007. He believes that “a lot of people want to get their news at that hour.”6
In Chicago, Fox-owned WFLD was planning a 10 p.m. newscast in March 2007 — going up against three other local newscasts and two Spanish-language newscasts.7
The story is similar in Boston, where in December 2006 the NBC affiliate WHDH-TV began producing a 10 p.m. newscast to air on the CW affiliate WLVI-TV. Both stations are owned by the Sunbeam television group. The 10 p.m. program on WLVI, an earlier edition of WHDH’s 11 p.m. newscast, will challenge the existing Fox newscast that is the time slot’s market leader.8 Those are examples of a single newsroom’s producing newscasts for more than one station, a trend that the researcher Bob Papper of Ball State University in Indiana estimates is occurring at 150 stations nationwide (see News Investment).
Local News in the Morning
Perhaps the starkest finding in the data concerns mornings. The time slot before the network news programs come on at 7 a.m. has been one of the bright spots for local news. While small compared to the evenings, audiences had been growing. The broader data set from 2006 suggests that that growth too, has ended, at least for now.
Ratings year-to-year fell 6% in February, 7% in May, 8% in July and 7% again in November.
Share didn’t fare any better in the morning. It dropped a substantial 17% in July and saw double-digit declines in May and November as well — approximately 11% in both months. February, the remaining sweep month, also saw share dip, by 5%.
Here again, local stations seem to be experimenting with their timing. Some stations are starting local newscasts later in the morning (at the other end of the strong network morning shows). In Philadelphia, for example, the Fox affiliate WTFX added an 11 a.m. newscast in October 2006. The market also has two 10 a.m. newscasts, run by competing stations.9
The move to produce newscasts on more than one station can have a double edge. On one level, it represents stations trying to deal with the pressure on audience and revenue by creating more content. Yet that in turn tends to spread news staffs thinner. Producing more news does not always translate into producing better-quality news.
Another problem for local TV in 2006 was the growing complexity of actually counting the television audience. The year saw Nielsen Media Research, the standard-bearer of TV audience measurement in the U.S., embark on some its most ambitious initiatives. Announcing its plan to remake television ratings to keep pace with media consumption — for network, cable and local TV — Nielsen launched what it calls the “Anytime Anywhere Media Measurement” or “A2/M2” initiative in June 2006.
The announcement had three significant components. First, the company said it would begin tracking viewership of TV commercials. Second, it introduced plans to fuse TV and Internet viewership. And third, it announced plans to eliminate the traditional paper diaries in local TV markets within five years.
Data on viewership of commercials: Nielsen Media Research plans to release numbers in May, 2007 that will show how many people actually sit through commercials on TV. That new yardstick could very possibly alter the economics of the TV marketplace, affecting how much advertisers will pay to air ads.
The release of such data was delayed twice in 2006 (from November 18 to December 11, and then to May 2007) because not everyone in the TV business is happy about the change.
The major broadcast networks and advertising agencies have signed on, but most major cable networks are giving it a pass.
When Nielsen Media first introduced the idea of measuring ad viewership in early 2006, the broadcast networks were the first to welcome it — mainly to show their advertisers that they would support the idea of better data on who is watching commercials. Then, however, the picture got more complicated. In the 2006 upfronts (the weeks when TV stations and advertisers decide advertising rates) broadcasters did not fare as well as they expected.
Nielsen began offering three streams of ratings data in April 200610, but advertisers and broadcasters immediately disagreed on which one they would like to use. The ad agencies succeeded in setting ad rates based only the stream they preferred – the number of “live viewers” (which tends to have the lowest number of viewers). The agencies argued that they should not have to pay for viewers who watch commercials after they are broadcast. These “live” rates were lower than what the broadcasters expected, and now the broadcast networks do not want a similar parsing out of data when the commercial ratings are released.
In a meeting with clients in December 2006, Nielsen announced a compromise plan: It would make available all the data needed for customers to create their own commercial ratings for any minute of viewing and any interval of digital-video-recorder (DVR) playback. The hope was to satisfy both advertisers and programmers (the broadcast and cable networks) who could then tailor the ratings according to their preference. It would also give them the option to include DVR playback at any interval up to seven days.11
According to trade reports, broadcasters are eager to make commercial ratings the standard for the TV marketplace and were said to be pushing Nielsen to have the new measurement ready to roll before the May 2007 upfront (which is more critical for national programs and advertisers than for local advertising rates). In January 2007, however, Nielsen announced that the data would be released on May 31, i.e., not until after the upfront presentations.12
The new system was proving an even harder sell to the cable networks. Among those who refused to take part in the first ad-ratings test were ESPN, NBC Universal, Turner Broadcasting, Discovery, Fox Entertainment, Lifetime, A & E and MTV.
Their outright rejection was tied to the fact that the difference between the audience for a show and the audience for commercials tends to be larger on cable than on broadcast TV. Cable is susceptible to the drop-off because it tends to carry many more commercials and for longer periods than the average broadcaster.
The cable advocates argued that Nielsen Media’s methodology for calculating the ratings was flawed. They called for the new process to be audited and accredited by the Media Ratings Council (MRC), and for it to pass a “practicality and usability” test.
One problem the cable people have is that Nielsen Media’s commercial minute cannot distinguish between national spots and local spots. If the commercial minute measured contains a local or regional ad, viewership will be low, and that might skew the data. They worry that advertisers might not pay them the rates they deserve because of deflated viewership numbers when a local commercial is on-air. Also, the minute might overlap between two programs, and that could skew viewership.
The research company has been trying to address all those concerns. According to Nielsen Media press releases, it is working on weeding out local ads from national ads using special codes, and emphasizes that the data it releases this year should not be relied upon for negotiating ad rates. It has also agreed to an audit and is offering the first batch of “evaluation data” for free to participating clients.
Analysts believe that despite all the back and forth, measuring the viewership of commercials is here to stay, and that those new ratings could take over the program ratings as the currency for advertising rates.
New Technology Plans: Part two of the June announcement was Nielsen Media’s effort to use the same audience pool to collect data on both Internet and TV use. This effort, which involves two sister companies, is proceeding according to plan.
The two audience research firms that combined their tools are Nielsen Media Research and the Internet /Web traffic-measurement company Nielsen/NetRatings. They are autonomous, but each is part of the larger Dutch media conglomerate, The Nielsen Company (better known as VNU before January 2007).13
The data fusion plan aims to remake the ratings system so that both units can combine their expertise and capture TV watching on multiple media — including out of home, on the Internet and on mobile devices such as iPods and cell phones. By the second half of 2007, Nielsen Media plans not just to fuse the TV and Internet panel data, but to have meters on both TV’s and PC’s so that “broadcast, cable and online viewing habits can be tracked precisely.”
In December 2006, it introduced its “Video on Demand” (VOD) measurement system, which can compare the performance of a program when it airs on the traditional outlet (broadcast or cable) to its performance on-demand. That could prove to be a boost for video-on-demand advertising support.14
A month earlier, the company had announced the launch of a “National TV/Internet Fusion Database.” It means a single, integrated sample that can measure TV and online use (and their relationship) simultaneously. Though the data set is still cumbersome, with the two separate datasets measured separately and then merged into one, Nielsen Media is working on a single-sample Internet/TV panel.15
As part of the release of that database, Nielsen Media provided findings from the data fusion trials in April 2006. Though the findings were limited (and conducted in-house), Nielsen Media used them to indicate “the wealth of analyses that are possible with the database.”16 Nielsen also announced that its first attempt at out-of-home viewing would begin in January 2007. That is when it began measuring TV viewing by college students — following the younger members of its national sample to their college dorms.17
No More Paper: Another change to Nielsen Media Research’s measurement system is its plan to eliminate handwritten diaries and extend its people meters.
Handwritten diaries are still the only source of viewership in the smallest TV markets, and continue to be used in the larger markets as well to get viewer demographic information. Those paper diaries, in which viewers voluntarily write down every program they watch, are obviously fallible, and broadcasters have long been calling for a better system. Nielsen Media’s aim is to remove them altogether and replace them with some sort of electronic measurement. Whether the company’s five-year timeline for the switchover is realistic is another matter; the shift to electronic measurement requires effort and money, and that requires much negotiation with advertisers and television networks in the affected markets.
The largest markets, meanwhile, received the more advanced Local People Meters (LPMs) as promised. The meters are electronic boxes that measure not just what is being watched, but who is watching it. By July 2006, they were operational in the top 10 television markets in the U.S. — Boston, New York, Los Angeles, Chicago, San Francisco, Philadelphia, Washington D.C., Detroit, Dallas-Fort Worth and Atlanta.18
The mid-sized markets, which make up the largest share of the local-station universe (210 markets) continued to be measured by either the original electronic meters, the handwritten diaries just mentioned, or a mix of the two.
How Nielsen Media Measures Viewers
Source: Nielsen Media Research Web site; Broadcasting & Cable, April 17, 2006
As it plans to eliminate paper diaries from the smallest markets, Nielsen Media will extend the reach of the Local People Meter into the next eight biggest markets. In October 2006, it announced that it would roll out the meters in Houston, Tampa-St. Petersburg, and Seattle-Tacoma by October 2007. In 2008, it plans to introduce the device first in Phoenix, then in Minneapolis-St. Paul and Cleveland (by August) and then in Miami and Denver by October.
Those people meters were the subject of much controversy when they were launched in 2004. Some of the biggest TV groups, including News Corp. and Tribune, strongly questioned their accuracy and even accused them of having an ethnically skewed sample. By 2006, however, the opposition gave way to a general acceptance of the Local People Meters as the trading currency for the markets they’re deployed in.
One of the biggest reasons for that shift was Nielsen Media’s détente with News Corp. In October 2006, the two announced an agreement whereby News Corp. signed on to use the audience rating service for eight years. According to the agreement, Nielsen Media will invest $50 million to improve the response rate of young and minority viewers and in turn, provide measuring services to 49 News Corp. TV entities.19 That is a big change from a few years ago, when News Corp. was highly critical of Nielsen Media’s methods and even funded a pressure group, ‘Don’t Count Us Out’, to block the Local People Meters. The group’s activities eased by the end of 2005 and ceased altogether in 2006.
The cessation of hostilities can be tied to two main developments. One, the creation of the CW network, led to huge shifts in local TV affiliations (see last year’s ownership section). Fox lost many of its UPN affiliates, which were largely the source of the allegations of ethnic skewing. Second, and probably more critically, Fox programs have been performing quite well in the people-meter markets.
In developing its new measurements, Nielsen Media Research decided not to partner with Arbitron, the largest U.S. radio ratings company, to create a central local-market-ratings service for both radio and television. In March 2006, it announced that it would work on its “own strategy to provide more accurate and complete measurement of TV ratings.”
Back in 2000, Nielsen Media Research and Arbitron had signed an agreement giving Nielsen Media an option to form a joint venture for the commercial deployment of Arbitron’s Portable People Meter (PPM) — which can encode and log any type of audio — including its potential use for measuring TV ratings.
Arbitron first tested the portable meter in Houston in September 2005. The results were similar to the ratings trends seen in Nielsen Media’s Local People Meter markets with one major exception: ratings were higher with the Arbitron device, in part because it could track out-of-home viewing. Those developments were welcomed by the industry, which also saw them as a way to temper Nielsen Media’s monopoly on the TV ratings market.20
Not surprisingly, though, Nielsen Media eventually decided not to add weight to the rival system.21 As Nielsen Media prepares to measure the viewing habits of the new media consumer, it also seems to realize that it cannot diminish the value of its core product — its national sample of television households, which forms the basis of its new initiatives as well.