Trying to Clean Up Sweeps

By Published on .

This past January, NBC began airing commercials to promote a guest appearance by Ellen DeGeneres on its popular sitcom Will & Grace. In a classic bit of stunt casting, the network hired the comedienne and out-of-the-closet lesbian to portray a nun, showing her cavorting wildly in the promo while dressed modestly in a clerical habit. Outrageous? Absolutely. Senseless? On the contrary, the move made perfect sense. Producers timed her cameo to air during the heart of February sweeps.

weeps, those special, month-long ratings periods in November, February, and May, represent the silly seasons of television programming, when no audience-luring gimmick, watch-and-win sweepstakes, or blockbuster miniseries is too salacious, provocative, or bizarre for public consumption. They're the months when Nielsen Media Research measures the viewership of local stations in all 210 U.S. television markets, to determine what rates advertisers must pay for spot and local TV commercials. With an estimated $24 billion in ad revenue at stake, network executives and their affiliate stations unleash all manner of programming mayhem in an effort to increase the size of the coveted 18- to 49-year-old demographic in individual markets. “To television producers, there's Christmas, there's Yom Kippur, and there's sweeps,� says Jay Kernis, a veteran producer at CBS's 60 Minutes. “You prepare for it like a religious experience. You either lose or save your soul depending on how successful you are during sweeps.�

But the hoopla surrounding sweeps masks a dark side to audience measurement. Many researchers view the current method of counting local viewers as flawed, outdated, and inaccurate. Nielsen, which has dominated TV ratings at the national and local level for more than three decades, measures the audience size of network shows via its so-called people meters, electronic set-top boxes that are placed in 5,000 homes around the country. To monitor the demographics of local stations, however, Nielsen goes low-tech, distributing paper diaries to 2 million volunteers over the course of a year, with requests that they record the programs watched during one week of a sweeps month. (Additionally, the company installs 400 to 500 set-top meters in the top 53 markets to help determine which households watch what programs, though not which people watch what shows inside those homes.)

The Demographics of Network Television Desperately seeking 18- to 49-year-olds.

ABC 47 $42,000 43% 57% 7% 5% 5% 12% 27% 45%
CBS 51 $36,000 43% 57% 5% 3% 4% 11% 25% 52%
NBC 46 $46,000 42% 58% 5% 4% 6% 15% 28% 43%
FOX 38 $44,000 59% 41% 11% 8% 8% 18% 27% 28%
WB 23 $39,000 45% 55% 25% 16% 11% 15% 20% 13%
UPN 29 $33,000 54% 46% 20% 14% 9% 15% 25% 18%
A&E 55 $52,000 38% 62% 3% 2% 3% 9% 23% 61%
ESPN 39 $54,000 75% 25% 5% 7% 10% 22% 25% 32%
HBO 38 $50,000 56% 44% 7% 8% 10% 19% 33% 23%
MTV 20 $54,000 50% 50% 10% 31% 22% 18% 13% 5%
Sources: Table provided by Universal McCann, network data from Nielsen, cable information supplied by Mediamark, 2000
Sweeps isn't just a numbers game. It's also a demographics game. Advertisers traditionally pay local stations for commercial time based on their ability to deliver an audience of 18- to 49-year-olds. But as the oldest Boomers reach their 50s, TV networks like CBS are encouraging advertisers to change the medium's favored target to 25- to 54- and even 25- to 64-year-olds. “The Baby Boomers have managed to cross over the 54-year-old threshold, and there's no evidence that their spending powers are diminishing,� says CBS research chief David Poltrack.

But sweeps and the diary method remain a demographics problem. Because local and spot advertisers need the demographic data to make sure stations are delivering their target audience, the antiquated diaries remain the measurement tool of choice in all 210 markets. And the extraordinary programming and promotion efforts during sweeps throws off the validity of audience measurement for the rest of the year. Networks and their affiliates design their stunts to pull in extra viewers for local stations during sweeps. Yet those exaggerated figures are the ones that stations use to charge advertisers for commercial time year-round — long after the fickle public has reverted to its old habits. One CBS study found that the overall audience for the four major networks drops 10 percent to 15 percent between periods of original programming and the repeats that dominate non-sweeps months. “Sweeps programming hurts us and hurts the public,� says David Poltrack, executive vice president for research and planning at CBS. “Television viewers dislike having to decide which of two major movies to watch in February while finding only repeats on in March.�

What's bad for viewers is even worse for advertisers. Media buyers have no way of knowing how many viewers stick around during the nine dreary months filled mostly with reruns. And the issue is about to get more complicated as the arrival of digital technology creates alternative ways to watch TV, and more accurate methods of counting viewers. “Sweeps are a travesty,� says Susan Nathan, senior vice president and director of media knowledge at Universal McCann, and chair of the media research committee for the American Association of Advertising Agencies (AAAA). “Advertisers buy time on stations 365 days a year, yet we have no idea what ratings are for most of the year when there aren't those hyped, big-event programs.� Nathan speaks for many in the advertising community when she concludes, “Sweeps should be done away with.�

Dispatching sweeps to the dustbin of TV history is hardly a new idea. Advertisers initiated broadcast ratings in 1930 in order to count radio listeners for the shows they produced and sponsored. The term sweeps dates back to the 1950s, when Nielsen began mailing diaries to households and reporting the results, first with East Coast markets before sweeping across the country. As TV networks gradually took control of programming, they needed better demographic numbers to calculate how much they could charge for commercials on specific shows. Local stations couldn't afford to track overnight ratings year-round, as the networks do, so they settled on the four monthlong sweeps (another occurs in July, but is less important due to relatively few summer viewers).

Gradually, Nielsen built a monopoly on the TV ratings business, knocking out its last major competitor, Arbitron, in 1993. More recently, various companies and consortia have tried to dethrone Nielsen's reign over the TV ratings business, but to no avail. In 1999, Statistical Research Inc. ended the latest major threat — the $40 million SMART TV initiative — after the broadcast networks dropped financial backing for the project. The initiative would have produced a digital TV coding system and the ability to give clients respondent-level data, features that Nielsen claimed would soon be incorporated into its systems. Indeed, the biggest complaint was that SMART was a “me-too� service that failed to deliver greater benefits than those offered by Nielsen.

But such failures haven't muted the cry for sweeps reform. Trade groups such as the AAAA have proposed increasing the number of weeks Nielsen measures, thereby reducing the likelihood that local stations will rely on programming stunts to artificially boost ratings. But Karen Kratz, Nielsen's director of communications, insists that the idea is too costly for small-market stations. “In a perfect world, putting people meters in all the markets is what we'd want to do,� she says. “But there's not enough advertising dollars in smaller markets to make it worth their while.�

And there's little incentive for local broadcast stations to push for changing the archaic system. At most stations, year-round monitoring will only result in lower ratings and ad revenues, says Universal McCann's Susan Nathan. Still, she regards that resistance as a short-sighted view for a long-term problem. “It's a Catch-22 situation,� she observes. “Nielsen should have been leading the charge to a better system but hasn't in the past. And the stations have been too afraid of year-round numbers hurting their bottom line. The blame goes all around.�

Must-See TV Lifestyles of





Bought jeans in last year 66.8% 71.1% 76.8%
Drink liquor 45.4% 50.0% 56.0%
Foreign travel in last three years 41.0% 41.0% 42.4%
Agree: “I enjoy taking risks� 38.4% 41.5% 45.2%
Agree: “On the whole, people get what they deserve� 38.3% 30.0% 34.2%
Drink diet cola 34.0% 36.0% 37.3%
Own foreign car 32.9% 33.7% 38.9%
Camping in last year 18.7% 22.5% 23.8%
Bought women's swimsuit in last year 13.1% 15.2% 22.5%
Drink espresso/cappuccino 11.6& 13.4% 13.2%
Source: Simmons Market Research Bureau-Fall 2000

Measurement Alternatives

Few in the industry deny that some kind of overhaul is needed. Even Nielsen executives say that their technology misses viewing that occurs outside the home — in bars and offices. The impending arrival of convergence technology means that Americans will soon routinely watch TV shows on their computers, Palm Pilots, and wireless devices. “You've got to be able to measure someone watching a TV show on their Palm Pilot on the train going to work,� says Howard Shimmel, president of Symmetrical Resources, a marketing research firm based in Deerfield Beach, Florida, and parent company of Simmons Market Research Bureau. “If we don't solve these issues sooner than later, it's going to get ugly.�

Toward that end, researchers are scrambling to develop new and improved ways of measuring TV audiences. Symmetrical recently launched TV Programming Clusters, a system that segments viewers based on Nielsen data and Simmons surveys. Media buyers can choose from among 23 distinct clusters: One segment, for instance, consists primarily of fans of cop/lawyer dramas, while another is filled with devotees of the Judge Judy- and Maury-type courtroom and talk shows. Advertisers then hone their media mix based on shows favored by their target market, using the TV clusters to supplement rather than to supplant Nielsen ratings. “The clusters allow advertising agencies to define the viewers that really matter to them,� explains Shimmel. “It shows subtleties where they may have been paying for high-priced inventory they didn't need or ignored some undervalued inventory that would work well for the viewers they want.�

To improve the measurement of smaller cable TV audiences, traditionally undercounted by Nielsen, a Florida-based company called ADcom has developed a set-top meter system named VVR, or viewership valuation record. After connecting viewership meters to every TV set in the homes of panelists, the company takes channel readings every five seconds to passively record which programs are being watched. ADcom then uses follow-up phone interviews to determine who was in a room at a particular time, allowing a company to produce year-round cable ratings by day part, target audience, and network ranking. Although the 5-year-old company has launched set-meter panels in only three markets so far — Jacksonville, Dallas, and San Francisco — it has recently begun an aggressive expansion. Backed by ad-agency Goliath, WPP, it hopes to offer cable measurement in the nation's 20 largest markets within two years.

An even bigger challenge to Nielsen recently arrived as a joint venture of shopping network QVC and its majority owner, Comcast, the cable system operator. Called TargetTV, the service relies on digital set-top boxes to monitor viewer behavior, recording “clickstream� data every five seconds in order to capture channel surfing. The service, launched last year, now operates in 60,000 homes in Philadelphia, with the goal of expanding to 750,000 homes nationwide. Though there are drawbacks — TargetTV fails to capture non-cable users and those with analog cable boxes — it found a receptive audience at this spring's Advertising Research Foundation (ARF) convention, and the deep pockets of its backers certainly improve its prospects for survival.

Another alternative to paper diaries involves a computer model that helps predict viewing habits based on the set-top meters Nielsen uses in the 50 largest markets. Funded by a $60,000 ARF grant, media consultant Erwin Ephron is now testing a methodology that gleans the demographic characteristics of viewers using household tuning data. As he explains, “You know that if the set in the kitchen is tuned to Oprah, it's probably a woman who's watching. If it's football in the den, it's probably a man. If the Cartoon Channel is on, it's probably a kid.� Ephron believes that such viewer modeling may be superior to both people meters and diaries in smaller markets because of the “noise� generated by low response rates and poor record keeping. But the project is still in the early stages and a year away from a national launch.

Perhaps the most promising new technology in TV ratings is Arbitron's “portable people meter� (PPM), which measures the person, not the appliance or the household. The pager-size device detects an inaudible code embedded in the audio signal of a TV or cable program, radio show, or Internet streaming video. At the end of the day, a person carrying around the device simply drops it into a bay station, which then sends the ratings information to Arbitron. Linda Dupree, senior vice president in advertiser agency services for Arbitron, maintains that the PPM is superior to the traditional set-top meter because it captures audience behavior year-round, it tracks media consumption outside the home, and it tracks all different kinds of media. “It blunts the whole sweeps promotion issue,� she says. “And it allows advertisers to make apples-to-apples comparisons across all media.�

But the device may not be ready for prime time for some time. After being tested for three years in Britain, Arbitron began PPM trials in Wilmington, Delaware, this spring, with plans to form a panel in the entire Philadelphia market in 2002. The goal is then to expand “aggressively� into the nation's top 10 markets, says Dupree, but there's no announced timetable for rolling out the service. The company has also skirted potential opposition from Nielsen by giving its one-time rival the option to develop PPM technology should it go into commercial production.

That type of co-opting partnership has become typical for Nielsen. In preparation for the media landscape of the future, the company has already formed an alliance with Intel for a “software meter� that can measure TV viewing on Windows 98. Other deals are tied to the development of interactive television and ITV viewership measurement. Earlier this year, Nielsen launched a demonstration project placing people meters in the Boston market, capturing year-round viewer demographics as well as household TV usage at the local station level. Eventually, the company would like to see the people meters in the nation's 10 largest metro markets.

But there's still no long-term plan to bring people meters to all 210 markets. And even with a successful test, any replacement for the Nielsen sweeps system would have to overcome stiff resistance. Local stations would likely be reluctant to adopt any system resulting in lower audience numbers and increased costs for year-round ratings. And ad agencies wouldn't be pleased if a new measurement system indicated that prior programming selections and fees were inappropriate. “There's a lot of money at stake,� says Symmetrical's Shimmel, who estimates that a completely new TV ratings service may require a couple of years and $100 million in start-up costs. “Switching from one measurement service to another is a monumental change.�

How Soon the Future?

Another wild card in the potential conversion to a new measurement system is the pace at which Americans may adopt “smarter� TV technology that allows personal recording and interactive video programming. The new generation of digital set-top boxes not only delivers cable and satellite programming but also has the capacity to track whatever shows a turned-on set is receiving. The downside is that personal video recording systems like TiVo are always turned on (and able to record programs throughout the week), but researchers currently don't know which shows are being recorded because TiVo strips all the personal data over who's watching what shows to avoid raising privacy concerns.

Hoping to answer these questions, ASI Entertainment, Nielsen, and TiVo recently created the National In-House TV Lab to explore how viewers use digital video recorders. Do they record one or five shows a day? Do they fast-forward through some or all of the commercials? Some of the answers will be known this year as ASI completes the formation of several regional panels to test TiVo usage. But the possibility of most viewers whizzing through commercials may have long-range implications. Dave Charmatz, ASI's executive vice president, believes that advertisers may soon create messages that will appear on your screen even as you fast-forward through a commercial. “It couldn't be subliminal because that's illegal,� he says. “But advertising is like Jurassic Park. Just as life finds a way to exist, advertising finds a way to get across its message.�

Jim Spaeth, president of the ARF, counters that TV advertising may eventually go retro, returning to the early days of the medium when advertisers relied on corporate sponsorships and ad placements to get their message across. Still others say that TV advertising will evolve into more futuristic forms, boosted by interactive technology that allows viewers to click on their screens and move from TV image to a related Web site through a process called Channel HyperLinking. According to the creator of the technology, Philadelphia-based WorldGate Communications, it allows viewers to click on, say, TV ads for General Motors and Kraft and then jump to company Web sites. At the same time, Wink Communications, based in Alameda, California, allows nearly 2.5 million cable viewers to reply to TV offers, buy products, and participate in games, using their remote controls during an enhanced program or commercial. A report last year stated that of the 2 percent who click on the enhancements, a whopping 42 percent follow through on the offers presented whether they accepted free brochures or purchased products. “Direct mail can't touch a response rate like that,� says Josh Bernoff, a principal analyst with Forrester Research, based in Cambridge, Massachusetts. “Soon, smarter TV will coexist with plain old TV.�

The success of such convergence technology may call into question traditional ways advertisers count demographically preferred viewers, which, after all, is what sweeps is all about. Given the many new platforms for delivering TV programs and commercials, advertisers may soon have to assign different values to viewers in different settings. The travelers who catch a few minutes of a show on a TV in an airport lounge may be worth less than the viewers of interactive systems like WebTV or RespondTV who click on their screens to respond to ads. Bernoff maintains that TV executives will soon follow the online pricing model of charging variable rates pegged to the viewer's response: One fee based on passive viewing, a higher one for those who click through, and the highest one charged for those moved to buy.

“An ad that stimulates a lot of response will be worth more than one that just sits there,� Bernoff says. “Interactive TV allows you to monetize a viewer's attention. You'll soon see a shift toward the value of TV based on viewer interaction, not just advertising impressions.�

But such lofty predictions have been heard before, and the timetable for such change is still in doubt. Last year, less than 1 percent of U.S. households had interactive TV. Forrester predicts explosive growth, stating that over 60 million households will be able to interact with their TV programming by 2005. But will viewers actually use TV in an interactive way? “That's the 64 million dollar question,� says ARF's Spaeth, “whether the couch potato will turn into a different vegetable. Will people actually use TV's interactive capability, or will it just become some more buttons on their remote control that they won't touch? No one knows.�

Given such uncertainty, trying to predict when sweeps may one day end is admittedly a fool's game. Several experts insisted that the archaic system will finally die in the next decade. Others weren't so sure. “I believe that sweeps will one day be gone, but probably when I'm retired on a beach in Hawaii,� says CBS research chief David Poltrack, 55. “And that's the optimistic forecast. The pessimistic forecast is that I will not be of this world.�

Of course, viewers may one day simply get tired of the same old manipulative tactics during sweeps and, like the characters in the movie Network, open their windows and scream that they won't take it any more. Market researchers would have fun trying to calculate the odds of that occurring. It's probably as likely as a lesbian comic undergoing a personality-change operation and turning into a nun.

Contributing Editor Michael J. Weiss is a journalist and marketing consultant. He recently completed a Knight-Bagehot Fellowship in Economics and Business Journalism at Columbia University.

Most Popular