'Estimated Gross' or Gross Estimate?
By Michael Getler
September 28, 2007
This was a big week for PBS. On Monday night, the Public Broadcasting Service won 10 "Emmy" Awards in the News and Documentary category, more than any broadcast or cable television network. And the night before, the highly-touted and much-publicized series "The War" by famed documentary filmmaker Ken Burns made its debut on hundreds of PBS-affiliated stations across the country.
But did PBS try to make what, by any measure, was a big — and justifiably good — week appear even bigger? The officials here say no, but it looks to me as though they did. I'm not well-educated on the confusing art of measuring and rating television audiences. But as someone who has seen a fair share of press releases, I'd say that the one put out by PBS last Monday, less than 24 hours after the first episode of the 15-hour/seven-day World War II series aired, was at least questionable, and could be viewed as misleading, as some have labeled it. PBS disagrees and explains itself farther down in this column.
The headline on the Sept. 24 release said: 18.7 MILLION ESTIMATED GROSS AUDIENCE TUNES IN TO SUNDAY BROADCASTS OF "THE WAR" ON PBS.
The first paragraph read: "An estimated 15.5 million viewers tuned in to the 8:00-10:30 p.m. premiere of Ken Burns and Lynn Novick's World War II film, THE WAR, last night on PBS, according to John Boland, PBS Chief Content Officer. The Gross Audience for the first night, which includes repeat airing from 10:30 p.m.-1:00 a.m., drew an estimated 18.7 million."
That's a lot of eyeballs and it made news, including an early Associated Press report, for example, leading with the PBS estimate of 18.7 million viewers who saw this first chapter.
But also among the press clippings was the TV column by Los Angeles Times TV staff writer Scott Collins, an experienced television critic much more familiar with ratings than I am, and it carried a headline that said: "PBS wins but not by a landslide. The network bragged about 'boffo' ratings, but the results were misleading."
Here's most of what Collins wrote: "PBS, which normally shies away from ratings boasts, hustled out a news release earlier this week trumpeting boffo numbers for the Sunday premiere of Ken Burns' 'The War' . . . The network bragged of an 'estimated gross audience' of 18.7 million viewers for two airings of the two-hour debut. At first glance, that would seem to make 'The War' an early winner in the fall season ratings derby.
"But wait. It turns out that PBS was using a different yardstick than the one other broadcast networks commonly use. That high 'gross audience' figure counted any viewer who watched at least six minutes of the program, a technique that invariably plumps the final tally. And the estimate was extrapolated from data in the 56 largest TV markets, not the entire nation, according to PBS spokeswoman Stephanie Aaronson.
"The preliminary 'total viewer' figure — in other words, the average number of viewers who were watching at any given moment, and the statistic virtually all network officials cite when discussing ratings — was in fact 7.3 million, Aaronson said. That's roughly double what PBS typically does with original programming. But it's a far cry from some of the big headlines PBS grabbed with its initial release on 'The War.'
"During special events such as the Academy Awards, broadcast networks sometimes advertise a so-called reach figure, which tallies anyone who watches a program even for a few minutes. But it's a statistic that can be misleading. Univision drew complaints this month after claiming that its telecast of a Democratic forum drew 4.6 million viewers, making it the most-watched presidential debate so far. In fact, the average total-viewer figure was 2.1 million."
In other reports on the initial viewing that I've seen, The Washington Post, the Arizona Republic and Reuters also used the 7.3 million viewer figure, while Daily Variety used the 18.7 million estimate but explained that "PBS is emphasizing the project's cumulative reach as opposed to average audience."
Another later and fuller Associated Press story, which may well have been seen by the widest portion of the public, said that the first of the seven-part series "attracted 15.5 million viewers to PBS on Sunday, according to Nielsen Media Research. Add in a re-run that was telecast immediately after it was done, and an estimated 18.7 million people saw it on Sunday." A spokesman for Nielsen said that "the proper way to attribute that would have been to say, 'according to PBS, which did an analysis of Nielsen Media Research ratings of that evening.'" PBS, in its internal newsletter for employees, also used the 18.7 million figure and attributed it to Nielsen Media Research. But a Nielsen official said, "We did not put out that number."
A Nielsen official said that the initial PBS press release properly attributed the number to PBS's top content officer, John Boland. "So they are probably being a little cute. It's not an egregious wrong, and probably a real number. But it is misleading to a casual reader. There could have been more explanation about how these numbers are derived at PBS. 'American Idol' gets 18 million and they (PBS) didn't get 'American Idol'" numbers, which he explained would be way above "The War" numbers using the same approach and analysis that PBS used. "That 7.3 million (for 'The War') is the average number of viewers watching at a particular time, and that's the most common yardstick," he said. "Everything in that Los Angeles Times article was correct," he added.
In a telephone interview, the LAT's Collins explained that the average number of viewers who are watching is a measurement taken every 15 minutes during prime time, "so when you are looking at 8 p.m., you are looking at the average audience over that hour because you want to eliminate statistical aberrations" such as someone who might have watched for only six minutes because they were on the same channel earlier, and then changed. "That's the average which is the industry standard. And 7.3 million was a good number (for PBS). There was no reason to hype it." There is, he adds, "a fairness and historical accuracy issue here" as well, "especially since you are talking about a historical documentary trying to get at the truth. The news release didn't seem that way to me. You should not have to call PBS" after the press release is issued, he said, to get that 7.3 million number.
In a written response to questions, Stephanie Aaronson, senior director of Media Relations at PBS, and Beth Walsh, the director of PBS Research, said:
"We know there are two parts to your question about the rating — why publish a cume (cumulative) and how we reached the audience estimates cited in the release.
"'Why' is really straightforward. We knew that media wanted to know if PBS believed the first night of THE WAR was a success based on Nielsen numbers (the first available set of data among everything PBS will evaluate). We did and the release explained why. The release accurately offered several data points — in addition to our 'cume' which is how PBS confirms that we are reaching the broadest possible audience (i.e. our measurement for success is an evaluation of the total audience), we also included the average rating. In Nielsen lingo this 'cume' is a measure of everyone who saw six minutes or more. Scott Collins is aware of this tally as he specifically called to ask if our number was factored based on 'anyone who watched for at least six minutes' which we confirmed. There are trade reporters who regularly cover ratings and prefer to use the average rating, which was a 5.0 HH (households) or 7.3 million people, which is a measure that reports how many people were watching at any given minute. We were happy to provide either and all data points.
"'How' we arrived at the cume estimates is through a set of formulas based on historical Nielsen data showing the typical ratios between our metered-market average rating and the national persons cume that will eventually be reported by Nielsen. Nielsen has approved this formula. Since we have to wait so long before the final national numbers arrive, we have time-tested formulas developed by this department that can be used to project the total audience with a high degree of certainty."
And, Furthermore . . .
John Boland, the chief content officer, adds this:
"First, all of the raw numbers used in all these various analyses of audience figures generally come from the same source — Nielsen. As far as the LAT writer's statement that 'virtually all network officials cite (average viewers) when discussing ratings,' this might be technically correct in that the commercial networks slice and dice the audience numbers in every imaginable way to serve a variety of purposes, usually related to advertising sales. They utilize overnight ratings, national ratings, average viewers, total viewers, cume viewers, and a wide range of age and demographic breakouts. So while 'average viewers' may be among the statistics they cite, it is not, to our knowledge, the generally accepted 'official' audience figure for any given program.
"It is a fact that PBS looks differently at the Nielsen numbers than our commercial broadcasting and ad-supported cable colleagues. And this is because we have different goals and serve different masters. The commercial and cable networks utilize the Nielsen's to support their ad sales efforts and much of the reporting Nielsen provides is tailored to serve those purposes. Commercial media and ad agencies are, after all, Nielsen's primary clients. The goal of commercial media is to effectively deliver the right audience to the right ad to satisfy their primary audience — advertisers.
"In that context, it makes sense that commercial media advertisers would be interested in knowing the average number of viewers at any given minute since that would indicate the average number of viewers watching their ads. The total number of viewers for a program would be of less interest because there is a reasonable expectation that a significant portion of the total viewers may not have been watching when the ad aired. And, the ad rates are driven by the audience size.
"PBS starts from an entirely different place. Our goal is not to serve advertisers, but rather to serve the American public. We receive a similar range of numbers from Nielsen but we are more interested in serving the broadest possible cross section of the American public than in how many viewers watched a particular ad or the specific target demographic of the viewership. All of this data is interesting to us, but we don't exist to deliver particular eyeballs to particulate advertisers, so we tend to focus on different numbers.
"We did not cite the total viewership of the first night's broadcast of 'The War' because we wanted to inflate the size of the audience. We did so because we are most interested in how many citizens we served, not how many were watching at a given minute. We define success differently than a commercial medium.
"As I noted in our press release on the subject of 'The War' ratings, the first night's audience is really just an early indicator for us — and in this case a positive one — but the real number that we care about will not be available to mid November — the total number of unduplicated viewers who watched 'The War' over the original broadcast period."
A Parting Thought
Well, that helps a lot, actually. Nevertheless, PBS, in my view, should have been more straightforward about its first ratings statement dealing with this very high-profile, high-stakes series. It should have specifically included the 7.3 million figure, which is derived from a standard practice approximating the numbers regularly used and reported by networks and media writers. And it should have made clear that it uses its own formulas to extrapolate to the larger figures, and what audience measurement (six minutes or more) those extrapolations are based upon. It would not have taken much to do that, in plain English, and avoid the sense of hype and disappointment that got attached, by some, to an event that didn't need to be hyped.