Is anybody out there?
I read the comments. What choice do I have? I work in podcasting, a media realm that for all its “metrics” and “listen time” data, is still mostly akin to leaving a message under a rock and hoping that someone finds it. Comments on stuff I’ve edited are at least proof that someone found what we made, and maybe even listened to and enjoyed it.
Read enough listener comments, and you start to see the patterns: Most are complaints about the ads. Unfortunately, we produce a free product that got stuck as free 100 years ago, a problem on par with climate change in terms of realistic solutions. Other than the ads, many complaints are about the presence of women’s voices. Thus far, podcasting is “solving” for that in the worst possible way.
But you can find useful feedback. I once worked in a place with a podcast whose host liked to show his enthusiasm by interrupting his guests’ answers during interviews. Listeners piled on, and he figured out how to hold his tongue.
Still, there’s a central mystery at the heart of audience feedback: the audience. What is it, really? The people who bother to make comments are only a small part of it. We producers spend a lot of time making our work, and we want to know how it is received by everyone. How are these conversations and stories reconstructed inside the minds of a wider cross-section of the people who encounter them?
Stanton-Lazarsfeld Program Analyzer brochure, Library of Congress.
Enter the analysts
In the beginning, broadcasters were shocked when the phones in the studios would ring after a program, or the volume of letters and telegrams the station would get from listeners. Then, as radio became more and more popular, broadcasters started to elicit those phone calls and letters. They used them to get advertising support, which was the only way US stations could keep operating.
But soon those advertisers wanted better data about who was listening, how many, and why. The anecdata from phone calls and letters didn’t cut it any more, so radio had to invest in actual research. Media historian Susan J. Douglas has a brilliant chapter in her book Listening In: Radio and the American Imagination about how radio had to figure out its audience.
As Douglas explains, audience research at first was crude and biased: starting in 1930, a polling firm would call a random sampling of listeners in a geographic area and ask what programs they’d heard in the past two hours, and if they could remember the sponsors. But people misremembered what they’d heard, and when. Not to mention, many Americans still did not have phones, so the overall sample data was skewed.
These were the kind of worries that faced a person with a new kind of job: head of audience research for CBS. Frank Stanton, who came to the network in the early 1930s with a PhD in psychology from Ohio State, soon met a brilliant Viennese social scientist named Paul Lazarsfeld, who was in the United States on a grant from the Rockefeller Foundation. With some more Rockefeller money, the two joined forces with Princeton psychologist Hadley Cantril and eventually started the Office of Radio Research (ORR).
If the challenge was to get at what radio was doing to people’s heads — as individuals and as members and shapers of a society — how would they proceed? It wasn’t just that researchers were unsure about radio’s impact on thoughts and actions; they didn’t know which techniques and answers might come close to producing answers.
Frank Stanton was the kind of guy who liked to build things, so after a chat with his ORR colleagues, he took on the task of constructing a contraption that could measure listener reactions in real time. They called it the Program Analyzer. But people at CBS soon called it Little Annie.
Little Annie was based on the same basic technology as a polygraph machine. (By the way, historian Jill Lepore has a fascinating book and podcast episode about one inventor of the polygraph, William Moulton Marston, who was interested in more kinds of poly- than just graph, if you catch my drift).
Anyway, here’s Frank Stanton in an oral history for Columbia University describing how his program analysis machine worked:
The buttons were wired to the machine so that you sat and held a red button in one hand and a green button in the other. If you liked something you pressed the green button, if you disliked it you pressed the red button, if you were indifferent you didn't press either one. And then on that piece of paper there, each one of those needles connected to our particular — or to the subject's response buttons, would record on a line either plus or minus or nothing or indifferent. And that tape moved at a steady rate so that you could calculate time by putting down a measuring device as you stretch — When the program was over, you would stretch out the tape and put down a ruler or a measuring device, and say, at one minute and thirteen seconds into the program, you switched from liking it to disliking it.
Yes, this was perhaps the first “like” button, one provided in a dim room along with cigarettes.
Stanton’s employer CBS soon started doing controlled audience research all the time, creating a bigger version (Big Annie) to poll more listeners at once. They’d put out the call for volunteers over the air to render their judgments on programs in development, filling up regular “program clinics” in the network’s Manhattan headquarters.
For The Reporter magazine in 1950, Thomas Whiteside sat in on one of these sessions. Each participant seat was assigned a number for easier identification of who produced each like and dislike.
The lights were lowered, a record of the show was started on a phonograph, and the executive retired to the next room to observe the reactions of Little Annie. As the show progressed, needles on the machine began to click, furiously registering approval here, disapproval there. "Number Four doesn't like the narrator," said the executive, scribbling frenziedly. "Number Eight's wavering." The needles clicked on. "Number Seven likes the heroine," said the executive. "But watch Number Four! See that needle? Number Four just can't stand the narrator."
A crucial next step after the listening session was the follow-up interview with each participant to ascertain the reasoning behind each reaction. Whiteside then found out what was going on with Number Four.
…the CBS psychologist asked, “How did you like the narrator, Number Four?” Number Four turned out to be a middle-aged lady with a high-pitched voice and a large hat. “Oh,” said Number Four, “I liked him fine.” She had evidently been pressing the wrong button. I don't know what happened to the narrator, but the damning evidence lay remorselessly entrenched somewhere in Little Annie's intricate intestines. I subsequently calculated that if CBS's statistical universe consisted only of the ninety-nine million people it claims listen to its network every week, Number Four had inadvertently brought down upon the narrator's head the curse of 1,320,000 of them.
The invented audience
Whiteside’s bemused account shows the paradox inside audience research: the whole concept of audience measurement quickly starts to evaporate when you get into the particulars. “The object of this scrutiny — the audience — was itself an invention, a construction that corralled a nation of individual listeners into a sometimes monolithic group that somehow knew what ‘it’ wanted from broadcasting,” Susan Douglas writes.
Still, audience research did create a lot of insights into how mass media was processed inside the brains of Americans. Another Viennese social scientist, Herta Herzog (who was Lazarsfeld’s wife for a time) was a master of this aspect of research for ORR. Her reports, like On Borrowed Time, about the “gratification” of soap opera listeners, are brilliant reads. She understood how much listening could be an emotional act. Just think about true-crime podcasts as you read some of her analysis:
The interest in other people’s misfortunes was also brought out in the answers to the question whether and about which incidents the respondents had ever been very much excited. Forty-one per cent of those who answered in the affirmative referred to murders, violent accidents, gangsters, and fires; 15 per cent more mentioned illness and dying; 26 per cent spoke of psychological conflicts, while only 18 per cent named incidents of a non-violent or non-catastrophic kind. The aggressive meaning of these answers was exemplified rather strikingly in the following comment of a listener who explained why she never had been really excited. Referring to [the soap] Woman in White, she said:
“I thought the murder would be exciting. But it was not. It happened abroad somewhere.”
It’s no surprise that in the 1940s, the research prowess fostered by radio followed the money into television, and specifically, to ad agencies. McCann-Erickson not only bought the Stanton-Lazersfeld Program Analyzer, they hired Herta Herzog as well. Fun fact: she’s widely thought to be the basis of the character Dr. Greta Guttman, a psychological researcher who talks about the “death wish” in the first episode of Mad Men. But better to watch the real thing as Herzog talks about her work in this 2010 interview with the BBC’s Adam Curtis. In the film, she’s back in Europe, 99 years old, and sharp as a tack.
So Herzog was one major inventor and refiner of the focus group interview, the backbone of market research (except for Doritos). Her innovation itself centered around listening: asking questions in both a methodical way for later analysis, but also building trust and responding to the cues of conversation.
Still, as a producer, I find some of the insights gleaned from those Little Annie “like/dislike” sessions fascinating. Most podcasts can’t really afford a full-on market research of the type radio got back in the day. And that research did reveal some sonic techniques favored by producers that listeners, when probed, admitted they did not like. CBS News Editor Paul White wrote about one of those production habits in 1947: the tendency to rely on a chorus of anonymous voices to create a sense of narrative “excitement.”
Actually there's some evidence from a Stanton-Lazarsfeld study that this effect, which is called a “montage,” is resented by the public. The different voices are unexpected and generally baffling.
No, Little Annie, do not take my montage sequences away! I am sure there are more truth-bombs of this nature in the old CBS Research archives, if they even exist. But besides being antique data now, in reality, there’s only so much that granular research can do for a show.
The radio drama Suspense was probably the network’s most audience tested. Listeners made it clear they strongly preferred a snappy opening, an obvious protagonist, and little digression from the main plot line — and those changes did increase the show’s popularity. Still, as Frank Krutnik points out, all the audience research in the world could not replace the work of the creative team behind Suspense.
Without the contributions of writers, performers, musicians, sound technicians, directors and producers, the Research Department would have had nothing to work with in the first place, and the program would have failed to attract loyal listeners. The research methods of the CBS program clinic also had significant limitations. Listening to the recording of a radio program in laboratory conditions was a very different experience from listening to it, live, at home.
Audience feedback can make a good thing better, in other words, but it cannot make a bad thing good.
Nor does it really challenge broadcasters to be better. This was the key insight — an unwelcome one — from the third major Viennese intellectual who briefly joined the Office for Radio Research: the musicologist, social critic and philosopher Theodor Adorno. Adorno was a refugee from the Nazi regime who was supposed to bring some critical rigor to audience research, especially when it came to the effects of music on listeners. Instead, as a dedicated Marxist, he mainly railed against the implicit (and explicit) goal of audience research at places like CBS — namely, to sell stuff. More interestingly, Adorno nailed the implicit one-way deal broadcasters had with listeners. The listener got free programming larded with commercials. They got to respond to that programming in only one of three ways: they could write or call in, change the channel, or turn off the radio. Adorno thought commercial broadcasting had made its audience so passive and bamboozled that it was pointless to ask them what they wanted.
The consumer is unwilling to recognize that he is totally dependent, and he likes to preserve the illusion of private initiative and free choice. Thus standardization in radio produces its veil of pseudo-individualism. It is this veil which enforces upon us scepticism with regard to any first-hand information from listeners. We must try to understand them better than they understand themselves. This brings us easily into conflict with common sense notions, such as “giving the people what they want.”
Adorno did not come to play. Soon he parted ways with the Office of Radio Research and decamped to sunny Los Angeles, leaving his fellow Viennese refugees a little mortified, though they swiftly recovered. But Adorno’s brief invitation to think bigger “was significant for communication research precisely because he was rejected,” writes media studies professor Josh Shepperd in his history of educational radio, Shadow of the New Deal (p. 148). Going forward, he notes, academic audience research would be tied not to theory but “to grant writing, advertising, metrics, and public relations discourses.”
Meanwhile, as the audience, we’re still trapped in the veil Adorno described. We now exercise our right to “like” stuff all day long. Those reactions are instantly turned into data so our personal algorithms can be used to sell us yet more stuff and predict our future purchases. We are heard and seen across comment boards, via avatars and arguments — and in large parts of Asia, danmaku “bullet” subtitling, probably something that will take over the world’s screens soon enough.
We are followed and counted all day long while left to feel, in our hearts, that we count for very little at all.