Asking People What They Listened To Gets the Answer Wrong
Memories are faulty, so ratings completely changed once radio started tracking real listening.
PodBiz | Where's The Money In Podcasting? · How Privacy-First Data Is Changing Podcast Ads with Emma Raz of NumberEightAsk someone what they listened to last week and you will probably get the wrong answer. Usually because remembering is hard.
That is the argument from Emma Raz, of NumberEight, a British company that supplies audience data for podcast advertising. She points to music and radio. For a long time, she says, their ratings systems relied on surveys. People were asked what they had listened to, and the answers set the rankings. Once the industry "moved to actually monitoring what people really listened to," in the words of Raz, the ratings completely changed. "People are not able to correctly identify and report back their action," she says. Raz says it is a human limitation.
In podcasting, Raz says the two methods currently available are matching by internet address and surveys. Surveys have two problems. They are expensive to conduct, so scaling them is costly, and they depend on memory. The typical question, what did you listen to in the past week, runs straight into the fact that listeners forget. The other method is matching by internet address, which at best identifies a household rather than an individual. That matters because demographic data, like age and gender, belongs to a person, not a living room.
The alternative is to stop asking and start observing. Instead of asking what someone consumed in the last week, NumberEight looks at what was actually listened to, streamed, or read in that week, and matches advertising to it. Across millions of users, those observations form what Raz calls podcast neighborhoods, or content consumption neighborhoods, groups of people clustered by real behavior.
The result, she says, is bigger than a survey and more reliable, because the underlying record comes from watching what people actually consume. Audiences are built from observed consumption.
Questions people ask
How did radio actually monitor what people listened to?
In the 1990s, U.S. radio moved in many large markets from listener diaries to Portable People Meters, small devices that passively detect inaudible codes in broadcasts. The change shifted credit from stations people remembered fondly to stations they actually had on, often boosting music and background listening.
Why is an IP address only a household match?
An IP address is assigned to your internet connection, which in a home is usually shared by phones, laptops, smart TVs, and everyone who lives there. Advertisers can tell roughly where the household is, but they cannot reliably tell which person in the house heard a particular episode.
Is it legal for a company to watch what I listen to or read?
Yes, with limits. Apps can share listening or reading history with measurement companies when users have logged in and given consent through the app's terms. Privacy laws in Europe and in several U.S. states set rules for that consent and for how the data can be used.