We Don’t Always Watch What We Want

A few weeks ago, Janko Roettgers of Variety published an article about which TV networks Americans would purchase if cable providers moved to an “a la carte” model.

Using a survey dataset from Digitalsmiths, a subsidiary of TiVo, Roettgers proclaimed that ESPN, “often quoted as the prime reason for people not cutting the cord”, barely made the cut for consumers’ Top 20 ‘most wanted’ networks.

According to the survey, consumers have clear stated preferences in networks outside of major broadcast networks, such as the Discovery Channel (ranked by survey respondents as number 2), History Channel (5), and Comedy Central (10). Networks like ESPN (20), The CW (39), and Fox News Channel (42), however, seem to rank much lower expected, given their reputation for having large audiences.

At Deep Root Analytics, we use data on media consumption behavior to inform, and often to challenge conventional wisdom. The Digitalsmiths data certainly did this.

This led us to ask the question: how do these stated preferences match up to what consumers actually watching?

To answer that question, we matched a large universe of “Swing” voters against our proprietary, multi-sourced platform of cable set-top box data to see which networks viewers actually watch on TV.

The Deep Root definition of a “Swing” voter includes individuals with a mid-level partisan score who are otherwise demographically similar to the general population.

Using the average target rating for each network, we assigned them rankings — best to worst — and visualized the findings in a Sankey Diagram below.

On the left, networks are listed in Digitalsmiths’ survey response order, ABC being first choice, Discovery Channel being second choice, and so on. On the right, we list how this same network matched up to the average “ranking” of the network across three major media markets — Chicago, Los Angeles, and New York.

Some notable observations:

  1. “The 26th Deadliest Catch.” The Discovery Channel may be the second choice network in the survey, but in these major markets, it is only the 26th most-watched program.
  2. “Diferencia Gigante.” Univision and Telemundo, primarily Spanish-language networks, placed dead last in DigitalSmiths’ survey. But, among swing voters in these major markets, Univision ranks 15, Telemundo 27.
  3. “The Guilty Pleasure Diaries.” More people watch the CW than say they would watch it. The CW did not rank highly in the Digitalsmiths survey, though it managed an average ranking of “4” among swing voters. Our data indicates that swing voters generally watch it much more than Discovery, PBS, and ESPN.
  4. “Number 20 in Your Hearts But Number 7 on Your Scorecard” ESPN ranked 7th among swing voters in our observed viewership, much higher than how it was ranked among what people said they would buy.
  5. “Wicked Tune-out.” National Geographic, number 6 on the “most wanted” list, simply lacks comparable rankings across our three selected markets.

Given the comparable demographics between “Swing” Voters and the general population, what are the factors that drive these differences?

  1. “Desirable Response.” It could be that consumers may say they want certain programming, when their behavior suggests other preferences. Individuals often give response answers they feel will be more “desirable” to the questioner in a survey setting. In the same way that some respondents lie about voting in the last election, some consumers may say they would subscribe to channels like PBS and National Geographic, though they actually would be more likely to subscribe to offerings from TLC or Cartoon Network.
  2. “Aspirational Viewership.” A consumer may want to watch PBS, but simply doesn’t make the time or the effort to do so.
  3. “Substitution.” Viewers may find some acceptable substitutes for their “most wanted” networks in the offerings currently available to them. When faced with only a certain number of options for what to watch on TV, consumers may spend time on networks they otherwise would not frequent, given a better option. The CW, though low on the list of what consumers say they would buy a la carte, is a commonly available broadcast offering. Though a consumer may want to be viewing the Disney Channel, without the right package, he or she may be content with reruns of Supernatural.
  4. “Network Association. ” Finally, consumers often associate a certain network with a certain show. Comedy Central’s The Daily Show and the former Colbert Report may draw entirely different audiences, than say, The Comedy Central Roast of Justin Bieber. When asked what networks he or she would buy, a consumer may very well be thinking in terms of certain programs available with each, rather than the quality of a network’s overall programming.

Each of these explanations speaks to the value of relying on real data to make crucial media buying decisions. When institutions rely on self-reported viewing histories, interested parties subject themselves to data influenced by the problems listed above. As a result, advertisers are forced to accept some uncertainty as to when their targeted customers will actually be watching.

Deep Root Analytics eschews national modeling approaches in favor of real data gathered from TV set-top boxes in every market across the country.

Rather than offer our clients a national top-down estimate of what similar viewers are likely to watch, we provide recommendations based on real, anonymized viewing history collected from individual households, building estimates up to the relevant DMA (designated marketing area), state, and national levels.

By bringing big data to big media, Deep Root Analytics overcomes the problems listed above to provide unique insights to marketing decision makers.

To learn more about Deep Root, visit www.deeprootanalytics.com and follow us at @DeepRootX.

One clap, two clap, three clap, forty?

By clapping more or less, you can signal to us which stories really stand out.