For cable outlets including CNN, ratings are a kind of deity. Every day, producers receive viewer numbers. What people have watched previously determines their coverage. If immigration and climate change don’t rate, they don’t get covered.
I decided to dig into this basic issue. I found that using ratings to determine news value is riddled with potential problems. Basing today’s content on what people wanted to watch yesterday means you inherently lose the “new” in news. What’s worse is that the ratings, at least the way they’re used by TV news producers, might not even present an accurate picture.
First, a quick explainer. CNN, and most other networks, use a company called Nielsen, which has been measuring TV audiences since 1950. They do so using volunteers who are demographically representative. It’s a form of survey, not a literal picture of who is actually watching.
Across the nation, there are 40,000 homes that participate in what is called the “Nielsen panel.” In the largest markets, representing about half of US households, Nielsen knows who is watching, how old they are, their race, what they watched, and for how long. Additionally, a company representative will visit in person periodically. It is somewhat of an intrusion and a commitment, which makes it likely the pool of participants is skewed toward the types of people who have the time and inclination. In smaller markets, Nielsen only knows what is playing on TV in the homes with monitors, which it then matches with demographic data for the initial results.
(During the pandemic, Variety reports, networks are increasingly wary of the audience numbers since members of the panels may not be replaced when they have moved or died, and Nielsen is not sending staff into people’s homes to test the technology. That could result in a “significant undercounting of audience,” according to a trade rep for the networks.)
Every morning Nielsen delivers a raw breakdown of the audience numbers for the day before, known in the industry as the “overnights.” Those numbers are based on the large markets where the numbers are relatively easy to turn around quickly. At 4pm eastern, Nielsen delivers a more comprehensive set of numbers that include smaller markets, and those results are fairly detailed, broken down by quarter-hour. They also present the numbers by age range, and each of the four news networks (FOX, CNN, MSNBC, and HLN) get a side-by-side comparison.
Based on those 4pm numbers, just in time for the biggest audiences of the day in primetime, shows often make changes to their rundowns to reflect what seemed popular the night before. Interestingly, the 4pm numbers often change significantly from the earlier “overnight” Nielsen numbers. CNN and MSNBC tend to do better in the early count based on urban households, but Fox pulls ahead by the time smaller markets are included in the totals.
The numbers change in the same way early and mail-in votes differed from election day votes in 2020. And this means producers are making decisions based on flawed information––they’re constantly, in my metaphor, calling Georgia for Donald Trump. And on top of that, it’s uncertain to what extent data from digital devices is being included in those 4pm reports––the ones that CNN producers are using to make editorial decisions.
“What if someone is watching Nicolle Wallace on MSNBC on their phone? They can’t integrate that data,” says Howard Shimmel, who was chief research officer for Turner Broadcasting from 2014 to 2018 and now is president of Janus Strategy and Insights.
In February 2020, Nielsen reported that about 20 percent of TV time is spent streaming––and that was before the pandemic when many people started cutting the cord. As of three years ago, two-thirds of people aged between 18 and 29 watched TV primarily online. And of all of those streaming, nearly 80 percent live in urban or suburban areas, according to a 2019 study. So if you aren’t counting digital use day-after, then those younger, more urban views aren’t being counted.
There are a couple of reasons it’s so challenging to integrate that data. A major factor is the needs of advertisers. They want to make sure their ads are being viewed as aired, and they also want to make sure that a single user is not being counted multiple times if they watch on multiple devices. The results, to use the jargon, need to be “deduplicated.”
The company says it does include some of the digital data too, but it’s not clear how much. It also says it has a plan for resolving these problems more fully. “With a single, deduplicated number from Nielsen ONE, you will have visibility into audiences by platform, as well as the unique demographic profile of each publisher,” it says on the Nielsen web site. Please note the use of the future tense.
I asked the company three key questions: what digital data is Nielsen capturing; of that data, how much is being integrated into the figures generated the next day at 4pm; and how do those “next day” figures change over time as they are fully able to integrate the data. A representative would only talk off the record, and had not provided formal responses by the time of publication.
I would prefer, personally, that CNN just allowed its excellent journalists to make editorial decisions, instead of desperately seeking the false certainty of numbers. But as that is never going to happen, it is worth noting that even under the current system CNN is horribly under-counting younger viewers in cities, as opposed to older, more rural ones. So that’s who the journalists at CNN are conditioned to produce for.
It might explain why the issues of yesterday are always favored over the issues of tomorrow. And why the younger, more urban parts of America might be better-served by CNN.
Ariana Pekary is the CJR public editor for CNN. She was an award-winning public radio and MSNBC journalist for two decades. Now she focuses on the systemic flaws of commercial broadcast news. She can be contacted at firstname.lastname@example.org or on Twitter @arianapekary.