Three days after Thanksgiving, CBS earned bragging rights for serving up the highest-rated NFL broadcast of the season, as the network’s national Sunday afternoon window averaged a 15.2 household rating. In fact, the 4:25 p.m. ET slot carved out a multiyear high, as the last time a regular-season game cracked the 15.0 HH mark was in 2019. Approximately 70% of CBS’ affiliates carried the Eagles’ 37-34 overtime win against the Bills, with the remainder of the stations locked into the concurrent Chiefs-Raiders game.

In what now stands as the most-watched non-Thanksgiving NFL window since 2007, CBS delivered 30.9 million viewers—a fitting capper to the network’s triumphant holiday weekend. On Turkey Day, CBS’ coverage of the one-sided Commanders-Cowboys game averaged 41.8 million viewers and an 11.9 rating. The weird discrepancy between the number of viewers for each broadcast and their respective ratings—10.9 million more people (+35%) tuned in to the Thursday afternoon game, yet Sunday’s window put up a HH rating that was 28% higher than the Tryptophan Bowl broadcast—is one of those things that often causes a fair amount of confusion among fans who keep track of the TV audience numbers. 

If much of the media math can be squared away with a look at the impact out-of-home viewing has had on the NFL’s Thanksgiving deliveries, there are plenty of other facets of the ratings-gathering process that could use a little explaining. Most people are aware that TV viewership data is gathered and processed by Nielsen—the same company has been counting the house since the 1950-51 broadcast season, back when the DuMont Television Network was still in business. But a good deal of the whys and wherefores remain shrouded in a sort of cloud-of-unknowing.

Now, unless you have a vested interest in the business of television, the practice of closely monitoring the Nielsen numbers is a bit like memorizing the weekend box office receipts. Outside of advertisers, network executives and league officials, nobody really needs to concern themselves with the daily ratings grind. Meanwhile, your awareness that the theatrical release of The Super Mario Bros. Movie grossed $1.36 billion isn’t going to make you any richer, nor should that vast sum necessarily serve as a condemnation of the world’s taste in cinema.

If our strange national obsession with audience metrics must endure, it is perhaps for the best if we all had a little methodological intel under our collective belts. What follows is a primer on how Nielsen goes about the thankless task of determining the TV ratings and what those numbers signify once they’re in circulation. As is the case with most statistical analysis, ratings-watching rewards patience and perspective. Sports media is all about the long view. While a few weeks of subpar deliveries isn’t going to get a sport yanked from the airwaves like some crummy sitcom, the consumption metrics play a key role in determining what you see and where you see it.

So how exactly does Nielsen know what I’m watching?

Unless you reside within one of the 42,000 U.S. households monitored by Nielsen, the word “know” is a bit off the mark here. In the absence of any sort of all-seeing omniscience, what Nielsen does is extrapolate its national audience estimates from a much more manageable statistical sample.

Let’s back up a little. What happens is, Nielsen recruits ordinary, TV-owning Americans to serve as the virtual eyes and ears of their TV-monitoring operation. (Back in the day, a Nielsen rep would sling what looked like a do-not-disturb door-hanger tag around your front doorknob, and while the means of recruitment is now a bit more sophisticated, the selection process hasn’t changed much. Rule of thumb: They draft you. This is not a volunteer army.) Should you agree to allow Nielsen to eavesdrop on your TV consumption, a rep will install a data hub and provide Portable People Meter devices to be worn/carried by each eligible member of the household.

What happens if I crack open one of the Portable People Meters and take a peek at what’s inside?

Hired muscle will come to your home, break both of your thumbs, and then they’ll throw all your TVs out onto the street. Kidding. But no—don’t do that. The entire point of the PPM interface is that the data-gathering is almost entirely passive, as the Nielsen gear stealthily picks up a series of inaudible tones embedded within each network’s signal. You just have to sit there like an inert lump of goo and let the TV rays silently wash over you, just like you always have, and Nielsen will take care of the rest. (The base station transmits your daily data to Nielsen’s servers in the wee small hours of the night, while you sleep. That’s how little is actually required of a member of a Nielsen household.)

“Portable People Meters,” huh? So you’re saying they can watch me wherever I go?

No. Nobody’s saying that. (If you’re that paranoid, perhaps being a Nielsen family isn’t for you.) The PPMs are designed to unobtrusively go about their business of monitoring your TV viewing at home, as well as in public venues such as bars, restaurants, gyms and hotels. Once your data is harvested and the relevant demographic information is secured, everything gets scrubbed and anonymized—which means no one will ever know about that you spend more time watching Dr. Pimple Popper than reading books or interacting with your children.

How did they spy on you before the days of technology?

In the old days—and for 140 of the 210 media markets, “the old days” ended in 2017—Nielsen homes were required to keep track of their TV viewing by way of paper diaries, which were to be filled out and returned to the company at the end of every week. Folded inside each diary was a one-dollar bill. Did people lie about what they watched, or otherwise misrepresent their TV consumption? What do you think? Setting aside the notion that $1 doesn’t buy a whole lot of probity, the task of actively monitoring one’s habits gets complicated exponentially in a home with multiple residents. Humans are forgetful creatures that lie all the time for no good reason; whenever possible, it’s best to keep them out of the data loop.    

But why does Nielsen go through all the hassle of surveillance in the first place?

Short answer: Because the TV business generates some $78.7 billion in annual advertising revenue, and that inventory is sold against performance guarantees. In other words, if you buy a 30-second unit to air during NBC’s Sunday Night Football for $900,000, you’re going to want that commercial to reach a very large number of people. More likely than not, an NBC sales rep will quote you a guarantee for total audience deliveries across all platforms, but you also can target specific demos (i.e., males 18-49, adults 25-54, etc.).

Without Nielsen, neither NBC nor anyone else knows how many people watched the ad you paid all that money for, which seems less than ideal. If NBC fulfills the terms of their guarantee, everyone walks away with a warm, fuzzy feeling. If the game stinks and half the East Coast hits the rack at 10 p.m., NBC will make up for the under-deliveries by way of “Audience Deficiency Units,” or ADUs, which is a nice little insurance policy that ensures that the advertiser gets what she paid for.

Ratings, share, deliveries: What do all of these seemingly interchangeable terms mean, and how are TV ratings calculated?

These terms refer to very specific measurements that are calculated by Nielsen, and they aren’t at all interchangeable. Let’s take a closer look through the lens of the biggest TV event of the year.

Fox’s broadcast of Super Bowl LVII drew a record 115.1 million viewers, which is the number Nielsen arrived at after averaging out the number of viewers age 2 and up who were watching during each one-minute interval. Collectively, this headcount can be referred to as “deliveries.” The broadcast averaged a 40.0 household rating, which means that 40% of all U.S. TV homes were tuned in to Fox. (FWIW, that works out to 50 million homes.) The game also did a 77 share, which is another way to say that 77% of all the TVs that were in use during the game were dialed in to Fox.

Demo measurements also can be expressed as either a quantity of viewers or a more abstract ratings figure; for example, Fox’s coverage of the Chiefs-Eagles tilt averaged 44.1 million adults 18-49, which translates to a 33.8 rating among that age group. (We’ll explain why advertisers are so enthralled by this seemingly arbitrary age range in an upcoming primer on demos.)

Incidentally, more than 24 million people watched the Super Bowl from a bar or at a party with friends. In out-of-home scenarios, the PPMs work in much the same way as they do within the confines of a Nielsen panelist’s residence. The OOH impressions are logged and blended with the standard in-home deliveries, but only if the embedded tone is detected by the wearable device. If ambient noise drowns out the signal or the panelist is too far from the audio source for the PPM to detect the tone, no OOH data is generated.

I’m a big Atlanta Falcons fan (rise up!), but they’re never on national TV. Why are the networks such haters?

Have you thought about trying to enjoy something that other people like? As is the case with most NFL teams that aren’t big national draws, the Falcons don’t get a lot of time in the spotlight. Their lone appearance in a coast-to-coast window this season was in an Oct. 1 London game that kicked off at 9:30 a.m. ET, which streamed on ESPN+. Nielsen doesn’t directly measure live streaming events, but rest assured that the NFL isn’t actively trying to hide its best matchups on a streaming service with 26 million subscribers. Not when there are 125 million TV homes in the U.S.

The last time Atlanta was slated in a nationally televised game was Oct. 10, 2021, when they beat the Jets at Tottenham Hotspur Stadium in front of an audience of 3.28 million viewers. Aside from serving as the currency for the TV ad sales market, Nielsen also provides the NFL and its network partners with a level-headed assessment of what works and what doesn’t; in other words, there’s a reason why the Cowboys are scheduled to play 13 of their 17 games in a national window.