Super Bowl LVI Ratings Comparisons: What is the Truth?

Network spin is nothing new, and nothing I haven't talked about in-depth before - most recently in regards to HBO and "Euphoria." However, there really isn't anything quite like Super Bowl-spin. It being by far the most-watched event of the year, its ratings are reported in basically any and every U.S. news publication you can think of, and this reporting is often lifted directly from the network press release. NBC's press release is actually not bad as it's pretty transparent with the ratings; my only qualm is with the historical comparisons.

Looking at the linear TV viewership, Nielsen reported 99.178 million viewers for the game, and NBC does acknowledge this figure in their release - rounding it to an even 99.2 million. Nielsen also reported, and NBC acknowledges, an additional 1.91 million viewers on Spanish-language network Telemundo. Together, this is a linear TV total of about 101.09 million viewers. All of that is fine. The issue comes with how the streaming audience is integrated into their Total Audience Delivery (TAD) figure.

Streaming figures for the Super Bowl have been available since 2012 and have always measured the actual number of streams versus the number of people watching through streaming (the latter is referred to as co-viewing). This is similar to the beginning of Nielsen's linear TV ratings which, at first, measured only households (HH) before later measuring total viewers. I believe that total viewers are a better and more specific measure of the linear audience than households, just as co-viewing data is a better, and more specific representation of the streaming audience. However, we just can't use this metric for historical comparisons.

NBC first reported that an average audience of 11.21 million people streamed the Super Bowl but then later clarified that this figure represented co-viewing - a metric that has never been reported for the Super Bowl before. The comparable figure - i.e. the average number of streams - is actually 6.0 million. This 6.0 million figure is still a nice increase from last year's 5.7 million, but the vast majority of the media are reporting on the 11.21 million figure and using that as the comparison tool to last year's 5.7 million. This is similar to comparing the HH figure from one year to the viewership figure from another. They can't be compared.

NBC includes the 11.21 million in their total audience figure which they report as 112.3 million viewers. "Total Audience" has historically been calculated in different ways, so these are difficult to compare. 

• Until 2012, the total audience was the same as the linear audience.
• In 2012 and 2013 it switched to include both the linear audience and the streaming audience.
• From 2014 through 2017 it included the linear audience, streaming audience, and the Spanish-language broadcast audience.
• From 2018 through 2020 it included the linear audience, streaming, Spanish, and out-of-home viewing.
• Starting for the 2021 Super Bowl, it still included streaming and Spanish-language, but Nielsen began combining the out-of-home audience with the linear audience and reporting them as one.

So when it comes to comparing a "total audience" figure, networks can be very deceiving. For example, NBC stated that - based on an 112.3 million TAD - Super Bowl LVI was the most-watched show in five years, since Super Bowl LI in 2017 (113.7 million). This is misleading on a number of levels, let's enumerate them:

1. As established, 112.3 million includes a unique streaming metric of which was not available for previous Super Bowls. In order to compare, we have to use the 6.0 million figure and with that, the comparable TAD for the game is 107.1 million.
2. The viewership figure they are comparing to from Super Bowls LIV (2020), LIII (2019) and LII (2018) are the ones that do not include the out-of-home viewing data. Because out-of-home viewing is now combined with linear viewing, we have to include it in the TAD from previous years in order to accurately compare. 
3. They can't accurately compare the TAD viewership to any Super Bowl before 2018. Out-of-home viewing for the Super Bowl has always been happening, but didn't start to be measured until 2018. Linear ratings from 2020 and before can all be compared, but it's deceiving to compare audience sizes between years when those audiences are measured in decidedly different ways.  

Year	Linear (millions)	Out-of- home (millions)	Spanish Viewers (millions)	Stream (millions)	Total Audience (millions)
2022	99.178		1.91	6.0	107.1
2021	95.2		0.647	5.7 or 4.6	101.5 or 99.7
2020	100.446	13.0	0.757	3.4	117.6
2019	98.477	12.0	0.473	2.6	113.5
2018	103.391	12.0	0.543	2.02	118.0
2017	111.317	-	0.65	1.72	113.69
2016	111.864	-	0.472	1.4	113.73
2015	114.442	-	0.368	0.973	115.78
2014	112.191	-	0.561	0.531	113.28
2013	108.693	-	-	0.508	109.2
2012	111.346	-	-	0.380	111.7

And unfortunately, when talking about last year's Super Bowl LV, things get even more complicated. Linear viewership was initially reported as 91.629 million viewers, and the total audience was reported as 96.4 million. But even that TAD never made sense to me because taking into account the 5.7 million CBS reported from streaming and the 647,000 Spanish-language viewers, it should've been around 98 million. As far as I can tell, only this Wall Street Journal article (and others who picked it up) ever reported the 98 million figure. Every other media outlet has run with 96.4 million.

It actually even gets more complicated than that. We later found out that Nielsen under-counted out-of-home viewing for a 15-month period which included last year's game, and Nielsen's revised number was 95.2 million for linear+out-of-home viewers, and a 99.7 million TAD. But this TAD also seems off to me. Summing the 95.2 million with the streaming audience and Spanish-language audience, it seems like the TAD of last year's game should be more along the lines of 101.5 million. So, I'd really like to know how and where 99.7 million came about. 

The only thing I can think of is that the streaming figure CBS reported was not placed into the correct context. A Variety article from last year seems to claim that the 5.7 million streams are included in the linear+out-of-home totals (i.e. as part of 91.629 million or 95.2 million). That same article also claims that the streaming total not counted by Nielsen was 4.6 million. By using 4.6 million as the streaming figure and by excluding the Spanish-language ratings, I was able to get to the 99.7 million figure, but I'm left unsure what to do about that. 

Regardless, the 2022 game beat last year's game, but let's be clear that this is very much was not: 112.3 million versus 96.4 million.

The total audience figures in the table from 2012-2020 have been verified from this 2021 Charlotte Observer article which used the same methodology to calculate TAD. Some of these numbers are remarkably different from what the network press releases said, but they are accurate.

Long story short: this year's game was definitely an improvement over last year, but it is by no means the "most-watched show in five years."

Source: Nielsen Media Research via ShowBuzzDaily, Michael Mulvihill, NBC Sports, CBS Sports, Fox Sports, Charlotte Observer, The Hollywood Reporter