Welcome back to “The Needle,” a ratings-focused column on Sports Media Watch that will break down the numbers, attempt to put some context behind the data, and discuss broader trends in measurement and television viewing.
There was once a time when the ratings were pretty easy to discuss. Nielsen’s methodology had not meaningfully changed in decades, and aside from population increases, a viewership figure in 2013 was not too different than in 1993. More importantly, ratings were a niche topic. It was the sort of content one might read about in a Rudy Martzke column, but not everybody was going to be reading a Rudy Martzke column. If you wrote about the ratings, or read about the ratings, you usually knew a bit about the ratings.
In the past decade or so, as social media has reconfigured the fragmented pieces of the internet into a deformed whole, the ratings have become an all-too-frequent topic of conversation, typically used to prove a point about something well beyond the playing field. This has occurred as the ratings themselves have become increasingly complicated. In less than a decade, Nielsen has adjusted its methodology considerably to address long-standing criticism that it was failing to adequately measure live sports viewing.
Nielsen began tracking out of home viewing in 2016 and started including that data in its official estimates four years later, addressing one of the persistent complaints by the networks — that the significant amount of sports viewing done outside the home was not being tracked at all, much less adequately. Nielsen began its “Big Data” era not long after, combining its existing panel with data from set-top boxes and smart TVs, and most notably, first-party data from participating streamers. That data became the official currency this past Monday.
It can hardly be questioned that the numbers Nielsen is reporting now are closer to accurate than the numbers being reported six-plus years ago. But it should go without saying that comparisons to six-plus years ago, or even one year ago, are less accurate.
It is worth noting that when Amazon’s Prime Video began reporting its “Big Data” figures for NASCAR over the summer, all of the comparisons were limited to the “Big Data” era — which began in 2023. Neither Amazon nor NASCAR compared the viewership figures to panel-only figures.
But according to an industry spokesperson, it is official Nielsen policy to compare the “Big Data” numbers to the panel-only figures from prior years. That means pretty much every event for the next year will have a built-in cushion for a year-over-year increase. While that does not guarantee gains — a Brewers-Blue Jays World Series probably is not catching Yankees-Dodgers no matter what methodological changes are at play — it certainly makes for increases on paper that may not always exist in reality. Let’s say a game increases 4% from last year, does that mean more people are watching? Or that the methodology changed?
These methodological changes have occurred at the same time that third-party reporting of the ratings has dropped off sharply. There was once a time when sports ratings were listed each week in places like Media Life Magazine and even ESPN.com — which ran a weekly “top ten” on its now defunct ‘ESPN the Life’ page — and daily broadcast and cable numbers were published on sites ranging from The Futon Critic to TV By the Numbers to ShowBuzz Daily.
While there are still some sites publishing daily ratings, it is not a coincidence that such publicly available information has dried up. In the absence of third-party reporting of the numbers, the networks are free to tell their own story — and one can take a wild guess at the kind of story being told by PR.
Even the most honest, transparent and communicative PR practitioner is not going to share all of the context surrounding a viewership figure. For example, Nielsen’s increases in out-of-home viewing are not something you will read about in a press release. It falls on the person circulating the PR-provided figure to have an understanding of that context and to convey it to readers. Yes, the networks’ viral graphics will generally include a “Nielsen Big Data + Panel” attribution in tiny font in the corner. But few are going to notice that, and even fewer will understand what it means. They will simply see the claims of ‘highest ever,’ and never know that there have been a few changes to how Nielsen runs the numbers.
And with sports ratings becoming an easy way to drive social media engagement, the numbers are in too many cases being circulated without an eye toward context — or even at times truth. At one point this year, social media circulated a wholly false stat that Caitlin Clark’s first game back from injury generated more than 20 million viewers on ABC (the actual number was 2.2 million). Clark, the biggest individual ratings driver in all of sports, is particularly prone to factually questionable ratings content. Have you heard that WNBA viewership has tanked without her? From what to what, over what period of time, and on which networks? All of that is immaterial. (For what it is worth, ABC and CBS have both said they had their most-watched WNBA seasons on record, and ION at last check said it was down a slight three percent year-over-year.)
So there is an odd combination at work. Social media is not bound to factual accuracy, and PR is not bound to full transparency. As a result, the ratings conversation has transformed into a series of abstract digits and lofty claims. 9.1 million! Up 58%! Down 55%! 14-year high! Fox last week published a PR graphic touting 3.8 million for “Big Noon Kickoff,” a figure that encompassed only the final hour of the show. The full, three-hour episode averaged about half that number, 2.0 million (a figure that Fox willingly disclosed to reporters). But how many people scrolling on social media would even think to notice the “starting at 11:00 AM ET” disclaimer in smaller font?
Last week’s opening weekend college football numbers were so good that they were celebrated by Pat McAfee at the top of this week’s “College Gameday.” McAfee, claiming that this is “the biggest college football season of all time,” noted that week one featured “more viewers on every game, more viewers on every show. The sport is hotter than it’s ever been.”
Indeed, the week one numbers were so strong that it is unlikely that they could be fully explained by Nielsen’s methodological changes. The 16.6 million for Texas-Ohio State is high enough that even taking out Big Data and out-of-home viewing, it is likely that the game would still have beaten, or at least challenged, Alabama-FSU in 2017 — which did not have either included in its numbers — to rank as the most-watched week one game on record.
But understand that every percentage gain, every record high, is compared to numbers that were measured differently. This is not comparing a home run record in a 162-game season to a record in a 154-game season. It is rather as if every home run counted for more in a 162-game season than in the 154-game era. Is college football really hotter than it’s ever been? Or would 2017 have had a real argument if the numbers were being measured then as they are now?
The impression given is that more people are watching than at any point in X years — or ever. Sometimes, the numbers will be so good that such a statement will ring true. Other times, it will be fair to wonder if more people are watching or if more viewing is merely being accurately measured.
And yet it would not be fair to say that the present-day numbers are inflated. They may be inflated relative to past years, but they are almost certainly still undercounted, as the NFL argued this week. Even with out-of-home, even with “Big Data,” it is fairly likely that the entirety of sports viewing has yet to be fully accounted for. (To begin with, illegal streams are not rated.)
But the real story of the television ratings has always been the comparison, more than the raw number. Without context, the ratings are just a number sitting out in space. You can mock up a Photoshop graphic with “5.9 million!” and a picture of Naomi Osaka and it really doesn’t mean much without knowing what the prior year’s match drew. (In the current social media era, that may still be enough to go viral.) The contextual cues are the primary way one gauges performance, and “apples and oranges” may be understating how skewed these comparisons have become.
In some ways, as the ratings become an ever-present part of the sports media discussion, they are less meaningful than they have ever been.
On a semi-related note, the only two network researchers anywhere with a public profile — ESPN’s Flora Kelly and Fox Sports’ Michael Mulvihll — were vocally critical of Nielsen in the hours ahead of YouTube’s exclusive NFL game Friday night. Kelly wrote on social media that the eventual YouTube number will not be “a fair comp” because Nielsen has created “a custom methodology” for the game that is not “the same approach as the rest of us, nor MRC accredited.”
Mulvihill added that the YouTube figure “will not be made available to other Nielsen clients,” which he called “a flagrant departure from Nielsen’s history of transparency and a slap in the face to long-standing clients. When it comes to streamers the rules simply don’t apply.”
“Nielsen’s strength isn’t in perfect accuracy,” Mulvihill wrote. “Their value comes from impartiality and transparency. When they move away from those tenets they undermine what makes them the only essential data source in media.”
This is far from the first time that Mulvihill has voiced complaints about how streaming viewership is handled. But if as Mulvihill claims this is an example of Nielsen moving away from transparency, if it is the case that data is now less available for third-party review and analysis, it is hard to be overly sympathetic. If this is where things are heading, it has been to the networks’ benefit.
