(Welcome to the Entertainment Strategy Guy, a newsletter on the entertainment industry and business strategy. I write a weekly Streaming Ratings Report and a bi-weekly strategy column, along with occasional deep dives into other topics, like today’s article. Please subscribe.)
Today, I’ve got a grab bag of data stories, featuring some items that got cut for space in my Streaming Ratings and “Most Important Story of the Week” columns over the last month or so but that I felt were well worth sharing anyway.
The theme? Politics.
But the analysis is non-political and non-partisan, mainly focused on data and the media. (It feels like the election happened a long, long time ago…but it’s only been, what, six weeks?)
In today’s article, I’ll share with you how many Americans actually care about politics (less than you think), election ratings (still big!), and explain what a “view” on YouTube actually is, since far too many pundits credulously take YouTube views at face value or compare them to traditional ratings. Then I’ll debunk the idea that this was the “podcast election” or that the left needs its own Joe Rogan. (All of this is coming from an early, early podcast adopter who is easily in the .001% of all time podcast usage…but I can separate my habits from everyone else’s.) Finally, I’ll reveal how consumer sentiment has (unsurprisingly) already rebounded post-election.
Again, the theme is politics, but the content isn’t political.
Let’s dive right in!
Most Importantly, Seriously, Most Americans Don’t Care About Politics
As I’ve noted before in my Ankler series on the average American viewer, when it comes to politics, Hollywood should remember that…
- Roughly one-third of the country votes Democrat
- Roughly one-third voted Republican
- The other third of adults don’t vote.
This year, the number of non-voters went up. In a country with an estimated 260 million adults, 75 million people voted for Harris, 77 million voted for Trump, and 100 million adults (or more!) didn’t vote. At this point, a plurality of Americans didn’t vote, which means a plurality of Americans don’t care about politics.
Elections Get Great Ratings
That said, a medium-sized subset of the country can yield huge ratings. According to Samba TV, 35.6 million households watched election night coverage (from 1:00 Eastern to 9:00 Eastern). For comparison, that’s only 3 million less than their Super Bowl rating.
According to Nielsen, 42.3 million people tuned in across 18 networks, a big drop from 2020’s 56.9 million viewers. That’s far less than Nielsen’s Super Bowl rating, but double this year’s Oscar rating, and 10 million more than the Macy’s Thanksgiving Day parade.
Let’s recap this year’s political TV show ratings (all according to Nielsen):
- 50 million people tuned in to the Biden-Trump debate.
- 24 million people tuned in to Biden’s post-debate press conference.
- 19 million people watched the news after Trump’s assassination attempt.
- 28 million people watched Trump’s convention speech.
- 28 million people watched Harris’ convention speech.
- 29 million people watched Biden drop out of the race.
- 67 million people watched the Trump-Harris debate.
Triangulating these numbers, probably 60 million people (30 million Democrats and 30 million Republicans) are really interested in politics, at least watching it on TV. (Most of these numbers include streaming.) I already know that some people are going to immediately rebut me, arguing “Well, everyone I know (or follow on social media) is into politics,” but even Nielsen reported that the election night coverage only had 22% market share; nearly 80% of TV viewers were watching something else.
What’s a View on YouTube Anyway?
Let’s talk YouTube “views”.
On Ezra Klein’s election recap podcast—in a section talking about Bernie Sander’s controversial appearance on The Joe Rogan Experience, and Harris not going on said show, but Trump getting interviewed on it—Klein mentioned that Trump’s episode had “46 million views on YouTube”. (As of today, it’s up to 52 million.)
But what does that actually mean?
As someone who writes a weekly Streaming Ratings Report, I worry that people equate YouTube views to streaming “views” (also called CVEs, or “Completed View Estimates”, as Netflix and some datecdotes from the streamers use, taking the total hours viewed and dividing it by runtime) when they couldn’t be more different. Or worse, they assume that 46 million Americans (emphasis on “Americans”) mostly watched the whole thing, equating it to Nielsen’s average minute audience, which it most emphatically is not.
If those 46 million views translated into customers watching all three hours of Trump’s Joe Rogan interview, that’d be 138 million hours, or the second most popular single show on streaming all year by a wide, wide margin. (Behind Bridgerton!) But it’s not even close to those numbers for a few reasons…
- YouTube “views” are vague and don’t tell you how long someone watched something. According to Google, er, Google’s AI overview, you only have to watch 30 seconds of a YouTube video for it to count as a view, but that stat comes from YumYumVideo, which doesn’t have a link to support this assertion. This support page goes quite in-depth on ad metrics but not views. Other people on Reddit believe a video loading is enough to count as a view. Plus, a viewer who stops watching a video and then returns to it later might count as multiple views. (For a three-hour video like Trump’s Rogan appearance, I bet this happened a lot.) Either way, it’s not clear or totally transparent how it’s calculated. Regardless, it’s not close to telling you how many customers watched something for how long.
- YouTube has a spam/bot problem. Google claims that they can track bots and remove their views from the YouTube tallies. I’ll leave it up to you if you trust that, since more ad views = more revenue from advertisers. (And other tech companies have gotten in trouble for this before.) And I’ve seen estimates that half of internet traffic is fake. (To be clear, I believe that most of these bots are crawling/scraping the web for information, including Google’s own bots indexing search sites! The problem is these bots still interact with web pages.) We just don’t know how many bots get counted as actual views. It could be up to 50%, so my confidence interval is anywhere from 15-50% of views are spam/bots.
- Nielsen’s Top Ten charts just show American viewership; YouTube stats are global. This is probably my biggest gripe with people citing YouTube views; 75% of the time, they’re making a point about America, American media, or American viewers, but they’re citing global data. (Writers, columnists and analysts do this all the time; I’ve literally seen it three times this week.) That’s not apples-to-apples. Now, Joe Rogan knows his foreign viewership numbers, but we don’t. There’s no way every one of those 46 million views came from the US. It might be as high as half of them!
- Trump’s appearance on Rogan was three hours long, but not every viewer watched all three hours. I bet a lot of people, especially on the left, clicked through to watch a short portion of this video but not the whole thing. Many news stories or social media posts also linked to this video and clips from it, and those viewers very, very likely didn’t watch the whole thing. Or viewers skipped some parts; as an avid podcaster, I do this all the time for most podcasts I subscribe to.
Translating Views into Hours Viewed
Knowing those factors, that’s why I heavily discount YouTube views in my head. Honestly, whenever I see a YouTube views number, I do this math right off the bat:
- I assume 50% of views are from outside the US.
- I assume 50% of the rest comes from bots.
- I assume most non-music videos are watched for about 2 to 5 minutes.
So if I read an article calling a given YouTuber “supremely popular”, most of the time, this math explains why they aren’t as big as their “views” imply. Say they have a “million views per episode”, here’s how that breaks down for me:
(Note: When I mention “spam” or “bots”, this is not to imply YouTubers manage these bots. I’m specifically saying the opposite; the internet/YouTube has a bot problem, and it impacts everyone equally.)
Of those 125K viewers, assuming a 70% completion rate (good for streaming, honestly), for a half-hour YouTube show, you’re looking at something like 48 thousand hours viewed. That’s the equivalent of a Nielsen rating of…96K average minute audience. That’d be one of the poorest-rated shows on TV. (This answers the question, “Why do so few advertisers sponsor hit YouTubers?”, frankly.)
I’d run the same math for Joe Rogan.
Using some back-of-the-envelope math—and assuming Joe Rogan bested my usual estimates—I’d subtract 10 million non-US views (which is generously low), another 5 million of bot traffic (also a number that could be much higher) and maybe estimate that another 20 million people who watched two minutes or less or watched clips non-consecutively and got counted as multiple views, so let’s say maybe 11 million people watched more than two minutes of this YouTube video. My traditional math would be even smaller: about 5.7 million viewers.
Comparing that to streaming, 12 to 23 million hours is good, but not one of the twenty-five biggest debut shows this year. Is it still big and influential? Absolutely yes! Did it single-handedly decide the election? That assertion doesn’t make sense to me.
Did Joe Rogan Bring Voters to Trump or Did Trump Deliver Viewers to Joe Rogan?
But actually, we can triangulate this even more. The common narrative is that the bro-cast-o-sphere or whatever brought voters to Trump, but what if Trump brought voters to Joe Rogan?
Again, getting causation is really important! Don’t mix it up for correlation at your peril.
So here goes. I tracked the views of Joe Rogan’s episodes going back to early October. Pulling out three guests, here’s how it looks:
Very rarely does Joe Rogan have guests that eclipse three million views within their first two months. (YouTube videos tend to keep growing over time as library content, remember.) To be blunt, that implies that the impact of going on Joe Rogan—outside of the show going viral—is to reach about 3 million average views.
Knowing that baseline, three of his guests stand out. Here they are:
In other words, it doesn’t look like Joe Rogan introduced Trump to 46 million people, but Trump drove 43 million extra views to Joe Rogan’s YouTube channel. That makes sense! Trump is very, very, very popular, had an entire campaign apparatus pushing his messaging, and it was the main news story for a day, meaning tons of links to this video. Honestly, it wouldn’t surprise me if Trump’s campaign used their marketing dollars to advertise this. I mean, J.D. Vance got 18 million views too!
Again, Accuracy Matters
The point isn’t that Rogan isn’t highly, highly influential; he is! Millions of Americans listened to this podcast or watched this YouTube video. (And I’ll leave it to other commentators to debate if Kamala should have gone on the show.) But we also shouldn’t compare Joe Rogan’s numbers to established media properties without the proper context.
But I can’t blame other outlets for not adding this context! It makes for boring prose. Adding a caveat every time you mention YouTube views (mentioning that it’s a poorly defined and vague global number) leads to more accurate but far less dynamic writing, so few people do it.
This is why I look at “hours viewed”. And right now, YouTube doesn’t provide that information publicly, likely because it would hurt their business model and perception. I wish they shared views and average view length; then we could triangulate this info way, way better. I wish YouTube shared viewership by country, too, but they won’t, probably because it might hurt advertising dollars.
I’d add, this isn’t a pedantic data point. Big coverage of news events influences how decision-makers—from campaign managers to media executives—make decisions. Too many people saw the Rogan YouTube views and drew strong, definitive conclusions from it. If you want to make better decisions, you need an accurate view of ratings, and this media coverage and analysis and punditry didn’t provide it.
Speaking of narratives misleading decision-makers…
No, This Wasn’t the “Podcast Election”
Post-election, I’ve read a ton of people talking about how 2024 was the “podcast election” or the “creator election” or repeating the old saw about how every election represents the new media landscape, and this one was podcasts. On the left, pundits complained that there’s no left-wing podcast ecosystem, compared to the influence of the aforementioned Joe Rogan.
I’m skeptical.
And this is coming from a voracious podcaster. Yes, I was an early, early podcast adopter (circa 2006), but even I’m incredibly skeptical about podcasts’ reach today and how much this affected the election, for four reasons:
We’re just getting started with this topic, but the rest is for paid subscribers of the Entertainment Strategy Guy, so if you’d like to know…
…Four reasons why this wasn’t the podcast election.
…and why this matters, strategically, if you care about winning elections.
…and how voters’ views of the economy have already flipped.
…please subscribe! We can only keep doing this great work with your support.
