But I can’t seem to find any data on actual Spotify viewership to pin these numbers down to be able make some apples to apples comparisons.
There are several third party research firms, most notably Edison, that compile their own data from the client side, akin to how Nielsen produces cross-platform numbers for television and streaming. And others compile data from different sources, including published ordinal rankings and publicly reported view counts, other metrics like mentions/shares on social media.
Podcasting historically predates always-online devices, so a lot of the interfaces count downloads rather than listens, as the tech companies try their best to push things towards data-tracking proprietary client apps. So it’s a non-trivial effort (usually requiring lots and lots of paid survey participants, which also introduces some statistical skew), to get listener data.
So there is data out there, almost all behind business-to-business paywalls/subscriptions. Those closely guarded datasets are what advertisers use to price ad campaigns on these channels, for example, where they have to decide whether and how to spend on one particular type of medium versus another (e.g., product mentions on podcasts versus YouTube versus twitch versus traditional television or film).
Also do people not “watch” podcasts?
The NYT article I linked describes the issue somewhat, that there’s a difference between viewing as a first screen or second screen, or listening as primary, or listening as background, and that advertisers and publishers are aware of how different shows perform at different functions. They describe the example of the political podcast, Pod Save America, having stats showing that about 20% of their audience is through video rather than audio-only.
As I understand it, through my own personal experience, podcasts are popular during commutes, as a primary audio experience for people driving alone in their cars or wearing headphones on the bus/train/sidewalk. Those are functions in which video is not desired, so “watching” would be a strange way to describe how people primarily consume the content, through listening only. And while I know official music videos get a lot of watches on YouTube, I don’t think counting those watches are a good way to rank or analyze how people “listen” to music overall.
It’s also why many of the most popular podcasts simply don’t do video (NYT’s popular The Daily), and stem from radio roots (the NPR podcasts, a bunch of iHeart or Sirius-published podcasts). Video isn’t a good proxy, especially for the podcasts that are tied exclusively to a non-YT platform.
Maybe SiriusXM overpaid for their deal, at $40 million per year. Maybe Spotify overpaid for their previous deal, at $30 million per year. But those numbers alone aren’t anything to scoff at, and in my opinion are the best proxy for cross-platform comparison: Smartless got about $33 million per year from SiriusXM, Dax Shepard got $27 million per year from Spotify, etc.
And maybe you’re right that podcasting itself isn’t ever going to be as big as video or streaming. The South Park guys got $900 million for 6 movies on Paramount+. Apple TV dropped more than 9 figures each on Killers of the Flower Moon and Napoleon.
But for advertisers, that specific podcast is a valuable property. I can’t comment on the quality or cultural comparisons because I literally have never heard any part of it (or other huge genres like true crime podcasts or celebrity podcasts or video game reviews or whatever other things people consume), but I watch the business side of things because I’m interested in that.
And I’m not trying to give any commentary on the cultural relevance of this particular Harris interview any more than I’d give commentary on her appearance on Howard Stern (who I might not have heard about in like 20 years) or Walz’s appearance in World of Warcraft or something I don’t actually understand the logistics of. Politics isn’t really something I care about, but tech and media businesses are.
So it seems like your argument is that podcasts on captured platforms use downloads, not views as rankings, we have to use that instead. But also, we don’t have access to that data. Because of this, it seems implied in your second paragraph (and you’ve mentioned it before) that we should use $$ as a proxy for viewership (since we can’t measure viewership directly) assuming our goal here is still to get these podcasts off of a discrete scale and into a more continuous scale.
Which is like totally fair, except that only very public shows/podcasts etc are going to publish those numbers. It’s a totally valid way to do it, but it comes to the same issue I identified previously, we only have limited access and we can’t put a number down for each data point. I know we can pretty easily map views to $$ on YouTube. And if we assume a $ is a $ and is roughly equivalent to a similar number of views, we can get to an apples to apples comparison.
I don’t agree with this:
And while I know official music videos get a lot of watches on YouTube, I don’t think counting those watches are a good way to rank or analyze how people “listen” to music overall.
I don’t agree. I think yt views are the least plagiarisable metric we have access to. And it’s the only one we have access to that’s a uniform proxy across all the discussants. Every single podcast or show also puts the same stuff in YT.
I would rank the reliability of continuous metrics for this conversation at YT views or comments first, then if accessable, downloads, then dollar amounts (published or derived from yt views).
I think we could pick a set of what we anecdotally consider ‘top’, ‘mid’, and ‘basic’, teir shows or podcasts (there’s no real difference in my mind; I’m watching the same seder show rn: but I’m not I’m listening to it.). If we can get maybe 3-6 for each tier, and we can get them for “Call Her Daddy”, we can evaluate the podcast via all three metrics. We could then weight the three metrics individually and average the two.
There are several third party research firms, most notably Edison, that compile their own data from the client side, akin to how Nielsen produces cross-platform numbers for television and streaming. And others compile data from different sources, including published ordinal rankings and publicly reported view counts, other metrics like mentions/shares on social media.
Podcasting historically predates always-online devices, so a lot of the interfaces count downloads rather than listens, as the tech companies try their best to push things towards data-tracking proprietary client apps. So it’s a non-trivial effort (usually requiring lots and lots of paid survey participants, which also introduces some statistical skew), to get listener data.
So there is data out there, almost all behind business-to-business paywalls/subscriptions. Those closely guarded datasets are what advertisers use to price ad campaigns on these channels, for example, where they have to decide whether and how to spend on one particular type of medium versus another (e.g., product mentions on podcasts versus YouTube versus twitch versus traditional television or film).
The NYT article I linked describes the issue somewhat, that there’s a difference between viewing as a first screen or second screen, or listening as primary, or listening as background, and that advertisers and publishers are aware of how different shows perform at different functions. They describe the example of the political podcast, Pod Save America, having stats showing that about 20% of their audience is through video rather than audio-only.
As I understand it, through my own personal experience, podcasts are popular during commutes, as a primary audio experience for people driving alone in their cars or wearing headphones on the bus/train/sidewalk. Those are functions in which video is not desired, so “watching” would be a strange way to describe how people primarily consume the content, through listening only. And while I know official music videos get a lot of watches on YouTube, I don’t think counting those watches are a good way to rank or analyze how people “listen” to music overall.
It’s also why many of the most popular podcasts simply don’t do video (NYT’s popular The Daily), and stem from radio roots (the NPR podcasts, a bunch of iHeart or Sirius-published podcasts). Video isn’t a good proxy, especially for the podcasts that are tied exclusively to a non-YT platform.
Maybe SiriusXM overpaid for their deal, at $40 million per year. Maybe Spotify overpaid for their previous deal, at $30 million per year. But those numbers alone aren’t anything to scoff at, and in my opinion are the best proxy for cross-platform comparison: Smartless got about $33 million per year from SiriusXM, Dax Shepard got $27 million per year from Spotify, etc.
And maybe you’re right that podcasting itself isn’t ever going to be as big as video or streaming. The South Park guys got $900 million for 6 movies on Paramount+. Apple TV dropped more than 9 figures each on Killers of the Flower Moon and Napoleon.
But for advertisers, that specific podcast is a valuable property. I can’t comment on the quality or cultural comparisons because I literally have never heard any part of it (or other huge genres like true crime podcasts or celebrity podcasts or video game reviews or whatever other things people consume), but I watch the business side of things because I’m interested in that.
And I’m not trying to give any commentary on the cultural relevance of this particular Harris interview any more than I’d give commentary on her appearance on Howard Stern (who I might not have heard about in like 20 years) or Walz’s appearance in World of Warcraft or something I don’t actually understand the logistics of. Politics isn’t really something I care about, but tech and media businesses are.
So it seems like your argument is that podcasts on captured platforms use downloads, not views as rankings, we have to use that instead. But also, we don’t have access to that data. Because of this, it seems implied in your second paragraph (and you’ve mentioned it before) that we should use $$ as a proxy for viewership (since we can’t measure viewership directly) assuming our goal here is still to get these podcasts off of a discrete scale and into a more continuous scale.
Which is like totally fair, except that only very public shows/podcasts etc are going to publish those numbers. It’s a totally valid way to do it, but it comes to the same issue I identified previously, we only have limited access and we can’t put a number down for each data point. I know we can pretty easily map views to $$ on YouTube. And if we assume a $ is a $ and is roughly equivalent to a similar number of views, we can get to an apples to apples comparison.
I don’t agree with this:
I don’t agree. I think yt views are the least plagiarisable metric we have access to. And it’s the only one we have access to that’s a uniform proxy across all the discussants. Every single podcast or show also puts the same stuff in YT.
I would rank the reliability of continuous metrics for this conversation at YT views or comments first, then if accessable, downloads, then dollar amounts (published or derived from yt views).
I think we could pick a set of what we anecdotally consider ‘top’, ‘mid’, and ‘basic’, teir shows or podcasts (there’s no real difference in my mind; I’m watching the same seder show rn: but I’m not I’m listening to it.). If we can get maybe 3-6 for each tier, and we can get them for “Call Her Daddy”, we can evaluate the podcast via all three metrics. We could then weight the three metrics individually and average the two.