The Apple Watch will no longer be counted in podcast listener numbers for Interactive Advertising Bureau (IAB) Tech Lab partners because it has been found to falsely inflate listener numbers.
Currently, when a podcast is automatically downloaded by an Apple Watch user, it is counted as two listeners; one from the Apple Watch, and one from its paired iPhone. Since the Apple Watch and iPhone download the same podcast episode by default, and they both report different device user agents, the podcast appears to be downloaded by two different people. This means that the Apple Watch falsely inflates podcast listener numbers.
Although the Apple Watch makes up only a small proportion of podcast listeners, the IAB explains that "Apple Watch devices enact an inherent behavior that triggers non-user initiated podcast downloads and results in significantly inflated download counts from this source." The issue is only present on Apple's own Podcasts app for the Apple Watch.
The IAB Tech Lab's participating members have agreed that by October 1, 2020, all traffic from the Apple Watch will no longer be counted toward any totals reported for the podcasting industry. While filtering out Apple Watch devices will result in reduced audience statistics for some podcasters, the new metrics will be more representative of actual human listeners.
The IAB says that it has "made efforts to work with Apple to support a means of differentiating automatic downloads from valid downloads coming from Apple Watch devices," and if Apple takes "corrective actions" it will revise or retract the guidance.
Top Rated Comments
Ahh--the old 'Dunder Mifflin Infinity' approach....Currently, when a podcast is automatically downloaded by an Apple Watch user, it is counted as two listeners; one from the Apple Watch, and one from its paired iPhone.
"Voyage into Pizzagate"...?
From the server’s perspective, there’s no way to tell whether a user actually listens to a podcast. It’s just an audio file on a server that your app downloads, and then you might listen to it or not. It’s a dirt-simple open standard consisting of an XML file pointing to a list of audio files, and that’s it, which is great.So the problem is counting downloads not listens... My laptop syncs podcasts every day all year long and I only listen to them when a few times a year.
That is exactly the problem, download counts are the most reliable way to measure podcast listeners. The issue is that podcasts are federated, there is no central repository that tracks play counts, and then you have to consider what constitutes a "playback" - if you hit the play button? Listen to 5 minutes? 50% of the podcast duration?I didn't even realize my Apple Watch had podcasts on it. I knew I could listen to them via it, but I thought I had to manually tell it to sync them or something for them to be there.
Counting downloads seems like a very poor idea to me. I have the same episodes downloaded by Apple Podcasts, Spotify, and the Tesla. I'll only listen to each episode from one device though.
There's also the issue that I have a large backlog that I may never actually get around to listening to.
I'd think whenever I download more episodes, it could just send an update on what I've actually listened to.
But then beyond that, every podcast player would have to agree to implement the analytics - overcast, tesla, apple, spotify, etc.