Tracking the impact of the auto-downloading changes to Apple Podcasts in iOS 17
(Updated with data through 2024-01-31)
Apple made a change to auto-downloading in Apple Podcasts in iOS 17, are we seeing a difference in downloads?
Subscribe for updates to all of our stories via our RSS feed
The change was made in the Apple Podcasts version that shipped with iOS 17, originally released to the public on September 18, 2023.
Since we wouldn’t expect to see any effect on downloads until listeners have updated to iOS 17 or higher, the first thing that would be interesting to know is how quickly they are updating.
Let’s take a look at the Apple Podcasts listeners to OP3, an open prefix service tracking over 10 million downloads a month from over 1000 shows of various types, so a nice small sample to work with.
Below is a graph of daily Apple Podcasts iOS version share by unique listeners.
We can see that iOS 17 is rolling out quite quickly, with a few beta testers in August, then an initial rise after September 18, a large increase starting November 19, plateauing in December/January to a level where over 4 out of 5 AP listeners have iOS 17+, crossing 80%.
We show pre-post 16.4 as well, since that was when Apple Podcasts made another change, no longer using the AppleCoreMedia user-agents for some requests.
Note that new iOS versions never reach 100%, some older devices are prevented from upgrading.
- iPhone X, and iPhone 8 / 8 Plus cannot upgrade to iOS 17.
- iPhone 7 / 7 Plus, iPhone SE (gen 1), iPhone 6s / 6s Plus cannot upgrade to iOS 16.
Even after listeners are using the latest version, we wouldn’t expect to see the effect metabolize as quickly as the 16.4 AppleCoreMedia change. That change took immediate effect on affected requests, this latest auto-downloading change will only be noticed when listeners re-subscribe and no longer automatically catch-up to old shows. It’s likely to be at least a year before we can say we have seen most of the effect.
That said, something is happening: to get a sense of the impact in a single graph, take a look at daily total download counts (not shares) for both Apple and Spotify from September 1st. The curves used to be distinct, but now they are almost identical.
This graph can be noisy though, distorted when new shows start using OP3 (and some apps redownload in greater proportion) - I chose a range here that safely avoids these distortions.
In general, looking at app share is a better lens. Let’s keep an eye on OP3’s overall podcast app share for all shows between August 1st (safely before the change) and today.
Some notes:
- OP3 added a few large shows in August and January, so Spotify will be over-represented in that month as Spotify redownloads episodes whenever the episode enclosure url changes. You can see exactly when in the daily graph.
- OP3 measures some NRC shows, and NRC has their own apps, so their app’s share temporarily squeezes every other app’s share when they release popular episodes. Included in the app list above so you can see the effect.
- Apple Podcasts does appear to have consistently lower share once iOS17 hit the majority of listeners. Something to keep an eye on in the coming weeks and months.
Narrowing it down
All shows are not the same, let’s think about the shows that are likely to experience some negative effect on downloads from the iOS 17 auto-downloading change:
- high percentage of Apple Podcasts listeners vs other apps
- release episodes regularly, high number of episodes in general
- listeners often unsubscribe/wait/resubscribe
With this in mind, let’s take a look at an Apple-heavy subset of shows, a handpicked sample of 38 OP3 shows with the following criteria:
- Over 500 total downloads between 2023-08-01 and 2023-08-10 and over 500 total downloads between 2023-12-11 and 2023-12-20
- Over 40% share of Apple Podcasts between 2023-08-01 and 2023-08-10 and over 40% share of Apple Podcasts between 2023-12-11 and 2023-12-20
The intention here is to filter out small shows, or shows that have started or stopped using OP3 or started or stopped publishing altogether during this entire period. Keeping things as constant as possible will better highlight the other effects we are looking for.
After isolating to shows likely to have the most dramatic effect, we see a drop in Apple Podcasts share of downloads of 9-11%, vs 6-8% drop in Apple Podcasts share across all shows.
The share of Apple Podcasts downloads for these 38 Apple-heavy shows went:
- from 65.52% in August to 63.11% in November (-2.41%)
- from 66.89% in September to 63.11% in November (-3.78%)
- from 65.52% in August to 57.01% in December (-8.51%)
- from 66.89% in September to 57.01% in December (-9.88%)
- from 65.52% in August to 56.46% in January (-9.06%)
- from 66.89% in September to 56.46% in January (-10.43%)
Once again, it’s very difficult to keep everything else constant when trying to identify the effect from the iOS 17 change alone. December is traditionally a slumping month for publishing and listening, but now that we’ve seen a full January we can say this change seems to be durable.
We’ll keep an eye on it, and update these numbers periodically.
Data collected and visualized by John Spurlock
Updated 2024-02-01, with eleven new days of data.
Updated 2024-01-21, with ten new days of data.
Updated 2024-01-12, with ten new days of data.
Updated 2024-01-01, with eleven new days of data.
Updated 2023-12-21, with ten new days of data, and after reprocessing. Filter for Apple-heavy shows changed to 40%, now 38 shows.
Updated 2023-12-11, with ten new days of data.
Updated 2023-11-27 thru 2023-12-01, with new days of data.
Updated 2023-11-26, clarified change pcts for Apple-heavy shows (straight share substraction vs pct of pct).
Updated 2023-11-26, with additional focus on 28 Apple-heavy shows, and a few new days of data.
Consider supporting this ongoing research if you find it valuable. We prioritize new areas of focus based on suggestions from our supporters.