Summary
iOS 17 changes how Apple Podcast users auto-download.
From Apple:
Before iOS 17, when a listener would unpause automatic downloads, the system would automatically download all unplayed episodes. With iOS 17, Apple Podcasts will not download previous episodes and will resume automatically downloading new episodes.
Apple Podcast listeners will now auto-download significantly fewer back-catalogue episodes, resulting in a 10-30% decrease in monthly downloads for most podcasts. The larger the podcast’s back-catalogue the greater the impact. For advertisers, this means fewer impressions will be served to Apple Podcasts auto-downloads that likely were not heard. This affects run-of-show / impression / back-catalogues buys more than new-episode / episodic / baked-in buys.
Prior, we would see most run-of-shows campaigns deliver up to 20-40% of impressions to ~3% of listeners, eg:
Now, we see the same campaigns delivering 15-25% of impressions to 3% of listeners — same campaign now:
High Concentration (definition). This is likely a positive for advertisers, since a greater percentage of impressions in back-catalogue campaigns will be listened to. The performance of these impressions that go to such a small percent of listeners has been a fraction of other impressions.
This also highlights the importance of trusted, 3rd party verification.
As of mid December 2023, it’s estimated at least 2/3 of iPhone users have updated to iOS17. Stay tuned here for adoption updates!
Back-story
Bloomberg tells this story better than we could.. but here is our abridged version.
We noticed a lot of impressions were going to a small percent of listeners at some point in 2022. We suspected we were just seeing many users with the same user-agent on noisy IPs, so disregarded it. However, we began to change our mind with new evidence:
- the IP types were mostly residential, not mobile or commercial
- almost all of the user-agents downloading dozens or hundreds of back-catalogue episodes were Apple podcasts. Where were all the Spotify user agents on this airplane, or in Times Square?
- these IP addresses soaking up a 20-50% of downloads had few conversions. If there were many people behind them, conversions should be roughly on par with those associated with other IPs. Eg. as shown below in a client’s campaign, the 29% of impressions going to 2% of listeners are only responsible for 9% of conversions. The 2% of listeners over-perform by 4x, however, the impressions under-perform by a factor of 3.
We told our most impacted advertisers, who then told publishers. Most, like us initially, were skeptical of our findings. As they inspected their own data, most soon arrived at our same conclusion — that a small subset of Apple Podcasts users were downloading dozens to hundreds of back-catalogue episodes inexplicably.
The IAB, hearing the industry, began to work on a solution to this. However, before an update to the IAB spec could be agreed upon, Apple released their update 🙂
Powered by