Is a 20% download drop normal?

It depends entirely on the shape. A 20% fall spread over several months is within normal variation for a show with a changing publishing rhythm or a seasonal audience. A 20% fall between one Tuesday and the next is not a variation, it is a change in what is being counted, and the cause is almost always on the measurement side.

Did Apple change something that reduced podcast downloads?

Yes, in iOS 17. Apple Podcasts now pauses automatic downloads when a device runs out of storage, or when a follower has not played a show in more than 15 days and has not played the latest five episodes. It also stopped downloading a backlog of unplayed episodes when a listener returns. Measured effects reported at the time ranged from roughly 10% for daily shows to around 30% for less frequent ones.

How do I tell a measurement problem from a real audience decline?

Look at the shape and the split. A measurement change produces a cliff on a specific date, often confined to one app or one platform in your breakdown. A real audience decline is gradual, shows up across every app at once, and is usually visible in completion rates and follower counts before it reaches your download total.

Why are my podcast downloads dropping?

If your podcast downloads are dropping, start by assuming your listeners did not leave. A sharp fall in downloads is far more often a change in what gets counted than a change in who is listening, and the checks that separate the two take about twenty minutes.

Work through these causes in order. They are ranked by how often each one turns out to be the answer, not by how alarming they sound.

1. Your host changed its filtering

This is the first thing to check for any drop that happened overnight, not over months.

Hosts periodically tighten the filters that strip bots, crawlers and duplicate requests out of your numbers, and those updates sometimes go wrong. In 2026 Podbean rolled out a filter intended to remove AI and crawler traffic and it marked every Apple Podcasts download as invalid, leaving some shows with plays visible in Apple Podcasts Connect and zero Apple downloads in their host dashboard. Our write-up of the Podbean Apple Podcasts filtering incident covers the audit to run when this happens.

Filter changes also arrive as scheduled compliance work. The IAB Tech Lab's podcast measurement guidelines have moved through versions 2.2 and into a 2.3 draft that adds guidance on URL prefix measurement, RSS enclosure URL changes and invalid traffic. Each tightening removes traffic that was previously counted.

The fix: check your host's status page and changelog for the week the drop started. Then compare the same window against a second source, such as Apple Podcasts Connect or Spotify for Creators. If your host fell and the platform dashboards did not, the problem is the filter, not the audience. Preserve the affected reporting window and do not restate sponsor numbers until the host confirms what happened.

2. Apple's automatic downloads paused for your lapsed followers

This one is structural and permanent, and it catches shows that have been running for years.

Since iOS 17, Apple states in its Podcasts for Creators documentation that "Automatic Downloads are paused when a device is out of available storage or when a listener hasn't played a show they follow (latest five episodes for more than 15 days)." Apple also stopped downloading the backlog: before iOS 17, a returning listener triggered downloads of every unplayed episode, and now the app resumes with new episodes only.

The effect is that followers who drifted away stop generating downloads. Reported impacts at the time varied by cadence, with National Public Media's Bryan Moffett describing around 10% monthly declines for daily shows and roughly 30% for longer shows on other schedules, while Podtrac suggested an average 15% drop for most large shows.

The fix: there is no setting to reverse. What you can do is re-engage lapsed followers so their downloads resume, which is genuinely what the change is measuring. A dormant follower who plays one episode restarts automatic downloads for that show.

3. Your feed or prefix broke

Two versions of this are common. A feed URL changed during a host migration without a permanent redirect, so directories keep polling an address that no longer updates. Or an analytics prefix was dropped when episodes were re-uploaded or the feed was rebuilt, so the requests bypass measurement entirely.

The fix: open your public RSS feed and check that the enclosure URLs still carry your prefix. Then request an episode file directly and confirm it resolves. A prefix problem usually shows as a drop to near zero, not a partial fall, which makes it easy to identify once you look. How to measure podcast downloads covers what should be in the chain.

4. You changed your publishing rhythm

Downloads follow episodes. A month with three episodes instead of five produces fewer downloads even with an identical audience, and a skipped week removes a release-day spike that would otherwise sit in the total.

The fix: compare downloads per episode at a fixed age, such as 30 days after publication, rather than downloads per month. If per-episode performance held steady, you published less, which is a scheduling decision, not a decline. Podcast download benchmarks sets out what normal per-episode movement looks like.

5. Seasonality caught you

Podcast listening tracks commuting and routine. Holiday weeks, summer in your main market, and the end of December reliably soften downloads for most shows, and the recovery is equally reliable.

The fix: compare against the same period last year rather than against last month. If last December dipped by a similar amount and recovered in January, you have a calendar, not a problem.

6. Your audience actually shrank

This is last for a reason. A genuine decline rarely produces a cliff. It looks like a slow slope across every app in your breakdown at once, and it usually shows in completion rates and follower growth before it reaches the download line.

The fix: check whether new listeners have stopped arriving, or whether existing ones have stopped finishing. Those are different problems. Falling acquisition is a discovery and promotion issue. Falling completion is an episode problem, and how to improve podcast downloads covers both sides.

How to stop guessing next time

Three habits make the next drop diagnosable in minutes.

Record the shape before you react. Screenshot the daily download chart on the day you notice a change, so the start date survives any later reprocessing.

Split by app every month. A drop confined to one app is a platform or filter change. A drop across all of them is either seasonal or real.

Keep one measurement point that does not move. Platform dashboards each count their own way and change their definitions without notice. What counts as a podcast download explains why a prefix-level baseline is the only number that stays comparable across all of them.

If the numbers are real and you want a plan

If you have worked through all six causes and the decline holds up, the next question is which part of the funnel broke, and that is worth a conversation, not another dashboard.

Book a free growth strategy call and we will look at your actual numbers with you.