How long should a podcast episode be?

No study answers "how long should a podcast episode be?" with a number. The two closest count different things, and neither changed a length to see what happened. So length is decided last. Decide what your show should be known for and the material behind it. That picks the format, and the format sets the range. Start from what shows in your format run, cut every stretch a listener wouldn't miss, starting with the open, and judge each new episode against your own past ones.

Narrated by a synthetic version of Lam Hoang's own voice.

Summary

• The two closest studies measure different things, who stays five minutes and who reaches the end. Neither sets a length.
• The 10-minute attention limit and the commute rule rest on less than they sound.
• The order is subject, then material, then format, then minutes. Shows in your format give you the starting range.
• Inside that range, cut by two tests: why would someone listen to this, and what new idea does it bring. Start with the opening.
• Judge length against your own past episodes, not against someone else's number.

Is there a right length?

When the 1,200 people in The Podcast Study: Growth were asked why they had left a podcast, length was not among the top answers. As Podnews reported the findings ↗︎, the top two were that "it started with too much meaningless chat" or that "it was just boring". Inside Radio's write-up ↗︎ adds that three in four want hosts to "dive right into the main topic or at least keep the chit-chat short". Eight in ten, it says, want podcasts to be an hour or less. That caps very long shows, and says nothing about ten minutes against thirty. Its headline, "Listeners Want Podcasts Shorter", goes further than the findings it quotes.

Two limits. These are stated reasons, not measured behavior, and the field dates and question wording are unpublished, so the date to give is when it was first presented, at Podcast Movement in August 2025. And the two firms behind it, Point-To-Point Marketing and Strategic Solutions Research, sell audience development and content analysis ↗︎.

The research on length itself holds two studies, and they point opposite ways.

Spotify's 2021 paper. Spotify's own researchers modeled listener engagement ↗︎ across a few thousand podcasts from publishers Spotify owns or operates. They took each show's most-streamed episode, left out episodes under 10 minutes, and counted streams to December 2020. Engagement meant the share of first-time listeners who stayed at least five minutes. Across every popularity band, the high-engagement shows "tend to be longer on the whole (contrary to advice to keep episodes short)". The authors warn that "care must be taken when generalizing" from Spotify's own publishers. It is a platform's paper, and a correlation.

A 2024 teaching podcast study. One medical series, Neurology Clinical Pearls, ran episodes of 8 to 17 minutes in 2021 and 2022. The study in Medical Science Educator ↗︎ had retention data for 20 of its 26 episodes. Each extra minute went with a drop of about 3.6% in the share of listeners still there at the end. The authors found no attention limit within 17 minutes, and say one series limits how far the result travels.

One counts who stays five minutes. The other counts who finishes. Neither changed a length to see what happened.

The 10-minute attention rule. It comes from lectures. A 2016 review by Neil Bradbury in Advances in Physiology Education ↗︎ found the primary data "do not support the concept of a 10- to 15-min attention limit". Attention varied most with the teacher, not the format. We read the abstract only, and it is about lectures, not podcasts. It still removes one reason given for going short.

The commute rule. In Edison Research's tracking of US weekly podcast listeners ↗︎, published January 2026, the share who listen most often at home rose from 48% in 2019 to 58% in 2020. Edison's current estimate is 63%. That is each person's main place, not how long they listen, and Edison links part of the rise to listening on TVs. Edison sells podcast measurement.

So when someone hands you a length rule, ask what it counted: who stayed a few minutes, or who reached the end.

What sets the length?

Length is not the big decision. I'd start with strategy.

A business now makes content for people and for the AI systems that search on their behalf. Both need to know which subject you are the expert on. That is the first decision, before any format. Then comes the material behind it. Often that is research or data you never had time to share, or what the experts on your team know and never get time to pass on. A podcast gets it out.

None of the evidence around this is about podcasts or length:

• In the 2024 LinkedIn and Edelman survey, as LinkedIn reported it ↗︎, 55% of decision-makers cited strong research and data as a top trait.
• In G2's survey of 1,076 software buyers ↗︎ in March 2026, 51% said they now start research with a chatbot more often than with Google, up from 29% in G2's 2025 report. LinkedIn and G2 both sell to businesses, and both surveys are self-reported.
• In Pew's analysis of 900 US adults' browsing ↗︎ in March 2025, people clicked a regular result on 8% of visits when a summary appeared and 15% when none did. They clicked a link inside the summary on 1%. That is Google only.

Nothing measures what a podcast adds to a business's place in search or in those answers.

If you want search to find you too, Google's guide to people-first content ↗︎ asks whether you write to a word count because Google has a preferred one, and answers "No, we don't." It asks for "original information, reporting, research, or analysis" and first-hand expertise, and it counts audio among a page's main content. Its page on AI features ↗︎ says AI Overviews and AI Mode need no special optimization, then points back to the fundamentals, among them making important content available as text. Neither page mentions podcasts. That a written version of each episode serves this is our reading.

Next comes the show type that carries your material: interviews, or episodes that teach what you know. Then the length follows.

That last link has the most support. In 2019 the Reuters Institute studied daily news podcasts ↗︎ in the UK, the US, Australia, France and Sweden, and interviewed about 30 publishers and platforms. It found three kinds: micro-bulletins of 1 to 5 minutes, round-ups of 6 to 15, and deep dives of 20 or more. Many publishers picked the format that played to what they already did well. The bands are the study's categories, not advice, and the work is about news.

Take true crime documentaries. I think they run long, and that is the nature of the content. The institute's 2026 report on news podcasting ↗︎ calls documentary-style shows "long form, episodic deep dives into a specific topic or themes (crime, economics history)". It describes them as expensive and highly produced, and conversational shows as longer and cheaper to make.

For business shows, the counts disagree by who is counted. One production agency's count of the top 100 business shows on Apple ↗︎ in 2025 puts their average episode at 46 minutes. That is who is at the top, not why they got there. The typical business show is shorter: CastFox's 2026 count ↗︎ puts its median at 29 minutes.

Source: One Stone Creative's report, as relayed ↗︎, CastFox ↗︎, WhichPodcast ↗︎, Edison Research and Ausha ↗︎, Pew Research Center ↗︎, Buzzsprout ↗︎.

Every row describes what exists, not what holds a listener. WhichPodcast's length research ↗︎ advises matching your category's median before experimenting, and Edison and Ausha's top 25 benchmark ↗︎ says "there is no magic number for episode length". Pew's audit of top shows ↗︎ found video shows ran longer, and Buzzsprout's stats ↗︎ cover only shows on its own platform. Short works too, when it is the format: NPR's The Indicator ↗︎ promises its episodes "in 10 minutes or less". So look up shows in your format, and start from their range.

The hours you have shape the format too. Mignon Fogarty launched Grammar Girl in 2006 at "probably more like a minute-and-a-half to two minutes". In a 2021 Podcast Movement interview ↗︎ she put it at "probably more like 12 to 17". Partly, "those are the kinds of shows I liked to listen to". Partly, her earlier show, an interview show with a co-host, took more production work than she could keep up. Her listeners always ask for more, and she adds that she cannot know how many listeners she misses by staying short. One show, her own account.

The episode on starting a podcast, How do you start a podcast for your business?, set a timing test: time one episode before picking a schedule. Length belongs in that same sum. The order holds when a podcast production service makes the show, too: the brief names the subject and the material, and the minutes come last.

Where do you cut?

Inside that range, the content still has to be worth the listener's time.

Broadcast coach Valerie Geller's rule, as journalism.co.uk reported it ↗︎ from a 2015 talk in London: "Never put anything on the air if you cannot answer why someone would be listening to this." Her point was about stories. I'd run her test on every stretch of an episode and cut what fails it.

Hrishikesh Hirway edits his podcast Song Exploder by a clock in his head. In an interview with Tape Op ↗︎ he calls it "some kind of clock that goes on in my head for how long [something can go on] before it starts to feel monotonous, or repetitive". For a song's second verse, he asks what new idea it introduces. The bracket is the magazine's. The question is his for a song. I'd ask it of every stretch of talk.

Then where to start cutting. Anything in the way of listening should be gone as fast as possible. The evidence on openings is consistent, if thin:

• The teaching podcast. Listeners left at about 10% a minute in the first 2.5 minutes, then about 1.8% a minute. The study's authors ↗︎ read the early drop as people realizing the topic was not for them.
• One host's intro. In Spotify's 2020 guide to episode performance ↗︎, a co-host of UnderUnderstood, a member of the guide's own team, showed an intro of about a minute losing around 10% of listeners at the start. A later season's few-second intro dropped "nowhere near as steep". The guide notes new shows can drop more early anyway.
• Spotify's 2021 paper. The higher-engagement shows had less non-speech in their first 10 minutes, and descriptions closer to those minutes ↗︎.
• What listeners say. In The Podcast Study 2026, 55% say they decide on a new podcast within five minutes ↗︎, and 14% within the first minute. Among the study's Superfans, 18% say they give up within the first minute ↗︎. Among its Prime Prospects, 90% named an execution problem ↗︎ that made them quit a podcast, most often too much meaningless talk at the start. All stated, not measured. Crooked Media, NPR and PRX sponsor the study.

Listeners should know what they are getting into. So say plainly what the episode covers, then start on it.

Is it working?

Compare each episode with your own past episodes. That is where the platforms point too:

• Spotify. As of June 2026, Spotify's creator analytics ↗︎ "provide benchmarks against your past episode performance", day by day against the rest of your catalog.
• YouTube. Its audience retention report ↗︎ compares your 10 latest videos of similar length, and separately all YouTube videos of similar length. Its intro measure is the share still watching after 30 seconds.
• Apple. Apple's analytics ↗︎ count an engaged listener as someone who played at least 20 minutes or 40% of an episode, show how much of each episode people played, and compare episodes against your median, average or top episode. It shows nothing until at least five unique listeners played in the period you pick.

The reason to compare with yourself is in Spotify's 2020 guide ↗︎. There, a show that usually ran short put out one episode of around 40 minutes, and it lost listeners steeply after a few minutes. The guide says "longer shows don't necessarily entail lower retention". It offers two readings: the audience did not take to the longer format, or it "might need a few more episodes to get used to it". So if you change length sharply, give it a few episodes before you judge it. A small audience also gives a staircase-shaped curve, so expect rough charts.

No platform sets a target. That your own episodes are the comparison that counts is my reading of what they offer.

What to do this week

1. Play back the first 10 minutes of your last episode with a timer running. Mark every stretch where nobody speaks and every stretch nobody would miss. That is where your cuts start. With no episodes yet, do it with a show in your format.
2. Write down the subject your show should own and the material only you have, before any number of minutes.
3. Find five shows in your format and note their episode lengths. Their range is your starting range.
4. Run two tests on each stretch of your next script or edit: why would someone listen to this, and what new idea does it bring.
5. Open your host's analytics and note how much of each recent episode people played. Set every new episode beside those numbers.

FAQ

Is there an ideal podcast episode length for a business show?

No. No study sets one, and the two closest measure different things. As a starting point, CastFox's 2026 count puts the median business show at 29 minutes, and One Stone Creative's count of Apple's top 100 business shows puts their average at 46.

Do listeners stop listening because episodes are too long?

Not mainly, by their own account. In The Podcast Study: Growth, people put a start full of meaningless chat and a boring show at the top of their reasons for leaving. Those are stated reasons, not measured behavior.

Should every episode be the same length?

No study shows that a steady length keeps listeners. One 2020 Spotify example, where a short show's long episode lost listeners early, is consistent with it but does not prove it. If you change length sharply, give the change a few episodes and judge it against your own past ones.