How Long Should a Podcast Episode Be? Format, Listening Context and File Size

Every new podcaster asks the same question and gets the same useless answer. “How long should my episodes be?” — “As long as they need to be.” That’s true and unhelpful. There is a defensible answer, but it isn’t a single number; it’s the overlap between two things you can actually reason about: what your format naturally runs, and where your audience listens. A daily news brief and a two-host comedy show have no business being the same length, and neither should be stretched to fill a slot or padded to hit a round number. The Podcast Episode Length Optimiser computes that overlap for you and estimates the resulting file size — here’s the reasoning underneath it.

Format sets the natural range

Different podcast formats have settled into ranges that listeners have come to expect, and fighting those expectations usually loses. A daily news brief lives or dies on being short — five to twelve minutes, something you finish before you’ve parked. An interview show has room to breathe: thirty-five to sixty-five minutes is the comfortable band, long enough to get past the guest’s press-tour talking points and into something real. Educational content sits in the middle, roughly eighteen to thirty-five minutes — long enough to teach one idea properly, short enough that it doesn’t become homework. Comedy and chat shows run longest, forty-five to ninety minutes, because the appeal is hang-out time with people you like, and cutting it short defeats the point. Documentary and narrative work lands around twenty-five to fifty-five, paced by the story rather than the clock. These aren’t rules, but they’re where audiences and successful shows have converged, and drifting far outside your format’s range is a decision you should make on purpose, not by accident.

Listening context is the second signal

The other half of the answer is physical: when is the earbud actually in? Listening data consistently shows people consume podcasts in the gaps of their day, and those gaps have characteristic lengths. A commute averages around twenty to forty minutes each way, so an episode that fits one leg of a drive gets finished — and a finished episode is worth far more than a half-listened one. A workout runs a bit longer, thirty to forty-five minutes, and listeners want something that carries them through the session without a mid-set cliffhanger. A deep listen — the person who puts on a long-form show for a road trip, a chore-filled afternoon, or a flight — will happily go sixty to a hundred and twenty minutes. The best episode length is the one where your format’s natural range and your audience’s available window overlap. If your interview show’s natural sixty minutes collides with a commuter audience that has forty, that tension is real information: either tighten the edit or accept that most listens will span two trips.

Ad load quietly eats your content time

Here’s the part that trips up planners: the length listeners experience is not the length of your actual content. Ad slots take real minutes off the top and the middle. A light load — one pre-roll, one mid-roll — costs around a minute and a half. A medium load with a second mid-roll pushes toward three minutes. A heavy load — multiple pre-rolls, three mid-rolls, a post-roll — can eat six minutes or more. That matters because if you’re aiming for a thirty-minute episode and running a heavy ad load, your genuine content is closer to twenty-four minutes. Plan the runtime as content-plus-ads, not content-equals-runtime, or you’ll either come up thin on substance or blow past your target window once the ads are stitched in. Dynamic ad insertion makes this worse over time, because a back-catalog episode can accumulate more ad slots months after you published it.

File size: the math that decides your hosting bill

Once you know the duration, file size is simple and worth estimating up front, because it drives your hosting bandwidth costs and your listeners’ data usage. The formula is just bitrate × duration. Audio bitrate is in kilobits per second; divide by 8 to get kilobytes per second, multiply by the number of seconds, and you have the size. A standard 128 kbps stereo MP3 works out to roughly 16 kilobytes per second, so a sixty-minute episode is about 16 × 3600 ≈ 57,600 KB — around 56 MB. Drop to 96 kbps mono (plenty for pure speech) and the same hour is about 42 MB; a talky show has no reason to encode at music bitrates. Push up to 192 or 256 kbps for a music-heavy show and the file grows proportionally. The relationship is perfectly linear: double the bitrate or double the length, double the size. Most podcast directories cap individual files around 500 MB, which you’d only approach with a very long episode at a high bitrate, but it’s worth knowing the ceiling exists.

Bitrate is a quality-versus-size trade, not a free lunch

The temptation is to crank the bitrate “to be safe,” but for spoken-word content it’s mostly wasted bytes — speech simply doesn’t carry the frequency detail that benefits from 256 kbps. A sensible default is 96–128 kbps for talk, reserving the higher tiers for shows where music and sound design are central. Every extra kbps multiplies across every download of every episode forever, so on a show with real audience numbers, a right-sized bitrate is a meaningful, recurring saving on bandwidth. If you also publish a video version, the same file-size logic applies but at far larger numbers — the Video Bitrate & File Size Calculator handles that side, and the Audio Sample Rate Calculator covers the encoding parameters upstream of the final export.

Let the overlap pick the number

Rather than agonizing over a target, pick your format, your primary listening context, and your ad load, choose an encoding bitrate, and let the Podcast Episode Length Optimiser return the recommended range, the real content time left after ads, and the estimated file size across common durations. It won’t tell you what to say — but it will keep your runtime honest, your content time realistic, and your hosting bill predictable, which is most of what “how long should it be” is really asking.