What a credible show costs to run, and where shows die

Ask three people what a podcast costs and you will get $200, $2,000, and $20,000. All three are honest answers to different questions.

The dollars are the easy part, and they are also the part everyone quotes. The number that ends shows is hours.

The dollar floor is low

One good microphone runs $50 to $100. Hosting runs from free to about $20 a month at the level a weekly show needs. Recording software is free, and cover art is a one-time expense.

So a show can start for a couple of hundred dollars and a small monthly fee. That floor is why podcast directories hold millions of feeds.

The floor tells you almost nothing about what your show will cost. The expensive part arrives every week after the first one.

What buying the work back costs

Freelance editing runs $50 to $200 an episode. A professionally edited hour-long episode averages around $199, on Castos's July 2026 pricing check. Full-service production starts around $400 an episode and climbs past $1,000. That tier adds show notes, clips, publishing, and sometimes guest booking.

Run those against a weekly cadence. Editing alone lands near $10,000 a year. A weekly show at the full-service tier reaches $21,000 to $52,000 a year. That puts a serious podcast in the same budget conversation as a junior hire.

Those quotes are not comparable to each other, which is why the range looks absurd. One vendor is selling you a cleaned audio file. Another is selling you a guest booked, a room run, an episode published, and a set of clips.

The hours nobody quotes

A do-it-yourself hour-long episode consumes four to ten hours. That count covers editing, show notes, artwork, uploading, metadata, and promotion. Practitioner estimates elsewhere land in the same range, clustering nearer six.

Multiply by a weekly cadence. That is 200 to 500 hours a year, or five to twelve working weeks.

Price those hours at whatever your work bills for. At $250 an hour, one episode costs $1,000 to $2,500 in work you did not do. That reframes every quote in the section above.

Guest coordination is the line item that behaves worst. It runs hours per guest, every guest, and a reschedule spends them twice.

Where shows die, and it is not episode three

You have probably met the figure that says 90% of podcasts stop after three episodes. It appears in the business press and in most articles on this subject.

That number describes the full feed universe. Millions of those feeds are tests, spam, and single-episode experiments. Others are private corporate feeds no listener was meant to find.

Measure the shows listeners can actually find and the picture changes. Across 36,900 indexed shows on the major directories, 96% reach ten episodes and 61% reach one hundred. The median publishing span is 2.4 years. Those figures were computed on August 24, 2026.

Both pictures are true, and they answer different questions. Launching a feed and quietly abandoning it is common. Among shows that clear distribution and reach a tenth episode, the attrition is slower and later.

Hold one more figure beside them. Podchaser counted the 153,767 shows that launched in the first half of 2026. By June 30, 41.7% had stopped publishing.

So the danger window is not the launch, which runs on enthusiasm. It sits somewhere between month nine and year two, which is when a working owner's calendar fills up again.

The four ways a show ends

Shows rarely announce their ending. They stop for one of four reasons, and three of them are decisions made before episode one.

The cadence was set for a good month. Weekly sounds serious. Every other week sounds unremarkable to listeners, and it survives a bad quarter. That makes it the better default for a working expert.

One person does everything. When the host also books, edits, writes, and publishes, any busy fortnight stops the show. A show with a single point of failure has a short expected life, whatever its quality.

Nothing sits behind the episodes. A show without a page you publish around each episode accumulates nothing a stranger can find. The work stops compounding, and the owner feels that before they can name it.

The chosen number never moves. Downloads on a new show stay flat for months. An owner watching only that number loses faith around episode twenty, even while the show is doing its job. Match the measure to the purpose. A show run for its guests is counted in conversations, introductions, and referrals.

What to buy first, if you buy anything

Buy the edit before anything else. It is the largest recurring block of hours and the one with the least of you in it.

Publishing and show notes come second. They take an hour a week, and they get skipped when the week is bad. Skipping them is what strands the episode.

Guest research and scheduling come third. They matter more than they look, because coordination is where the cancellations live.

Keep two things for yourself. The recording, and the choice of who sits in the chair. Those two are the show.

The whole production can also sit inside an arrangement you already have. Someone who already produces your content each month can absorb it. The work does not change, and the question is whose calendar it lands on.

What money will not buy

Money buys polish, consistency, and hours back. It stops there.

It does not buy an audience. A $2,000 episode and a $200 episode reach the same number of people in month one. The difference between them shows up over years.

It does not buy discovery. Podcast apps favor established shows. A production budget will not move you up a chart that barely functions as one.

And it does not buy the thing that keeps a show alive. That is an owner who still wants to record in month fourteen.

Price the year before you buy the microphone. Fifty-two recordings, the hours around them, and one bad quarter in the middle.

Then ask which of those hours only you can supply. Those are the ones to keep. The rest are for sale, and they cost less than the show being over.