Should you start a podcast, or keep guesting on other people's shows?

You come off an interview on someone else's show and it goes well. The host is pleased, the questions were good, and a listener writes to you that week. Then the thought arrives. You should have your own show.

Hold it for a minute, because guesting and hosting are two separate decisions. They look like one channel from outside. They cost different amounts and they pay in different ways. A third question sits underneath both: whether audio is the right long game for your buyers.

Guesting puts you in front of an audience that already exists. Hosting builds an audience of your own, slowly, and asks for a standing commitment while you wait.

For most expert practices the order runs guesting first, hosting second. The show is the second thing you buy. The first one has more room left in it than most owners think.

Who is paid by which answer

Before you read further advice on this question, look at who wrote it. Production companies sell shows, so their answer is a show. Booking agencies sell guest placements, so their answer is guesting. Both do good work and both make a case worth reading, and both have a side.

The people with no side are the owners who have run each one for a year. Their reports are quieter, and easier to miss.

What a seat on someone else's show buys

When a host books you, she lends you something she spent years building. Her listeners extend you the trust they already give her. You did not earn it, and for forty minutes you get to use it.

Speed is the other half. You can pitch a show this month and record next month. There is no feed to set up, no cover art, no schedule to keep, and no editor to pay.

Getting booked has a craft of its own. A seat on the right show is worth more than a seat on a larger one.

What your own show buys, and what it charges

Your own show gives you three things a guest slot cannot. You choose every guest. You keep the archive on a domain you control. And you set the questions, which means you decide what subject your name sits next to.

The cost comes due weekly. The show needs the next episode whether or not the month went well. I produced more than 200 episodes of the Deliberate Leaders Podcast. The show did not care whether the week was convenient.

That recurring appointment is what ends most shows. The hours a weekly show takes can be counted before you commit to a single one.

A new show starts at zero, and most stay close to it

One number settles much of this argument. The median podcast episode gets 27 downloads in its first seven days. That is Buzzsprout's platform figure for July 2026, drawn from every show it hosts.

Picture that audience. It would fit in a small meeting room, and half the episodes published draw fewer than that.

The median includes hobby shows and feeds abandoned after episode four, so your show would likely beat it. It would not beat it in month one. Month one is when most owners privately decide whether the show was a good idea.

Guesting hands you a room with people already sitting in it. Hosting asks you to build the room first, then fill it.

The first test: has your material worked in someone else's room?

Publishing your own show makes little sense until the material has moved strangers. Guesting tells you that cheaply, and it tells you fast.

After four or five interviews, look for movement you can name. A host who sent you a referral. A listener who wrote and mentioned the episode. A guest appearance that turned into a conversation about work.

If the interviews produced polite silence, your own show repeats the same material in front of fewer people. The material is the thing to fix first, and another channel will not fix it.

The second test: can your calendar hold a year of it?

Count a year of episodes, and count it against your worst month rather than a quiet one. Shows are rarely killed by the launch. They die in the months after, when a client crisis lands on the day you were going to record.

Owners usually estimate the recording and forget the rest. Booking guests, chasing reschedules, prepping questions, editing, writing notes, and publishing all sit inside the weekly number.

Guest coordination is the piece that refuses to compress. You can shorten an edit. You cannot shorten a guest who cancels twice.

Most of that weekly number can be bought back. A producer, or a monthly engagement that already runs your content, can take the booking, the editing, the notes, and the publishing. What stays with you is the recording, and the recording is the part only you can do.

What a show produces now that it did not produce in 2019

Search engines and AI assistants read text. They do not sit and listen to your audio.

So the durable asset from an episode is the page you publish around it. That page holds the transcript, the notes, and the written piece the conversation earned. It can be found, quoted, and cited months later. The audio file on its own mostly cannot.

This changes who should host. If you will publish the page, the show becomes an engine that feeds your site every week. The material on it belongs to nobody else. If you will not, the show produces good conversations and leaves nothing a stranger can find.

The one reason to start before both tests pass

There is one exception. A show built around the guest rather than the audience can pay from the first episode. An invitation to be interviewed opens doors that outreach does not.

Run that way, audience size stops being the point, and both tests matter much less.

What a show will not do

A show will not create demand for what you sell. It reaches people who are already curious about the subject and gives them a reason to trust you sooner.

It will not make you findable by itself. Podcast apps are poor at discovery, and most listeners arrive from a link somebody sent them.

It will not repair a presence that does not convert. Someone who hears you looks you up next, and your site does the deciding from there.

And it will not tell you when it is failing. Downloads drift, the gap between episodes stretches, and the show ends quietly without anyone making a decision.

Guesting and hosting both work. They work in an order.

Use the borrowed audience until it has shown you what your material can do in front of strangers. Then build the show, once two things are true. Your calendar can hold a year of episodes. The interviews have produced something you can point to.

Many owners run that order backward. The cost of the wrong order is a year spent talking to two dozen people.