How to start a podcast

You can be published on Apple Podcasts and Spotify inside a week, for under $100, with gear you may already own. That part is genuinely easy now and most guides are right about it. What they are quiet about is the part that decides whether the show exists in six months, which is not the microphone. This is the whole path in order, with real numbers, and an honest account of what each step actually costs you in time.

The short version

1. Pick a topic and format narrow enough that you can do fifty of them.
2. Buy one USB microphone and a pair of wired headphones. Nothing else.
3. Record in the smallest soft room you have.
4. Edit lightly. Cut mistakes, level the audio, leave the rest.
5. Make square cover art between 1400 and 3000 pixels.
6. Write show notes and a title that make sense out of context.
7. Pick a host, upload the episode, get your RSS feed URL.
8. Submit that one feed URL to Apple and Spotify.
9. Publish on a schedule you can hold for six months.
10. Grow it with clips, guest appearances, and consistency, in that order.

Steps one, nine, and ten are the hard ones. Everything else is a checklist. The rest of this page is the detail on each.

Step 1: pick a topic and format you can sustain

Most podcasts stop before episode ten. The usual cause is not audience or audio quality, it is that the show was defined so broadly the host had to reinvent it every week, or so narrowly they ran out of things to say by episode six.

A useful test before you record anything: write down fifty episode titles. Not ideas, titles, the actual sentence that would appear in a feed. If you stall at twelve, the topic is too narrow or you have not found its angle yet. If you can write two hundred without effort, it is too broad, which sounds like a nice problem and is not: a show about “business” competes with everything, and a show about “how small agencies price their work” competes with almost nothing and gets recommended by name.

Then pick a format, and understand what each one costs you every week:

• Solo. Cheapest to schedule, hardest to sustain. No guest to carry a flat day, and it requires you to be interesting on demand. Great for teaching a subject you already know cold. Episodes tend to run 15 to 30 minutes, and the prep is where the work lives.
• Interview. Easiest to fill an hour, most operational overhead. Every episode needs a booking, a reschedule, a reminder, and a tech check. Budget roughly an hour of admin per guest on top of the recording. The upside is real: guests bring their own audiences and give you a reason to reach out to people you want to know.
• Co-hosted conversation. The most durable format, because you show up for each other on the weeks you would have skipped alone. The risk is scheduling two people forever, and that the show dies the moment one of you loses interest.
• Narrative or documentary. The best listening experience and by far the most work. A twenty minute narrative episode can take twenty or thirty hours to produce. It is a genuine craft and a poor fit for a weekly cadence you are fitting around a job. Do not start here.

Pick a length last, and pick it from your format rather than a rule. There is no ideal podcast length. There is only the length at which your episode stops being good, which for most conversational shows is somewhere between 25 and 60 minutes.

Naming: the name should say the subject or be memorable enough to search for, and it needs to survive being said out loud. If you are stuck, the podcast name generator is free and gives you ten directions with the angle each one takes.

Step 2: the minimum real gear

You need a microphone and headphones. That is the list. Everything else on the gear pages is an upgrade you can make in a year if the show is still running, and buying it now is the most common way people spend $600 to never publish.

• A USB microphone, roughly $60 to $100. A dynamic USB mic in that range is genuinely good enough to publish with, and the gap between it and a $400 setup is smaller than the gap between it and your laptop’s built-in mic. Dynamic rather than condenser matters more than the brand does, because a dynamic mic hears mostly you and a condenser hears mostly your room.
• Wired headphones, anything you own. Wired so there is no Bluetooth latency, and worn so your guest’s voice does not leak into your microphone and make the audio impossible to edit cleanly. Any pair works. This costs nothing and is the most commonly skipped step.

Two upgrades worth it later, in this order: a boom arm or a stand that gets the mic close to your mouth, and a proper audio interface with an XLR mic if you eventually record two people in one room. A mixer, a cloudlifter, a shock mount, and a $300 mic are all real products that will not make your first fifty episodes better.

The one rule that matters more than the gear. Put the microphone close, roughly a hand’s width from your mouth, slightly off to the side so plosives pass by rather than into it. A cheap mic four inches away beats an expensive mic two feet away, every time, and it is free.

Step 3: the room, which matters more than the microphone

The difference between amateur-sounding and professional-sounding audio is mostly reverberation, and reverberation is a property of the room. You cannot fix it in editing. Nothing removes room echo convincingly after the fact.

What actually works, in order of effect per dollar:

• Pick the smallest room with the most soft things. A bedroom with a carpet, a bed, and curtains is a better recording space than any large room with hard floors. A walk-in closet with clothes in it is, unglamorously, excellent.
• Avoid parallel bare walls and avoid corners with nothing in them. If you clap and hear a ring, the microphone hears it too.
• Turn things off. Fridge, air conditioning, fan, dehumidifier, phone on the desk. A constant hum is removable but always leaves the voice slightly processed, so it is cheaper to not record it.
• Hang a duvet or moving blanket behind the microphone if the room is still live. This is the whole of acoustic treatment for a hundredth of the price, and nobody can hear the difference between it and foam panels.

Step 4: recording

Free software is sufficient forever. Audacity is free and cross-platform, GarageBand is free on a Mac, and any of the browser-based remote recording tools will do the job for interviews. You do not need to pay for a digital audio workstation to make a podcast.

Get these five right and the rest is taste:

• Levels. Aim for your loudest moments to peak somewhere around minus 6 on the meter, with the normal conversation sitting well below that. Too quiet is fixable, clipping is not. If the meter hits the top and turns red, the audio is permanently damaged, so record quieter than feels right.
• Record each person to a separate track. This is the single most useful thing you can do for the edit and for transcription. When two people talk over each other on one track it can never be cleanly separated. On two tracks it is trivial.
• Record locally, not just through the call. Remote recording tools that capture each participant’s audio on their own machine and upload it afterwards give you full quality regardless of the connection. Recording the call itself gives you a recording of the internet having a bad day.
• Say the mistake, pause, and redo the sentence. Do not restart the episode. A clean two second gap is easy to find in a waveform and easy to cut.
• Back up the file before you touch it. Copy the raw recording somewhere else the moment you stop. Nearly every podcaster has lost one episode this way, once.

Do a full sixty second test recording and listen back on headphones before every session, including with your guest. Two minutes here has saved more episodes than any other habit on this page.

Step 5: editing, and how little of it you need

New podcasters routinely spend four hours editing a one hour episode, and almost all of that time goes into things listeners do not notice. The goal is a clean listen, not a perfect one. Conversation is supposed to sound like conversation.

Do this much:

• Trim the top and tail. Get to the point. The first fifteen seconds decide whether someone stays, and “so, how’s your week been” is not those fifteen seconds.
• Cut the real mistakes. Restarts, long derailments, the two minutes where somebody’s dog arrived, dead air over about three seconds.
• Level the voices against each other so your guest is not half your volume, then normalize the finished episode to about minus 16 LUFS, which is the loudness target Apple Podcasts expects for stereo. An episode that is much quieter than everything else in someone’s feed reads as amateur before a word is said.
• Add intro and outro if you have them, and keep the intro under fifteen seconds. Long musical intros are a habit inherited from radio and they lose listeners.

Skip this much: removing every “um,” tightening every pause, noise reduction on audio that is already fine, and music beds under speech. Over-edited conversation sounds airless and it takes hours to achieve.

Step 6: cover art

Your artwork is the only visual most listeners will ever see, at about the size of a postage stamp. Apple enforces hard requirements and will reject a feed that misses them, so get these right the first time.

• Square, between 1400 and 3000 pixels per side. This is the single most common rejection reason. A 1200 pixel image fails. A rectangular crop fails.
• JPG or PNG, under 5 MB, RGB colour.
• Readable at 55 pixels. Shrink your design to thumbnail size and look at it. If the show name is not legible, redo it. This usually means three or four words maximum, heavy weight type, and high contrast.
• No episode numbers, no “new episodes weekly,” no podcast microphone. The first two date the artwork, and the third makes you look like the twelve thousand other shows using the same clip art.

Design it yourself in any free design tool, or pay someone $50 to $150. Both are reasonable. What is not reasonable is stretching a photo you like into a square and hoping.

Step 7: titles and show notes

Episode titles do more work than any other text you write. They appear in a directory next to a hundred others, with no context, to someone who has never heard of you. A title that reads “Episode 14: Sarah” tells a stranger nothing. A title that reads “How Sarah priced her agency out of the hourly trap” tells them everything.

Put the subject in the title, keep the episode number at the end or leave it out, and write for the person who has never listened rather than the person who has heard every episode. The episode title generator will give you options from a description if you are staring at a blank field.

Show notes want four things, in this order:

• Two or three sentences on what the episode covers, written to be read cold.
• Links to everything mentioned, because people genuinely come looking for these.
• Timestamps or chapters for the main sections.
• Who the guest is and where to find them.

Chapters are worth ten minutes each episode and almost nobody does them. They show up in Apple Podcasts and on YouTube, they let people skip to what they came for, and they make an hour-long episode feel navigable rather than daunting.

Step 8: a host and an RSS feed

This is the step that confuses people the most, so here is the mechanism plainly. You do not upload episodes to Apple Podcasts or Spotify. You upload them to a podcast host. The host stores the audio files and generates one RSS feed, which is an XML document listing your show and every episode. Apple, Spotify, Overcast, Pocket Casts, and everything else read that one feed. Publish to the feed and every app gets the episode automatically.

What a host actually does for the $12 to $30 a month it charges:

• Serves the audio from a CDN with range request support, so scrubbing works.
• Generates a valid feed with the Apple-specific tags the directories require.
• Counts downloads in a way the industry agrees on (the IAB standard).
• Keeps the feed URL stable, which matters more than everything else here.

You can technically host a feed yourself on any static file host. It is a bad trade for most people, because the failure modes (a feed that validates but breaks in one app, an enclosure URL that refuses byte ranges, download counts that are not comparable to anything) all surface after you have subscribers.

The one irreversible decision. Your feed URL is your show’s identity in every directory. If it changes without a permanent redirect from the old address, every subscriber you have is orphaned and they simply stop receiving episodes, silently. Before you pick a host, check that it supports feed redirects on the way out. Whatever you choose, run the URL through the RSS feed validator before you submit it anywhere.

Step 9: submitting to Apple Podcasts and Spotify

You need one published episode, artwork at the right size, a category, and an email address in the feed’s owner field that you can actually read, because both directories mail an ownership code there. Then you paste the same feed URL into Apple Podcasts Connect and Spotify for Creators, once each, and never again.

Spotify is usually live within hours. Apple takes a few days for review, which is normal rather than a sign something went wrong. The exact click path for both, and the four things that actually get feeds rejected, are in the Apple and Spotify submission guide.

Step 10: publishing consistently

Cadence beats volume and consistency beats both. A show that publishes every second Tuesday for a year is in far better shape than one that publishes three times a week for a month and stops.

Practical launch mechanics that measurably help:

• Record three or four episodes before you launch anything. This proves the format works before it is public, and it gives you a buffer for the first week you get sick.
• Launch with three episodes at once. Someone who finds you and likes episode one can immediately listen to two more, which is how casual listeners become subscribers. It also gives the directories more to work with.
• Pick a day and keep it. Weekly if you genuinely can, every two weeks if you cannot. Choose the slower cadence you will actually hold over the faster one you are optimistic about.
• Batch the parts you dislike. Record two episodes in one sitting, or write all your show notes on the same afternoon. The work that kills shows is not the recording, it is the twelve small tasks after it.

Expect the admin around each episode (editing, notes, artwork, scheduling, posting) to take longer than the recording itself, often two to four hours for a one hour episode. Budgeting for that honestly at the start is most of what separates shows that survive from shows that do not.

What actually grows a show early

Discovery inside podcast apps is weak. Nobody browses Apple Podcasts looking for new shows the way they browse Netflix, so almost all early growth comes from somewhere other than the podcast app. Three things work, and they work in roughly this order.

• Clips on social video. Short vertical video is where podcast discovery actually happens now. One good 45 second moment, cropped to 1080 by 1920, with burned-in captions, reaches more new people than any amount of directory optimisation. Three to five clips per episode is a realistic target, and the process for choosing and cutting them is in turning one episode into a week of content.
• Being a guest on other people’s shows. The highest conversion rate of anything available to you, because the audience is already listening to podcasts and has just heard forty minutes of you being worth listening to. Aim for shows about your size rather than the biggest one in your category, because those people answer their email.
• Consistency. Unglamorous and it compounds. Back catalogue is an asset: every episode is a permanent, searchable entry point, and a show with forty episodes gets found in ways a show with four cannot. Most of the growth in year one arrives in the second half of it.

Also worth doing, in a lower tier: ask for ratings and reviews explicitly at the end of an episode (it works, and almost nobody asks), get a transcript published for each episode so the words are searchable, put every episode on YouTube as video or an audiogram, and make sure your show is easy to find by name because word of mouth is spoken.

Not worth doing early: paid ads (you cannot measure podcast conversion well enough to spend on it), buying followers, posting the same clip to six platforms you do not use, chasing chart placement, and redesigning your artwork for the third time.

What the numbers really look like

The published percentile data from the major hosts puts the median episode at roughly 30 downloads in its first week. Around 130 downloads puts an episode in the top quarter of all podcasts, and a few hundred puts it in the top tenth. Those figures move a little year to year, and they are worth knowing for one reason: your first episode will get somewhere between five and fifty downloads, and that is completely normal rather than evidence the show has failed.

A realistic timeline for a weekly show done properly: a few dozen listeners for the first couple of months, mostly people who know you; slow growth through months three to six as the back catalogue and clips accumulate; and any meaningful audience arriving after episode twenty or thirty. Shows that grow faster than that usually brought an audience with them from somewhere else.

Total cost to be live: $0 to $100 for a microphone, $0 to $150 for artwork, and $12 to $30 a month for hosting. Total time for the first episode, including everything on this page: a weekend. Every episode after that: three to five hours until you get quick at it.

A first month that works

1. Week 1. Write the fifty titles, pick the format and the name, order one microphone.
2. Week 2. Record a throwaway test episode you will never publish. Listen to all of it. Fix your room and your mic distance based on what you hear.
3. Week 3. Record episodes one, two, and three. Make the artwork while they are fresh.
4. Week 4. Edit, write notes and titles, upload to a host, validate the feed, submit to Apple and Spotify, and publish all three the day Apple approves you.

Notice that the throwaway episode is not optional. It is the cheapest way to find out that your room rings, your levels are hot, or your format does not hold, and it costs an hour that you would otherwise spend discovering it in public.

Where PodPilot360 fits

Nothing on this page requires us. You can do all of it with a $70 microphone, Audacity, a $15 host, and your own time, and plenty of good shows do exactly that.

What we do is the part after you stop recording, which is where the hours actually go. You upload one audio file and get back loudness-normalized audio at the Apple standard, a word-level transcript with speaker names, three to six vertical clips with word-timed captions, a longform video, titles, descriptions and hashtags, chapters, a drafted blog post and social captions, hosting with an Apple-compliant RSS feed, your own episode page, download analytics, and a release-anchored publishing schedule. That replaces a host, a clipping tool, and a couple of hours a week of metadata work, which is where the pricing comes from on the pricing page.

The honest boundary: we do not record for you, we do not book your guests, and we cannot make the show worth listening to. Steps one through four on this page are yours no matter what you pay for.

Everything else we have written

The guides and free tools referenced above, in one place:

• Getting listed on Apple Podcasts and Spotify, with the four things that get feeds rejected
• Turning one episode into a week of content, the clip selection test and the posting schedule
• Podcast transcription, the options, what accuracy really means, and why word-level timing matters
• Podcast name generator and episode title generator, both free with no account
• RSS feed validator, to check a feed before a directory rejects it
• Opus Clip alternatives and PodPilot360 vs Opus Clip, for when clipping is the specific problem