Work out which search you're losing
Run three searches before you change anything. Type your exact show title into Apple Podcasts, then into Spotify, then type your last guest's name into Google. Those are three separate indexes that share nothing with each other, and the answer from each one points at a different guide underneath this page.
If the show is missing from an app entirely, that's a distribution problem. It has nothing to do with ranking. On Spotify it usually means the show was never claimed, or the 8-digit verification code went to an address published in your feed that nobody reads. A show that isn't in the catalog can't be ranked inside it, so nothing you type into a title field matters until that's fixed.
Present but sitting below bigger shows? That's placement, and placement is metadata and categories. Absent from Google when somebody searches your guest is the third problem, and it's the big one, because it isn't a podcast problem at all. Your episode isn't a web page. A search engine or an assistant has nothing to return.
Do the free platform afternoon first
Start with the first category in your RSS feed, because it takes five minutes and it decides which chart you're eligible for at all. Spotify reads the first category listed in the feed. Open your show settings at your host and move the category you actually want into first position, then load the feed URL in a browser and read the order there, not in the form you just typed into.
Then the titles. Apple names its own ranking inputs, and the only one you type is the show name, the channel name and the episode title. Go through the back catalog, kill the duplicate episode titles first, cut the emojis, and get the episode numbers and pipes out of the front of every title. Apple picks up a feed edit within 24 hours, so by tomorrow you'll know whether it left your host.
Then the one thing Apple actually lets you submit. Editorial featuring takes a pitch and rank doesn't. Send the request at least two weeks ahead of the date you want, ask for a Monday, and file a separate one for every country you care about. None of this has a fee, and nobody should invoice you for editing four fields and a category.
Pick one number, and know what counts it
Only your host counts downloads. Apple reports Plays, Listeners and Engaged Listeners, Spotify reports Plays, and those three numbers will never reconcile, so an afternoon spent making them agree is an afternoon gone.
Know what a download actually is before you chase it. It's a request that pulled the header plus enough audio to play for one minute, deduplicated on IP address and user agent inside a 24-hour window. Pull your own episode ten times from one browser on one connection and it counts once. The number moves when more separate devices get a minute into the audio, and that's the whole reason the open-web half of this page exists.
A blank analytics panel is usually a floor, not a fault. Apple holds the numbers back until at least five unique listeners have played something inside the window you selected, so a new show or a quiet week shows you nothing, and it reads like a bug.
Then build the pages, because that's the other half
One episode, one page, one stable URL, on a domain you own. Say the address you'd send somebody is on podcasts.apple.com, on open.spotify.com, or on your host's share domain, then you don't have an episode page. You have a listing on somebody else's site, and a search engine or an assistant has nothing to read there.
The transcript goes in the page's HTML, rendered from the server. Google's list of indexable file types carries no audio format at all, and the transcript is the only thing that turns a conversation into something a crawler can hold. A linked VTT file won't do it, and neither will a PDF or a panel that loads when somebody clicks it, so request the page with a plain HTTP GET and search the response for a sentence from the middle of the episode.
The transcripts Apple and Spotify already made for you don't count. They live inside those apps and never reach the open web. Fetch either company's public page for your show and the text isn't in it. Whatever machine transcript you start from has to be read and corrected before you publish it, because speech recognition mangles exactly the words you'd want quoted back: guest names, company names and figures.
A web address is not a website
Neither store requires a site you own. Apple and Spotify want a working web address inside the show's link tag in your feed, and the page your host already publishes satisfies that as completely as a domain you paid for. Open your feed, read the address sitting in that tag, and open it. If it loads a page a person can read, you're compliant. There's nothing to do here today.
If it loads raw XML, a parked domain or nothing, that's the one platform fix worth making this week. Make it in your host's show settings, not by hand-editing a feed your host regenerates, which means the decision about a real site is a search decision and nothing else. You want the show findable outside the two apps, and you want pages you can edit instead of a directory listing you can't.
Don't pay anybody for podcast schema. Google's developer page for podcast structured data returns a 404 and Podcast doesn't appear in the current gallery of supported types. There's no podcast-specific rich result left to earn, though ordinary markup describing your organization and your pages still does its job.
Write for the guest, the subject and the question
Your show title is the one query you're already winning, and hardly anybody types it. People type a guest's name, the company they run, the thing they're known for, and the exact question you spent an hour answering. Those are the pages worth building, and every one of them is already sitting in an episode you've published.
A wall of raw transcript clears the indexing bar and still isn't worth citing. Give the page the layer a person needs: the episode title, a real summary, guest names and what they actually do, timestamps, topic headings, and links to whatever got discussed. One kind of page contains words. The other answers something.
Then check that the answering crawlers are allowed in, one named agent at a time. That's OAI-SearchBot for ChatGPT search, Claude-SearchBot and Claude-User, PerplexityBot, and Googlebot for Google's AI features. Blocking a training crawler is not the same as blocking a search crawler, and a blanket AI block added in a hurry usually cuts off the surface you wanted to be on. Read the CDN and the firewall rules too, more than the file.
What none of this moves
Rank isn't for sale in either app. Apple states that rankings cannot be purchased or sold, and that ratings and reviews aren't factored into Search, so a review drive persuades the human reading your show page and does nothing to your position. Spotify publishes no ranking control for creators and says results are often impacted by a show's play count, so anyone quoting you a price for a chart position is selling something the platform says doesn't exist.
The two halves don't feed each other the way people hope. A YouTube play never touches your host's server, so it never becomes a download. A page on the open web doesn't move your Apple chart position or your Spotify search order, and nothing you edit inside a listening app puts your episode in front of the person searching your guest's name. Each surface pays out on its own terms, and you have to work them separately.
And nobody publishes a time to first citation. Google's bar for AI Overviews and AI Mode is that the page is indexed and snippet-eligible with no additional technical requirements, and it also says indexing is never guaranteed. So a promised date, a citation count or a ranking inside an assistant is invented. What the work does is stop the honest answer to why you're invisible being that there was nothing there to read.