Podcast SEO: The Definitive UK Guide to Getting Your Show Found

Most podcasts are never discovered. They are shared one listener at a time by hosts posting clips into the void and hoping. Podcast SEO is the alternative. It is the practice of making your show findable everywhere people actually search for podcasts: Apple Podcasts, Spotify, YouTube, Google and, increasingly, AI assistants like ChatGPT and Perplexity. Done well, it turns your back catalogue into a compounding source of new listeners. Ignored, it leaves your best work sitting in a feed that only your existing audience will ever open.

This is the definitive UK guide. It covers how search actually works on each platform, what to do about it in practical terms, and the angle almost nobody has written about properly: how large language models decide which podcasts to recommend. Discoverability is built into every show we make as part of our podcast production services, and what follows is the playbook we use.

What Podcast SEO Actually Is

Podcast SEO is the work of making a show findable wherever people search for one. That used to mean Apple's search box and not much else. It now means five surfaces, each with its own rules:

• In-app search on Apple Podcasts and Spotify, where listeners type a topic or a name and pick from a short list of results.
• YouTube search and recommendations, which now matter more for podcast discovery than any audio directory.
• Google web search, which ranks your website, episode pages and transcripts, not your RSS feed.
• AI assistants like ChatGPT, Claude, Gemini and Perplexity, which answer "what should I listen to about X" with a shortlist you are either on or not.
• In-app recommendation systems, the algorithmic shelves and "you might also like" rows, which feed on the same metadata and engagement signals as everything above.

One principle sits underneath all of it: machines cannot listen. Apple's search engine has never heard your episode on procurement. Google has not enjoyed your host's interviewing style. ChatGPT does not know your guest was brilliant in minute 40. Every one of these systems reads text: titles, descriptions, transcripts, captions, web pages and whatever other people have written about you. Podcast SEO is the discipline of turning what was said into text that machines can find, in the places they actually look.

It is also not the same thing as making a good show. A good show retains listeners. Discoverability gets them there in the first place. You need both, but only one of them is usually missing.

Why It Matters More Now Than Two Years Ago

Three shifts changed the game, and all of them favour shows that take discoverability seriously.

Google Podcasts is dead. Google shut it down in 2024 and moved podcasting into YouTube. That was not a tidy product decision, it was a signal. The web's biggest gatekeeper decided podcasts are a YouTube format now. If your show has no video presence, it is invisible on the platform Google chose as podcasting's home.

The apps transcribe everything. Apple Podcasts began auto-generating transcripts in 2024, and Spotify offers them too. Your spoken words are now searchable text inside the apps whether you planned for it or not. Shows that talk clearly about specific topics get found for them. Shows that meander get found for nothing.

AI assistants became a recommendation layer. A growing share of "what should I listen to" questions are answered by ChatGPT, Claude, Gemini or Perplexity rather than a search results page. Those answers are short. Three to five shows, delivered with total confidence. You are either on the list or you do not exist, and the factors that put you there are text, structure and reputation. More on this below, because it is the section most guides skip entirely.

There is a quieter reason too, and it matters most for B2B shows. Your target audience is not "podcast listeners". It is a few thousand specific people with job titles and budgets. Charts will not reach them, because charts reward volume. Search rewards relevance, and relevance is a game a niche show can actually win.

Episode Titles and How Apple and Spotify Search Work

Titles are the highest-leverage metadata you control, and most shows waste them on episode numbers and in-jokes.

How Apple Podcasts search works

Apple's search leans on a small set of metadata fields, with the show title and author field carrying the most weight, followed by episode titles. Episode descriptions do relatively little for Apple search, which surprises people who have spent hours polishing them. (They matter elsewhere. Hold on.)

Two practical consequences. First, your show title should say what the show is about, or at least not fight it. "The CFO Playbook" works because the audience is in the name. Second, do not stuff. Apple's guidelines explicitly warn against keyword-stuffed show names, and titles like "Marketing | Growth | SaaS | B2B | Leads Podcast" risk rejection rather than rankings.

The author field is quietly useful. It is searchable, so it should carry your company or host name exactly as people search for it, not an internal brand name nobody types.

How Spotify search works

Spotify matches queries against show names, episode titles and descriptions, and it has been steadily moving towards semantic search: matching what a query means rather than the exact string typed. Descriptions pull real weight here, which is one reason identical boilerplate under every episode is such a waste. Spotify also personalises results based on listening history, so two people searching the same phrase see different orders. You cannot control that. You can control whether your metadata gives the algorithm something to match.

Writing episode titles that get found

• Lead with the topic, not the number. "Ep 47:" tells a search engine nothing and burns the characters most likely to be read. Episode numbers belong in the metadata field built for them, or nowhere.
• Write what a listener would type. "How to build a finance team from your first hire" will be found. "A wonderful chat with Sarah" will not, even if the chat was wonderful.
• Include the guest's name and company. People search for people. A guest with an audience is a search term with an audience attached.
• Front-load the meaning. Apps truncate titles on mobile, so the first few words do the work. "Pricing strategy for SaaS, with Jane Smith (Acme)" survives truncation. "In this week's episode we sat down with..." does not.
• One idea per title. If the episode covers five things, title it for the strongest one and put the rest in chapters.
• Drop your show name from episode titles. It is already attached to every episode. Repeating it crowds out words that could earn a search impression.

Show Notes and Descriptions That Earn Their Keep

There are two descriptions, and they do different jobs.

Your show description is the shop window. Its first sentence should say who the show is for and what they get from listening, in the words that audience would use. It is indexed by Spotify, displayed by every app, and read by every AI assistant that ever summarises your show. If it opens with "Join hosts Mike and Dave on a journey", it is doing none of those jobs.

Your episode descriptions have a harder brief, because most apps show only the first line or two before a "more" tap. So:

• Open with a plain-English summary: the guest's name, their company, and the specific questions the episode answers. Two sentences. No teasing.
• Follow with chapters or timestamps. They help humans skim and give platforms a topical map of the episode.
• Then links: the guest's profile, resources mentioned, your show page.

Three habits to break. Identical boilerplate pasted under every episode, which turns your most indexable field into dead weight. Sponsor links above the summary, which push the useful text below the fold. And the mystery tease ("you won't believe what happens at minute 34"). They will believe it. They just will never find the episode.

Transcripts: The Highest-Leverage Text You Are Not Publishing

The apps now transcribe your show automatically, which is excellent for accessibility and in-app search, and precisely no help to your website. In-app transcripts live inside the platform. They do not rank on Google, they do not build your domain's authority, and they cannot be cited by an AI assistant browsing the open web.

So publish your own. An edited transcript on your own episode page turns a 45-minute conversation into a few thousand words of indexable, quotable text. That is the difference between an episode existing as a row in an app and existing as a document the entire text-reading internet can find. It catches long-tail searches no title could ever target: the specific phrase your guest used, the framework they named, the objection they answered in passing.

Edited is the key word:

• Fix names, companies and product terms. Speech-to-text mangles all three, and a transcript that misspells your guest is worse than no transcript.
• Break it into sections with descriptive headings that mirror your chapters.
• Pull the two or three best quotes out as emphasis. Skimmers and machines both love a highlight.
• Cut the "so, yeah, I mean" scaffolding. Nobody links to a wall of filler.

The transcript is also the raw material for everything else: clips, LinkedIn posts, articles, newsletters. It is the cheapest asset in your entire pipeline and the one most shows throw away.

YouTube Is a Podcast Search Engine. Treat It Like One

When Google closed Google Podcasts, it moved podcasting into YouTube and YouTube Music. That makes YouTube the place where Google-scale search meets podcast content, and it behaves nothing like an audio directory. Nobody browses YouTube by category. People search questions and get fed recommendations, and both systems run on signals you control.

What a video podcast gets that an audio feed cannot:

• Search real estate. Full episodes, clips and Shorts can all rank separately. One recording becomes a dozen chances to appear for a dozen different queries.
• Captions as text. YouTube indexes captions, so the words spoken in your episode count towards what it ranks for. Speaking clearly about specific topics is, once again, a search strategy.
• Google's front page. YouTube videos appear in Google's main results. Your episode about procurement can sit on page one for a procurement query while an audio-only feed sits in the dark.

Treat the platform on its own terms:

• Titles follow YouTube norms: clear, specific, a touch of curiosity, no episode numbers.
• Custom thumbnails with faces and a few large words. A screenshot of two people in headphones stops nobody's scroll.
• Chapters on every episode. They help viewers skip, help YouTube understand each segment, and can surface directly in search results.
• Descriptions written as descriptions, not link dumps.
• Clips titled as standalone searches, never "Clip from Ep 12".

This is why Earworm is video-first. Every recording we make becomes YouTube episodes, LinkedIn clips and the audio feed, because producing audio-only in 2026 means opting out of podcasting's biggest search engine. It is the foundation of how we approach B2B podcast production.

Your Website: Where Podcast SEO Compounds

Everything so far happens on rented land. Apple and Spotify decide what surfaces. YouTube decides what gets recommended. Your website is the one surface you fully control, and it happens to be the primary thing both Google and AI assistants actually read. This is where podcast SEO stops being a series of tweaks and starts compounding.

The show page

Every show needs one canonical home: a single URL with the show name in the page title, a plain description of who it is for, an embedded player, subscribe links for every platform, and links to every episode. This page is what ranks when someone searches your show's name, and it is what an assistant cites when it recommends you. If your show's only home is an Apple listing, you have outsourced your own front door.

Episode pages

One URL per episode. Each carries the episode title as its heading, the player, proper show notes, the full edited transcript, and links to related episodes. Over a year that is fifty-odd substantial pages on your domain, each targeting queries your buyers actually type. No other marketing channel quietly builds an asset like that as a by-product of showing up.

Structured data

Add PodcastSeries and PodcastEpisode schema markup to those pages. Structured data tells machines exactly what they are looking at (this is a show, this is an episode of it, this person hosts it, this person was the guest) instead of leaving them to infer it. Google uses it, and AI retrieval systems benefit from it. It is an hour of developer time for a permanent gain in clarity.

Articles around the show

The compounding step: turn strong episodes into written guides that target the search query behind the conversation, with the episode embedded. An episode on hiring becomes an article on hiring that ranks, and the article feeds listeners back to the show. This loop is the engine behind our content playbook: the podcast is the source, the website is the distribution asset.

Done consistently, this is what separates shows that grow from shows that merely persist. Soldo's The CFO Playbook is a useful example of a show built around a precisely defined audience from day one. When the audience definition is that clear, every episode page knows exactly which searches it exists to answer.

How ChatGPT, Claude and Perplexity Recommend Podcasts

Ask ChatGPT for the best podcasts for UK finance leaders and you get a confident shortlist in seconds. Ask Perplexity and you get one with citations. These answers are already shaping what people subscribe to, and almost nothing useful has been written about how to earn a place in them. So here is what actually determines it.

An LLM has never heard your show

Language models do not listen to audio. A model's entire knowledge of your podcast is text: your show description, your episode titles, your transcripts if they are on the open web, and everything anyone else has ever written about you. If your show exists only as audio inside apps, then as far as an LLM is concerned you barely exist at all. Your text surface area is your AI discoverability. That sentence deserves a sticky note above every producer's desk.

Where the recommendations come from

Assistant answers draw on two pools, and you can influence both.

Training data. Models learn which shows exist and what they are about from the text they were trained on: press coverage, "best podcasts for X" roundups, directories, reviews, Reddit threads, newsletters. Shows that are frequently and consistently described in text become strong entities the model can recall and recommend. Shows that were never written about are simply not in the room when the question gets asked.

Live retrieval. ChatGPT's search mode, Perplexity and Claude with web access fetch current pages at answer time and summarise what they find. In practice they lean heavily on pages that already rank for the query, which for podcast recommendations usually means listicles and roundups. Classic SEO does not disappear in the AI era. It becomes the supply chain for AI answers. If the pages an assistant retrieves all mention your show, so does the assistant.

What actually moves the needle

• Publish transcripts and episode pages. Retrieval needs something to retrieve. A show with fifty substantive, crawlable pages gives an assistant fifty reasons to mention it.
• Keep your entity consistent. Same show name, same host name, same one-line description everywhere: your site, the directories, guest appearances, social profiles. Models connect mentions by matching names. Casual rebrands and inconsistent descriptions fragment you into several weak entities instead of one strong one.
• Get written about. Pitching your show into credible roundups used to be a nice PR win. It is now the most direct route into both training data and retrieved answers. Guest appearances on other shows do double duty, since each one generates a page describing who you are and what you know. This is a large part of why guest booking is a growth channel rather than an admin chore.
• Use structured data. Schema markup will not make a model love you, but it makes every crawl of your site unambiguous, and retrieval systems reward unambiguous.
• Write answer-shaped pages. Assistants assemble answers from pages that already look like answers. A page structured around "what is the best way to X" is far easier to quote than a fragment of raw transcript.

An honest caveat: nobody outside the AI labs knows the exact weightings, and anyone selling you a guaranteed ChatGPT ranking is selling weather. But the direction of travel is unmistakable. Text-rich, consistently named, frequently cited shows get recommended. Audio-only shows with a two-line description do not.

The Podcast SEO Checklist

Print this, or at least stop nodding and start doing it.

Every episode

• Title leads with the topic, includes the guest's name and company, and survives truncation.
• First two lines of the description summarise the episode in plain English.
• Chapters or timestamps included.
• Edited transcript published on its own episode page, broken up with headings.
• YouTube version live, with a custom thumbnail, chapters and a real description.
• Two or three clips published, each titled as a standalone search.
• Episode page links to related episodes and to your show page.

Once, at show level

• Show title and author fields match what people actually search, without stuffing.
• Show description leads with who it is for and what they get.
• Canonical show page live on your own domain, with subscribe links for every platform.
• PodcastSeries and PodcastEpisode schema in place.
• Show name, host name and description identical across every directory and profile.
• Listed on Apple Podcasts, Spotify, YouTube Music and the smaller directories.

Every quarter

• Review Search Console: which queries reach your episode pages, and which pages deserve a refresh.
• Ask ChatGPT and Perplexity for recommendations in your niche. Note who appears and which sources get cited.
• Pitch one or two credible roundups or publications.
• Update your best-performing episode pages with new internal links and sharper summaries.

Make Your Show Impossible to Miss

Every show Earworm produces gets this treatment as standard: video-first recording, edited transcripts, episode pages, clips and multi-platform distribution, from £1,500 a month. Explore our podcast production services to see how the pieces fit together, or book a call and we will look at your show's discoverability with you.