Industries · Media & Entertainment

Platforms win “where do listen.” Your page wins “what did he say.”

A show, a channel, a catalogue — the audio and video live on platforms you cannot edit, generated from a feed. Which raises a fair question about what your own site is even for. It is for the one thing those platforms structurally cannot do: hold the words. An assistant asked what somebody said in an interview cannot quote audio. It quotes a transcript, and it cites the page that transcript sits on, by the exact title and chapter heading you wrote for it.

The failure we keep finding: the show’s own name, nowhere as text

The most common defect in this industry is invisible to everyone who works on the site, because to a person the page looks finished. The show’s name is there. The host’s name is there. They are just not text. They are inside logo files and artwork, with empty alt attributes, under a title tag that reads like a navigation label rather than the name of anything.

The result is a page that a search engine cannot connect to the show it is about. Search the host’s name alongside the word podcast and you get other people’s shows — the ones they guested on, whose pages did write their name out. The guest appearance outranks the home show, which is a strange thing to explain to a client and an easy thing to fix.

The same audit usually turns up the rest of it at once: no sitemap, an empty robots file, no canonicals, no share image, no structured data anywhere, and a hero image several megabytes large marked to load lazily even though it is the first thing on the screen. None of that is exotic. It is what happens when a site is designed and then shipped, without anyone asking what a machine reading it would come away with.

Anatomy of an episode page

This is the artifact the whole engagement turns on, so it is worth being specific about what is on it. The guest’s name first in the title and the show’s name after it, kept short enough to survive a results page. A description written from the hooks the conversation actually contains rather than from a template. A summary. Five to seven takeaways someone could read instead of listening. Chapter titles with their timestamps. A short bio for the guest. The full transcript. The episode’s own trailer. And outbound links to every platform it is on.

Each of those earns its place for a different reason. The takeaways and the chapters are what an assistant lifts. The transcript is what makes the page the only place on the internet where those sentences exist as text. The platform links are what stop the page competing with the distribution it depends on. And the title is where the search demand is won or lost, which is why it is written last, after the research rather than before it.

A page in that shape is not a marketing page for an episode. It is the episode, in the form a machine can read.

The schema that makes an interview citable

Structured data on a media site is doing something more specific than it does elsewhere: it is telling a machine that this page is an episode, that the episode belongs to a series, that a video exists on it, and that the person being interviewed is a person with a name that means something. Those are four different declarations, and a page carrying all of them is dramatically easier to quote correctly than one carrying none.

We validate every block before it ships and we keep it describing what is actually visible. Structured data that claims a transcript on a page with no transcript is worse than none, because it is the kind of mismatch that gets a site’s markup discounted wholesale rather than page by page.

No hidden text, ever

A transcript is long, and the temptation to tuck it away is immediate. A collapsed section the reader can open is fine — the text is in the page, the reader chose not to expand it, and every machine reading the page still sees it. Text hidden with CSS so that nobody can ever reach it is not fine, and the distinction is not a stylistic preference.

The rule we work to is simple enough to hand to a designer: if a human being cannot get to it by clicking something, it does not go on the page.

Check the intent before you write the title

Guest names look like easy search targets and frequently are not. The clearest example we have run into: for a well-known retail founder, searches pairing the name with the word interview are overwhelmingly job-applicant intent. The suggestions that come back are about interview questions, the interview process, what to wear, what the starting pay is. An episode page targeting that phrase would have competed with a careers page for an audience that was never going to listen to anything.

The demand that actually existed was for the story — the origin, how the thing got built, what happened when it nearly did not. Different words, a different page, and a difference you only find by checking what a search box suggests before you commit a title to it.

The same check cuts the other way and saves arguments. A framing the team loved turned out to have essentially no search demand behind it, which is a useful thing to learn in ten minutes rather than after a quarter of publishing. That framing was ours, not the audience’s. Meanwhile, for another guest, people were searching for him together with the name of a specific show — so putting the show’s name in the title was doing real work rather than branding.

Intake, write, publish

The reason most media sites do not have pages like this is not that nobody wants them. It is that the work lands on release day, when there is no time. So the process starts two weeks ahead of release, with a hard floor of one week, and it asks for the handful of things that cannot be derived: the guest’s name and any other spellings of it, every company or show or team they are known for, anything they are currently in the news for, anything launching around the episode, the most memorable moments with rough timestamps, and anything off-limits.

Almost everything else comes out of the transcript. Once the transcript and the artwork exist, the title, description, summary, takeaways, chapters and bio are draftable, and the page can be staged through the content system by API rather than assembled by hand on the morning of release.

Two safeguards make that safe to run at speed: the tools do nothing to the live site unless explicitly told to, and the publish step refuses outright if a placeholder is still sitting in the page. A lorem ipsum episode indexed and live is not a hypothetical failure — it is one we have found on a client’s site before we started.

What we control, and what we only advise on

The platform pages are generated from your feed. We cannot restructure them, and anyone who tells you otherwise is selling you something. What we can do is tell you what to put in the fields the feed does control, and then make the site the part of your presence that is genuinely yours.

That boundary matters more in this industry than in most, because so much of the audience relationship happens somewhere you do not own. The site is what we control, and the words are what the site is for.

We also work in property management, finance and sports & recreation. The foundations are shared; what changes per industry is which query gets typed and which page type answers it.

FAQ

Media Entertainment — questions we get

Our episodes are already on YouTube, Spotify and Apple. Why do we need pages?

Because those platforms answer a different question. They win “where do I listen,” and they should — we link to all of them from every episode page. What they cannot answer is “what did she actually say about that,” because an assistant cannot quote audio. It quotes text, and cites the page the text is on. If the only text about your episode is a two-line description, that is the most anyone can cite.

Do we have to publish the whole transcript?

It is the single highest-value thing on the page, so we push hard for it. A full transcript turns one episode into a page that can match hundreds of different phrasings of a question, and it is the thing an assistant quotes from. It can sit inside a collapsed section a reader opens — that is fine, the text is still in the page. What it cannot be is hidden with CSS so nobody can reach it.

Who writes the episode page — you or us?

We draft it from the transcript, and a person on your side approves it before it goes live. Most of the page is derivable: the summary, the takeaways, the chapter titles, the bio. What we need from you is the short list only you know — alternate spellings of the guest’s name, what they are currently in the news for, anything launching around the episode, and anything off-limits.

How fast can a page go up after an episode drops?

Same day, if the transcript and the artwork arrive ahead of release. That is why the intake runs two weeks out with a hard floor of one. The work that cannot be compressed is the research behind the title, and that is the part that decides whether the page finds anyone.

We migrated platforms and our search traffic collapsed. What happened?

Almost certainly your old URLs stopped resolving. We have inventoried a migration where every single one of forty-two article URLs returned an error afterwards — years of accumulated links and rankings pointing at nothing. It is recoverable: map the old addresses to their closest new equivalents, redirect each one, and use patterns for the sections that moved wholesale. The sooner it is done the more of it comes back.

How do we rank for a guest’s name when bigger sites have covered them?

You usually do not beat a major outlet on the bare name, and you do not need to. You win the long tail the outlet never wrote: what that person said about one specific thing. A chapter title naming the topic, sitting above the transcript of that exact stretch of conversation, exists in one place on the internet. That is a question your page answers better than anybody, and it is asked far more often than the name alone.

Should the show live on its own domain or on the company’s?

On the company’s, in almost every case. A separate domain starts from zero and splits whatever authority the company already has in half. A section of an established site inherits it. The exception is a show that is genuinely a separate business with its own commercial life, and that is a judgement about the business rather than about search.

How do we get an AI assistant to quote our interview?

Give it something quotable and tell it what the page is. A specific claim needs to sit in a sentence that survives being lifted out of the paragraph around it, which is a writing decision. Chapter titles and takeaways do that naturally. Then declare it in structured data: this page is an episode, of this series, with this video, featuring this person. There is no lever beyond that — and be wary of anyone who claims one.

Does this only work for podcasts?

No. The argument holds anywhere the media lives on a platform and the words live nowhere: video series, conference talks, interview programmes, archived broadcasts, live events with a recording. The page types differ, the principle does not — own the text, because the text is what gets quoted.