The Free Report reads your feed. This one goes and looks.

Every link, opened and followed

Not inspected. Opened. A few hundred addresses, one at a time, politely.

• Dead links — a guest’s site that has gone, a page that now 404s, a domain that expired quietly two years ago
• Links that end up somewhere other than they point — and where, minus the tracking parameters the site adds on the way
• Every mis-targeted link, named — where the words say one thing and the link goes somewhere else, listed episode by episode
• Addresses that cannot open at all — usually invisible characters picked up by pasting from a document. The link looks perfectly normal on the page, which is why nobody finds these alone. You get the address it was meant to be
• Links still using http:// — they redirect, so they work, but every listener pays an extra round trip and some apps flag them as insecure. Usually one of them is your own site
• Links to sites that answer very slowly — anything taking more than five seconds. A listener tapping one on a phone assumes it is broken before it loads, and you would never know unless you timed it
• Sponsor and affiliate links flagged — a dead one is money rather than tidiness, so they are called out separately
• An honest verdict on each — works, definitely gone, or couldn’t be verified. Instagram, Facebook and Amazon refuse automated checks, so those are flagged as needing a human click rather than quietly called broken

Every image, fetched and measured

The XML tells you an address. It tells you nothing about the picture.

• Artwork that doesn’t load — addresses in the feed pointing at files that are no longer there
• Anything under Apple’s 1400-pixel minimum, which can hold up a submission
• Crops that aren’t square, and files heavy enough to load slowly on a phone
• Every episode’s artwork, shown — the whole catalogue as a grid, so you can see for yourself how many designs are actually in play. Counting image files tells you nothing: several hosts mint a new address every time an episode is saved

Every file, opened and weighed

A tag saying there is a transcript is not the same as there being a transcript. And a file that downloads is not the same as a file that ought to be that size.

• Whether your audio is far bigger than it needs to be — the real file size against the running time gives the bitrate. Speech mastered at 320 kbps is roughly two and a half times the size it needs to be: a 79-minute episode arrives as 180MB where 72MB would sound identical on the earbuds and phone speakers people actually use. Your listeners pay for every one of those megabytes in mobile data and loading time, and nothing in the feed reveals it
• Episodes encoded at a different quality from the rest — an average hides an odd one out. An episode at half the bitrate of its neighbours does not sound like a smaller file, it sounds like a worse recording, and the listener moving between them hears the change
• Transcripts opened — confirming the file is really there, and that what comes back is a transcript rather than an error page wearing one’s name
• Chapter files opened — because malformed chapter data renders in no app at all and looks perfectly fine in the feed
• The declared file size against the real one — many feeds publish zero. That is usually what dynamic ad insertion does, and it stops apps showing a download bar or pre-caching properly
• The analytics chain traced — every prefix and redirect sitting between a listener and your audio, named. Each one is measuring something for somebody, and each is a thing that can be slow or stop existing

What the directories are actually showing

Apps do not read your feed when someone opens your show. They keep their own copy, refreshed on their own schedule, and it drifts from what you published.

The Free Report asks the same two directories and tells you whether each one has your show. This audit says what they are holding — which is the part that drifts.

• The feed address Apple is really reading — matched by address, not by name, so it is your show and not one with a similar title. If Apple is on an old address after a change of host, everything you publish reaches its listeners or it does not, and nothing tells you which
• Apple’s episode count against your feed’s, and how many days behind its newest episode is. A day or two is ordinary. A fortnight is something stopping Apple reading the feed
• The categories Apple has you filed under — which is what browse and chart pages use, and is whatever your feed said the last time Apple read it
• Whether the open index has you at all — Podcast Index is what Overcast, Pocket Casts, Podcast Addict, Fountain and most independent apps read. A show missing from it is not ranked low there, it is absent, and it is free to fix
• What the index’s last read of your feed returned — the answer your server gave it, and any errors it hit parsing what came back
• What your feed itself declares about where it is distributed — the <podcast:id> field, which names the show’s id or page on each platform it is published to. Most feeds carry none, and apps that follow the open standard then have only the title to go on when connecting the same show across services
• Nothing about Spotify. Spotify does not publish the feed address anywhere in its system, so no tool can say which Spotify show belongs to a feed — a match by name alone would be a guess, and a guess is not a finding. This audit says nothing about Spotify rather than something that might be about a different show. The Free Report lets you point at your own show on Spotify if you want the link on record; that is your identification, not a check.