Podcast news roundup: 12 November 2026

Seven days, 5 to 11 November 2026, and the same argument turned up three times wearing different clothes: how much of a video podcast is anybody looking at? RAJAR put a UK number on it. Podnews ran a week-long test on its own daily show and published the lot, including the parts that did not move. AdsWizz started selling ads against the answer. Around that, five hosting companies signed up to a code that should make copied feeds harder to cash in, and Meta handed everyone a machine for making shows with nobody in them. The edit desk is about a note, eight lines long, that I wish came with every recording.

How many of your video viewers are watching?

John Carroll, RAJAR's Director of Research and Audiences, presented new video podcast research at the ASI Audio Conference in Venice the week before, and Adam Bowie wrote it up on 10 November. The figures come from RAJAR's Podcast Tracker sample, which is not public, so I am taking them from Bowie's account, alongside RAJAR's own summary.

In a given week, 34% of the podcast audience flips between audio and video. 37% stay audio only and 29% video only. Then the finding that should be pinned above every edit suite: of the people who use video podcasts, 20% mostly watch. 35% mainly listen and only occasionally look. 43% split it fairly evenly. Add the last two and 78% are only partly watching. What are they doing instead? Household chores for 41%, and other things on their phone for 37%. Video podcasts are also significantly more likely to be seen on a phone than on a TV. Podnews, covering the research on 11 November, added the scale: 105 million hours of podcasts a week in the UK.

Bowie's own reading is that video assets are close to essential for discovery, while a lot of the viewing is secondary, closer to daytime television. I agree, and I would push it further. If four in five viewers are folding laundry, the picture is the shop window and the sound is the product. A visual gag nobody explains, a chart held up to camera, a guest pointing at something off screen: each of those is a hole in the episode for most of the audience. Say what is on the screen. We make the same case in video podcast or audio podcast, and now there are UK numbers behind it.

Did a week of video change anything for Podnews?

James Cridland published the results on 10 November. For one week in October, Liquid Studios turned each Podnews Daily episode into a landscape YouTube video, a vertical cut and chapter markers. Cridland is upfront that it cost Podnews nothing, that no generative AI images were used (stock library pictures filled the gaps), and that a human editor at Liquid Studios checked each first cut.

YouTube views rose sevenfold across the week. Launch day had a press release behind it, so the fairer figure is the other four editions, which averaged almost four times a normal day. Engaged views went to 82% from 69%. Impressions in the first 24 hours roughly doubled, a median rise of 121%. Average view duration crept up about 5%. LinkedIn's five videos drew more views than YouTube did, though there is no baseline to compare. And the half that matters as much: Spotify plays and impressions were flat, audio downloads were not meaningfully different, and when the pilot ended daily views dropped back to where they started.

Cridland concludes the lift came from the format. It is one show, one week, and a supplier's pilot of its own product, so treat it as a strong hint. My read: video bought reach on the platforms built for video and nothing anywhere else, and the reach left the day the video stopped. That is an argument for a format you can keep up every episode over a splash you cannot. It is also worth noticing where the person sat in that process. Somebody looked at every first cut before it went out.

UK listening: the car is radio's, the morning is yours

The same Bowie post covers RAJAR's MIDAS Autumn 2026 survey, which he says is not comparable with the quarterly RAJAR figures. 96% of UK adults listen to some audio each week, 27.8 hours apiece. Podcast reach in MIDAS is 26% and has been flat for about a year. The quarterly RAJAR says 30.7% and Ofcom's 2026 survey says 25%. Three methods, three answers; our UK podcast listening statistics page is where we keep these as they land.

The details are more useful than the headline. Only 11% of podcast listening happens in the car, against 25% of radio listening on RAJAR's Q3 2026 figures. Podcast listening peaks at 8.30am. Most people hear three or fewer episodes a week, and 6% hear eleven or more. Word of mouth is the top route to a new show, social media second, and suggestions from players and charts third. So you are competing for one of three weekly slots, and the thing most likely to win it is a friend saying your name. Make the episode somebody would recommend before you worry about the algorithm.

Will your host sign the code on copied shows?

Last week's edition covered the copied shows Podnews found. On 10 November Podnews reported that Spreaker has committed to changes to fight unauthorised copies, and Podnews has drafted a Code of Practice for podcast migrations. Its code page lists Captivate, RedCircle, RSS.com, Spreaker and Transistor as committed to follow it.

The code asks three things of a host importing a show: respect the podcast:locked tag, tell the original creator when an import begins, and allow no monetisation and no one-click distribution to Apple, Spotify and the rest until the redirect from the old feed is confirmed. That third rule is the one with teeth, because it removes the money. Two jobs for you this week: check your feed is locked, and ask your host whether it has signed.

A studio nobody is sitting in

Podnews reported on 11 November that Meta's Muse assistant, available since September, can research a subject, record a podcast with artificial voices and host it on Meta's servers. Podnews found more than 20 shows already hosted there. Its longer piece follows a pitch from a man calling himself Charles Mayfield, who says he runs a podcast network. Podnews reports that the photo supplied with the pitch was AI-generated, carries a digital watermark saying so, and shows a studio with a twelve-channel mixing desk he is not using.

I am not worried about these shows taking your listeners. Go back to the MIDAS finding: people find podcasts through other people. Nobody tells a friend about a show with no one in it. I am worried about directories filling with noise, and that makes a real voice, a real face and a real name on the lower third worth more than they were a month ago.

The ad money: more spend, and a video banner

AdsWizz said on 6 November that it has opened its video podcast monetisation for Apple Podcasts to publishers and creators, after early testing with select publishers. Advertisers can extend an existing audio campaign into video by pairing the audio ad with a branded companion banner, with no new video ad to produce. Analytics come through Simplecast. AdsWizz claims it powers one in four ads across the global open ecosystem and has monetised more than 1.3 trillion impressions in a decade; those are the company's own numbers. Read it next to RAJAR's 78%: an audio ad with a banner is an honest fit for an audience that mostly listens.

Per Podnews on 10 November, Magellan AI's Q3 2026 benchmark report, which sits behind a registration wall, has podcast ad spend up 18% year on year. Ad load was 8.86% in Q3, from 8.75% in Q2 and 8.34% a year ago, roughly 20 seconds more advertising than a year ago in a typical hour.

From the US, Nielsen's Q3 2026 edition of The Record says radio and podcasts take 83% of daily ad-supported audio time and streaming music 15%, using Edison Research's Share of Ear data, which leaves out music on YouTube. Radio reaches 92% of US adults monthly. Podnews, on 9 November, reads the report as podcasts at 21% of that time against radio's 61%, and, among 18 to 34s, podcasts at 33% of all audio time against radio's 42%, and notes Edison data showing US podcast listening time up 386% in ten years.

In brief

The Independent Podcast Awards 2026 were held in London, per Podnews on 6 November. Independent Podcast of the Year went to Great Days Out on the bus from Buxton by Maria Passingham. The runner-up was War Hall, on the wartime history of Walthamstow's Assembly Hall, whose co-host Angela Allen MBE, aged 96, got a standing ovation.

Apple says listener growth in Japan is nine times the global average, new Japanese shows are up 70% year on year, and more than four in five new shows come from independent and non-traditional creators, per Podnews on 5 November.

Podnews reported on 11 November that Message Heard, a UK podcast company, has gone into voluntary liquidation. If you freelance, it is a reminder to know exactly who you are contracting with before the work starts.

Also from Podnews: Flightcast is now certified under the IAB Podcast Measurement Guidelines v2.3 for download, listener and ad delivery (11 November). And Chrome is showing red errors on RSS feeds that use XSLT; Transistor has made a CSS stylesheet alternative (9 November).

The numbers

From the edit desk: what should you send your editor along with the recording?

I have edited well over a thousand episodes, and a great many arrived the same way: a folder of files and the word "enjoy". Then the first cut goes back, and the notes that follow are all things the host knew on the day. The guest's surname is spelt wrong on screen. The sponsor read lost a line the sponsor insists on. The story the host loved has gone, and the one they hated is the opening.

None of that is an editing problem. The editor was guessing. A short handover note, written in the five minutes after you stop recording while it is still in your head, removes the guessing.

The mistakes I see most, in order. Names: a company said aloud can be spelt four ways, and captions and lower thirds need the right one. Cuts: you know the moment the guest asked to strike something, so say "cut that" and clap once when it happens, then list it. Files: nobody labels which camera is which, or mentions that the second audio file is only a backup. And purpose: if I do not know what the episode is for, I cannot tell which tangent to keep.

Copy this, fill it in, and send it with the files:

1. This episode is for: (one sentence the listener should leave with)
2. Running order: (the main beats, in the order you want them)
3. Open on: (the moment you want first, roughly where it is)
4. Names and spellings: (every guest, job title and company said aloud)
5. Cut these: (each "cut that" and clap, with a rough time and why)
6. Sponsor reads: (where each one sits, and any line that must stay word for word)
7. Files: (which camera and mic is whose, and which files are backups)
8. The short: (the one clip you already know you want)

Line 1 does the most work. This week's RAJAR research is the reason: if most of the audience is only half watching, the point has to land in the sound, and I can only protect it if I know what it is. Line 3 matters nearly as much; there is more on choosing that moment in our guide to the podcast cold open. Eight lines, and the first cut comes back looking like the episode you thought you recorded.

So what goes in the folder with your next episode?

Decide before you press record. Either the files go over on their own and your editor guesses, or they go with eight lines and the cut matches what was in your head. If you would like those eight lines read by a human editor who then builds the episode for the 78% who only partly watch as well as the 20% who watch properly, a full episode edit costs £280. We bill every four weeks, with no minimum and no VAT, and revisions carry on until you are happy with it. Tell us about your show and paste your note in. Back next Thursday.

If you'd like a hand with your own show, here's our full episode edit and what it costs.