Authored by Chris Norton, Managing Director, Prohibition in Leeds, United Kingdom
For the first time on record, more UK podcast listeners pick YouTube as their go-to platform than Spotify. The UK Podcast Consumer 2026 report from Edison Research puts YouTube on 29% of weekly listeners and Spotify on 28%. BBC Sounds sits back on 15%.
In the US the picture is starker still. Sounds Profitable’s Podcast Landscape 2026 found that 40% of podcast consumers name YouTube as their most-used platform. Apple Podcasts, the app our whole industry was built around, is now the main home for just 8%.
So here is my prediction for the next three years. YouTube will not just be a podcast platform; it will be THE podcast platform. Most agencies are still selling shows as if it were 2019, and that gap is where the opportunity sits.
We called this one early
A few years ago, my business partner Will and I recorded an episode of Embracing Marketing Mistakes about video. I said on air that YouTube was having a rebirth. We talked about how longer videos outperform shorter ones there, which felt counterintuitive at the time, because every other platform was racing towards fifteen-second clips.
I’m not claiming we’re psychic. The signs were just there if you were making a show yourself. Our YouTube channel kept growing while the audio numbers held steady, and the comments were coming in on YouTube, not Apple. The audience had moved before the industry reports noticed and caught up.
The data now backs it up. Sounds Profitable found that 44% of podcast consumers watched a video podcast in the past week, up from 40% a year earlier. Only 17% still expect a podcast to be audio-only. Netflix is even getting in on it, with 21% of consumers saying they’ve used it for podcasts in the past year, which is nothing short of bizarre if you remember what a podcast used to be.
YouTube rewards testing, not perfection
The big shift for agencies is that YouTube is a discovery engine, not a directory. Apple and Spotify mostly serve people who already know your show exists. YouTube puts your episode in front of people who have never heard of you, provided the title and thumbnail earn the click.
That changes the job. Leroy Ter Braak, a YouTube strategist, made this point brilliantly when he came on the podcast. He said his biggest mistake was not testing more, especially on videos that were already doing well. When something was pulling in a hundred thousand views a month, he left it alone out of something close to superstition. He reckons that habit cost him over a hundred million views.
I think most agencies make the same mistake with client podcasts. We upload the episode, write a decent title, and move on to the next one. On YouTube, the episode is never finished. The title, the thumbnail and the first thirty seconds are things you keep testing for months.
Leroy also said something every strategist should hear. He gets paid to be right more often than average, not to be right every time. That’s a healthy way to sell podcast work to clients, actually, because nobody can promise a hit.
Downloads still matter, whatever the reports say
This is where I part company with a lot of the trend pieces I’ve been reading. The fashionable line is that downloads are dead and agencies should only talk about pipeline and attribution. I simply don’t buy that argument.
Downloads are a reality check, and reality checks are underrated in our industry. One guest told me about a 2007 Nokia campaign to launch a music service, built around exclusive rights to a new Kylie Minogue album. The client spent around fifteen million pounds on media, technology and rights. Thirty-seven people downloaded the app.
He’d warned them that most viewers didn’t own a handset that could run it. Someone had already made the business case, though, so nobody could stop it. Another guest joked that they could have paid for a private Kylie gig in every buyer’s living room. Without that download number, the campaign would probably have been reported as a success on reach alone.
So yes, track pipeline and influenced deals. But don’t let anyone tell you the raw number is irrelevant. Download velocity in the first week still decides chart positions, and charts still win guests, sponsors and new business. For context, 428 downloads in the first seven days puts a show in the top 10% globally, according to benchmarks from Share Your Genius. Most B2B clients have no idea that figure is so low, so set expectations with it early.
AI belongs in the edit suite, not the host’s chair
AI is already changing how we make shows. We use OpusClip to find and cut short clips, and Claude or ChatGPT for guest research, question prep and show notes. It saves hours every week, and I wouldn’t go back.
It also gets things wrong, and I learned that the hard way. I once asked a guest about a band he’d been in, using research that had come back from an AI tool. He looked at me, slightly puzzled, and told me he’d never been in that band. It was the guy from Yorkshire Tea. The AI had merged two different people into one confident biography, and I’d read it out on air without checking.
Funny now, mortifying at the time. The lesson is simple: AI can do the legwork, but a human has to check every fact before a guest hears it.
Listeners draw the line in a similar place. Edison’s UK research found that 66% of weekly listeners are happy for AI to be used in research, and 64% for brainstorming. But 85% disapprove of AI hosts, and 72% see AI as a threat to the credibility of the medium. My prediction is that synthetic hosts will exist by 2029, but as a niche gimmick. They won’t win trust, and they won’t win B2B budgets.
What this means for agencies over the next three years If you run podcasts for clients, here is where I’d put the effort.
- Plan every show for YouTube first, then distribute the audio everywhere else.
- Budget time for testing titles and thumbnails after launch, not just before it.
- Report downloads alongside pipeline, because each keeps the other honest.
- Use AI for clips and research, then fact-check everything it produces.
- Keep a real human voice at the centre, because trust is the whole asset.
The concentration problem makes all of this more urgent. Edison found that the top 29 shows now reach half of all UK weekly podcast consumers. The big names are getting bigger, so a brand show has to be found by people who weren’t looking for it. On current evidence, that happens on YouTube.
Which leaves one question for every agency selling podcast work: is your client’s show built for the platform where people are actually watching?
