Sound to Sleep To: Inside the Quietly Enormous World of Sleep Podcasts
Sleep podcasts are third by hours streamed and command top-quartile CPMs, yet critics rarely review them. A craft-led look at six major shows, the production techniques that separate them, and why the genre's invisibility is its design.
The category most critics never listen all the way through
Open Spotify's annual data dump, Apple's top fifty or BBC Sounds' charts and a strange asymmetry shows up. The podcasts everyone writes about — the prestige interviews, the chart-topping news shows, the football panels — are not the shows people press play on most often. The longest cumulative sessions, the highest repeat-play rates and a startling share of overnight minutes belong to a genre most critics have never reviewed properly: sleep podcasts.
It is an old joke inside production circles that you cannot write a 1,500-word feature about a show whose job is to render you incapable of finishing the second paragraph. But the craft inside Get Sleepy, Nothing Much Happens or BBC Radio 3's Slow Radio strands is, to anyone who has tried to record a thirty-minute murmur without tipping into self-parody, every bit as exacting as the cold open of a Goalhanger show. Different end state, same engineering discipline.
This piece is an attempt at the missing review — the genre as a craft, the major shows compared on their own terms, and the question of why the most-listened-to corner of the medium remains its least-discussed.
How big is "big"?
A few numbers, drawn from public platform reports and recent industry surveys, to set the scale.
- Edison Infinite Dial 2025 reported that 22% of weekly US podcast listeners use the medium specifically to fall asleep, up from 14% in 2022.
- Spotify Wrapped 2024 flagged Sleep & Relaxation as the third-largest podcast category by hours streamed, behind only News and Comedy.
- Get Sleepy, launched by British host Tom Jones in 2019, passed 300 million downloads in early 2026 — placing it inside the top thirty UK shows by all-time downloads, alongside The Rest Is Politics and No Such Thing as a Fish.
- Apple's 2025 retention data, presented at Hot Pod Summit in March, showed sleep-genre episodes averaging 41 minutes of listening per session against a podcast-wide median of 26.
- Acast's Wind-Down Wednesday packages, sold from late 2024, command CPMs roughly 30% higher than the platform's general-comedy inventory — the ad-spot pool is shallow and advertisers will pay to be the only mattress brand a listener hears all night.
A category that is third by hours and commands top-quartile ad rates is not a fringe one. It is simply read by a different audience to the one critics belong to.
Six shows, six theories of how to put a listener to sleep
| Show | Host / Network | Format | Episode length | Production notes |
|---|---|---|---|---|
| Get Sleepy | Tom Jones, Slumber Studios (UK) | Original guided story with meditation top-and-tail | 45–60 min | Multiple narrators in rotation; music beds under speech, never above; ~3 releases per week |
| Nothing Much Happens | Kathryn Nicolai, iHeart (US) | Cosy low-stakes narrative read twice — once at normal pace, once 15% slower | 28–32 min | Single voice; no music under the story itself; explicit "the story is going to repeat now" turn |
| Sleep With Me | Drew Ackerman, "Scooter" (US) | Rambling, deliberately boring monologue | 60–90 min | The originator (2013); ASMR-adjacent vocal fry; structured tangents; no music whatsoever |
| Slow Radio | BBC Radio 3 | Ambient field recordings — moorland, cathedrals, forest paths — minimal speech | 30–90 min | Binaural and spaced-pair mic techniques; commissioned producer pieces; archived on BBC Sounds |
| The Sleepy Bookshelf | Slumber Studios (UK) | Public-domain novels read slowly, in 30-minute instalments | ~30 min | Same network as Get Sleepy; quieter, less performative read than an audiobook |
| Boring Books for Bedtime | Sharon, solo (US) | Obscure 19th-century texts read deadpan | 30–45 min | One mic, one chair, no edit beyond noise cleanup; commitment to genuinely dull source material is the bit |
Six different theories of what makes audio sleep-friendly. The disagreement is the interesting part: Get Sleepy trusts music as a sedative; Nothing Much Happens will not have it; Sleep With Me refuses both music and a coherent plot; Slow Radio refuses speech altogether. They cannot all be right. They are all, by their listeners' lights, working.
The narrator problem
The first decision a sleep-podcast producer makes is voice. And the conventional rules of good podcast narration — projected, varied, alive — are the wrong ones. What you actually want is closer to what radio-drama directors call flat-affect intimacy: a voice close enough to feel proximate but not so emphatic that the brain leans in.
Get Sleepy keeps a roster of British and Irish narrators in rotation. Tom Jones records the meditative book-ends himself; the long middle stories rotate between voices including Arif Hodzic and Esther O'Moore Donohoe. The rotation matters: listeners report developing strong preferences and using the host filter on the show's app like a sleep-mask choice, not a content one.
Nothing Much Happens takes the opposite view. Kathryn Nicolai has now read more than 400 of her own stories; she trained as a yoga teacher and her cadence is, to a fault, identical episode to episode. The single-voice approach is the show's central craft decision — predictability is the sedative.
Sleep With Me's Drew Ackerman is the outlier, and the genre's most-imitated voice. His delivery is high vocal fry, slow, with an audible smile and constant "you know"s. It is not technically beautiful. It works because the listener cannot anticipate where the sentence is going, gives up trying, and disengages.
The production palette
In a normal interview show the producer's job is to lift voice into clarity — high-pass at 100 Hz, light compression, de-essing, a touch of room reverb. Sleep podcasts invert nearly every dial.
- Compression: very gentle, sometimes none. The goal is to preserve the dynamic ramp from audible to barely-audible so the listener's brain stops scanning for new information.
- EQ: low-cut higher than normal — often 120 Hz — but more importantly a mild 3–5 kHz dip to soften consonant attack.
- Loudness: targets around -21 LUFS integrated, well below the -16 LUFS Spotify normalises to. Producers know their audience nudges volume up at the start and want headroom to fall away.
- Music beds: where present, mixed 14–18 dB below voice. Get Sleepy's long-running cue is a single 6/8 piano figure looped under each story; it is deliberately repetitive in a way a music supervisor on any other show would reject.
- Edits: silence is left long. A standard narrative show might cut a breath gap to 250 ms; Nothing Much Happens and The Sleepy Bookshelf often leave full seconds.
None of this is accidental. Several Slumber Studios producers have spoken publicly about A/B testing edit lengths against drop-off curves; the data, unsurprisingly, suggests that longer pauses and quieter mixes correlate with longer sessions. The category has built listening-data feedback loops that the prestige end of the industry only paid lip service to until very recently.
A British counter-current: Slow Radio
Worth pausing on the BBC's contribution, because Slow Radio is the only mainstream entrant in this space that flatly refuses the rules above. There is no narrator. There is no story. The 2024 commission A Walk Up Skiddaw is a 78-minute binaural field recording of a walk in the Lake District. Choral Evensong from Wells, broadcast and podcast-archived since 2020, is exactly what it sounds like.
The argument inside Radio 3, when the strand was first proposed, was that Slow Radio was a luxury commission for a small audience. Eight years later, the figures justify it — the strand's BBC Sounds catalogue passed 65 million on-demand requests in 2024, the bulk of those after 10pm. It is, by some distance, the most-listened-to thing Radio 3 has ever made.
What Slow Radio shows is that the genre's success is not really about story or voice. It is about a particular relationship to attention — content that lets the listener withdraw rather than demanding they lean in. Every other sleep podcast is solving the same problem with words. The BBC is solving it without them.
Why critics keep missing it
Three structural reasons, none of them flattering to the press.
- Reviewers listen for craft they can describe. Sleep audio's craft is invisible by design — the better the production, the less you remember it. Critics' job descriptions reward the opposite.
- The advertiser ecosystem is unfashionable. Mattresses, magnesium gummies, weighted blankets. None of it scans as prestigious; none of it gets a feature in The Guardian's podcast column.
- The audience is largely female and largely between 11pm and 2am. Awards juries and trade-press readerships are largely neither.
The result is a genre that has quietly become a structural pillar of the medium — high engagement, high CPMs, high download counts — without commensurate cultural attention. If you are in the business of listener-experience design, that gap is where the interesting work is.
Where to start, if you've never properly tried one
Three suggestions, ordered by likely cultural shock from a regular podcast listener.
- Start with Nothing Much Happens. Choose any episode whose synopsis mentions "garden centre" or "tea shop". You will not finish it on a first listen. That is the point.
- Graduate to Slow Radio. The Choral Evensong archive on BBC Sounds is the entry point.
- Save Sleep With Me for when you are genuinely struggling. It is an acquired taste even within the genre, but the case for it is strongest at 3am.
The shows do not need defending. They have the audience. The interesting question is whether the rest of the industry — the prestige documentary makers, the news-podcast operators, the panel-show producers — has anything to learn from a category that has been quietly winning the retention war for a decade.
Probably yes. The first lesson would be that producing for attention and producing against it are not opposites. They are the same craft, pointed the other way.