The Business of Podcasting: Subscription Networks vs Ad-Supported Models
Advertising rewards the show a million people hear once. Subscription rewards the show ten thousand people cannot do without. Almost nothing about running them is the same.
Podcasting is repeatedly described as a single industry with a single problem, usually monetisation. It is two industries. One sells attention to advertisers and is priced on how many people can be reached. The other sells access to listeners and is priced on how much they care. The operational decisions that follow from each are close to opposites, and most disappointing audio businesses are running one while measuring themselves with the other's numbers.
Two economics, one medium
| Ad-supported | Subscription | |
|---|---|---|
| What is sold | Audience attention, per thousand | Access, per listener per month |
| Audience shape needed | Broad and predictable | Narrow and devoted |
| Key metric | Deliverable impressions | Churn and lifetime value |
| Cost of a bad episode | Low — the next one resets it | High — cancellations are permanent |
| Scales with | Audience growth | Pricing and retention |
| Revenue volatility | High, follows ad cycles | Low once the base is built |
The differences compound. An advertising-funded show benefits from every incremental listener, however casual, because inventory is sold by volume. A subscription show gains nothing from a casual listener and everything from a devoted one, because the same content is sold once per person regardless of how often it is consumed.
That is why growth tactics transfer badly between the two. Chasing broad reach for a subscription product adds cost without adding revenue, and narrowing a show to a devoted niche removes the scale an advertiser is paying for.
The download problem
Podcast advertising is still transacted largely on downloads, and the download is a poor proxy for a listen. It records that a file was requested — by a person, by an app pre-fetching episodes, by a device that will never play it. Industry measurement standards have improved the situation substantially by filtering automated requests and defining what counts, but the metric still describes supply rather than attention.
The practical consequence for buyers is that podcast CPMs are high relative to other audio while the underlying attention is imperfectly verified. The practical consequence for publishers is that certified measurement is not a technicality: it is the difference between being bought by a media agency and being bought by an enthusiast.
The honest defence of the medium is that attention within a listen is unusually strong — the format is long, sequential and consumed alone, and host-read advertising carries transfer of trust that few other placements do. That is a real advantage, and it is not the same as knowing how many people heard it.
What dynamic insertion changed
Advertising used to be baked into the audio file at production, which meant an episode carried the same advertiser forever and old episodes earned nothing new. Dynamic insertion stitches advertising at the moment of download, and its effect on the business has been larger than its technical description suggests.
The archive became inventory. A show with four years of episodes suddenly had a saleable long tail, which is why catalogue depth started appearing in valuations. Campaign flexibility improved, targeting by region and device became possible, and buying moved closer to how digital audio is otherwise transacted.
It also cost something. Host-read endorsement loses force when it is obviously stitched, mid-roll placement in an archive episode has no editorial context, and the format drifted towards the interchangeable inventory it was originally valued for not being.
Where subscription genuinely works
- Utility over entertainment. Shows that save a listener time or money — analysis, professional briefing, market coverage — convert at rates general-interest shows rarely reach.
- A back catalogue worth having. Subscriptions are bought partly for the archive. A show with three hundred episodes has a stronger offer than one with thirty, at the same audience size.
- Something the free feed cannot be. Ad-free versions of the same episodes convert poorly. Extra episodes, full interviews, written companions and community access convert considerably better.
- An audience that has already paid for something. Listeners arriving from a paid newsletter or a conference convert several times better than listeners arriving from a chart position.
The hybrid that works
Most durable audio businesses run both, but not as the same product at two prices. The free show is a marketing asset carrying advertising and doing discovery work; the paid tier is a different product with a different promise. When the paid tier is simply the free show without ads, the publisher is charging listeners to remove something the publisher chose to add, which is a weak proposition and converts accordingly.
The European position adds a constraint worth planning around: podcast advertising markets outside the largest countries are thin, and a show in a smaller language market may find that subscription is not a strategic preference but the only viable model at its scale.
Note on data. Podcast CPMs, conversion rates and churn figures vary widely by market, language and category, and most publishers do not disclose them. Measurement standards continue to evolve, which means download figures from different periods or providers are not directly comparable. Treat any single benchmark circulating in this category as indicative.
Sources
The claims in this article rest on the documents below. Each is linked to what it establishes, so you can check any statement against its origin rather than taking ours for it.
- IAB, US Podcast Advertising Revenue Study (PDF) — the industry's standing measurement of the ad-funded side of the business
- Radio Ink, on IAB's 2025 digital audio figures — podcast advertising reaching roughly $2.9bn, growing faster than digital audio overall
- Adweek, IAB podcast upfront — how the ad market itself frames the subscription-versus-advertising choice
Frequently asked questions
Is podcast advertising or subscription more profitable?
Neither, in the abstract — they suit different audiences. Advertising monetises breadth and pays regardless of listener commitment. Subscription monetises intensity and can produce more revenue from a far smaller audience. The wrong model for your audience shape underperforms whichever it is.
Why are podcast downloads considered an unreliable metric?
Because a download records a file request rather than a listen. Automated pre-fetching by apps and devices inflates counts, and while measurement standards now filter much of this, the metric still describes distribution rather than attention. It is the currency the market uses, not a measure of who listened.
What is dynamic ad insertion and why does it matter?
It places advertising into an episode at the moment of download rather than baking it into the file at production. Its major effect was making back catalogues saleable, which turned archives into inventory and changed how audio businesses are valued.
Should a paid podcast tier just remove the ads?
It converts poorly. Charging listeners to remove something you chose to insert is a weak offer. Tiers that add genuinely different value — extra episodes, unedited interviews, written analysis, community access — convert considerably better at the same price.
Does subscription work for podcasts outside the largest markets?
Often it is the only model that works. Advertising markets in smaller language territories are thin, and a show that could never sell enough impressions to fund itself may support a modest subscriber base comfortably. Scale constraints push smaller markets towards direct revenue earlier.