---
title: "Podcast Discovery Platform: Every Episode Findable"
description: Podcast discovery runs on friends and YouTube. The engine that moved multi-page sessions +46% on publisher web makes every episode findable, cross-show.
url: https://preview.artificialpoets.com/use-cases/podcast-discovery/
site: Artificial Poets
type: page
date: 2026-08-09T19:46:31+00:00
modified: 2026-08-18T02:10:14+00:00
---
# Every Episode Findable

**The same engine, on new surfaces**

"We make great shows nobody finds — how does a listener discover the next episode, or the next show?" Demand is not the problem: 58% of Americans listen monthly, but discovery happens through people they know (56%) and YouTube (52%), not on surfaces you own (Sounds Profitable). The Platform vectorizes every episode through its transcript and surfaces the right one per listener across shows and formats, proven on publisher web pages and, here, the same engine on a new surface.

[See it on my numbers](/request-a-demo/)

Horse & Rider · Ride TV · Equine Network · Equus Magazine · The Score · Equine Network Lockup

## The problem, in your numbers

Listening is at a record. Discovery is outsourced to friends and someone else's algorithm. The gap between the two is your back catalogue.

- **72%** Podcast creators naming discoverability and audience growth their top challenge (Independent Podcaster Report, via The Podcast Host)
- **56%** Discover new shows via people they know; 52% via YouTube — the top named sources (Sounds Profitable)
- **44%** Streaming fans who discover on social but consume elsewhere — a black box you can't attribute (Deloitte)
- **58%** Americans 12+ listening monthly — an all-time record (Edison Research, The Infinite Dial 2026)

## What the fingerprint connects

One listener's recommended path. Today each show's audience lives inside its own feed; the fingerprint draws the routes between them — from the episode they just finished on Show A to a relevant episode of Show C, and to a related article. Cross-show, cross-format, chosen per listener.

- **Show A — just finished** — The episode the listener has just completed, inside its own feed.
- **Listener fingerprint** — One anonymous interest profile, matched against the whole catalogue rather than a single show.
- **Show C — recommended** — The episode of another show this listener would choose, if anyone had ever introduced them.
- **Article** — The related written piece, because articles and episodes share one index.

## The evidence, at its honest tier

Our measured results are on publisher web surfaces. There is no measured podcast deployment yet — this application runs the same mechanism on a new surface, and we won't blur that line. What the engine did on web:

- **Our measured results are on publisher web surfaces.** — Podcast discovery runs the same mechanism, vectorization, fingerprinting, intent matching and orchestration, on a surface that carries no measured claim yet. The web figures below are the mechanism's evidence, not a podcast result.
- **Multi-page sessions +46% on enabled titles** — (10.3% → 15.1%) while comparable titles fell 16% (12.0% → 10.1%) over the same nine months
- **Largest single title** — +19.1% in the first quarter after enablement, +4.0% at steady state — both figures printed, always
- **Fewer than 0.2% of readers on the multi-title network we measure touch a second title** — the cross-property discovery this use case builds is, today, almost entirely unbuilt
- **A first podcast deployment gets the same measurement design** — four-week baseline, frozen metrics, a comparison group of shows held back

The web deployment behind the mechanism, not a podcast result: [two titles grew multi-page sessions 46% while the comparison group fell 16%](/customers/equine-network-rollout/).

**Introducing**

## Artificial Poets Platform

One engine behind every solution on this site. It learns your archive and your readers, then acts inside your CMS, your templates and your ad stack.

- It learns your archive: **Every story you have published, current again** — The engine understands each piece by what it is about, not when it ran or where it was filed. A feature from 2019 competes for the next slot on merit with one from this morning.
- It reads the visit: **What a reader wants, without asking** — Interest builds from what someone actually does in the session. No login, no third-party cookies, nothing leaving your domain. Useful on the second pageview, not the tenth visit.
- It chooses: **The right next read, not the popular one** — Someone comparing products and someone following a running story want different things. A most-read list gives both the same five links and serves neither.
- It serves: **There before the reader leaves** — The feed, the recommendations, the search answer and the signup ask all run on the same engine, in your templates and your ad stack. Any slot that arrives with them is yours to sell.
- It proves: **A lift you can defend, or we say so** — Every deployment runs beside titles that did not get it, plus a serving pause. That is how a result becomes a number you can take to a board instead of a vendor claim.

## FAQ

### Almost none of our listening happens on our site. What is a module on our own pages supposed to change?

It works where you own the surface, and that is a smaller share than the pitch usually admits: website mentions account for a single-digit share of how listeners find a favourite show, against YouTube in the high thirties (Sounds Profitable, 2026). What your own pages can do is connect readers who are already deep in a subject to the show about it.

### We have a few hundred episodes, not a Netflix-scale catalogue — what will this surface that our producers can't hand-pick?

Producers pick per show; the engine picks per listener. A few hundred episodes across a handful of shows is tens of thousands of listener-to-episode pairs nobody has time to curate — and the pairs that matter cross show boundaries, where no producer is looking. Matching runs on meaning, not popularity, so a small catalogue is matchable from day one, and the back catalogue is where it earns: an episode recorded three years ago resurfaces when a fingerprint says it is relevant now.

### If a listener acts on a recommendation inside Spotify, how would either of us know?

Neither of us would, and anyone claiming otherwise is mismeasuring. Platform analytics are aggregate, anonymised and platform-scoped, with no join key back to a site session. We can measure what happens on surfaces you own. We will not convert that into a claim about listening we cannot see.

### There is no cookie in an RSS feed. How do you personalise for a listener you cannot identify?

On your own surfaces, where a first-party anonymous profile exists. In the feed itself there is no identity to personalise against, and podcast telemetry is device-shaped and delayed by design. This is why the honest scope here is discovery on properties you control, not personalisation inside the apps.

### We run on a hosted stack — a Megaphone- or Art19-class platform — with no dev team. What does integration take?

Hosting stays where it is. Serving runs on the web surfaces you already control — show pages, episode pages, the editorial site — as a JavaScript integration carried out by our team. Feeds are untouched; where the graph suggests a promo swap or a feed drop, your producers run it in the tools they already use.

### We have four thousand episodes and no transcripts. What is the lift on our side?

Transcripts are what makes an episode matchable by subject, so producing them for the back catalogue is the real preparatory work. Adoption is low industry-wide, so this is a common starting point rather than a disqualifier, and it is worth scoping before anything else is discussed.

### Show me incremental lift, not engagement vanity metrics. How would you prove this worked?

The way the web numbers were produced: a four-week measurement-only baseline, metrics frozen before enablement — listen-through rate, second-episode conversion, cross-show starts — a comparison group of shows held back, and the failure condition agreed before the switch is flipped. No podcast deployment has been measured yet; the first one is designed so its numbers survive your analyst.

## Related

- [Your Reader, Viewer and Listener Are One Person.](https://preview.artificialpoets.com/use-cases/cross-media-graph/) — "Why does my podcast app have no idea what this person reads?"
- [One Audience. Every Property.](https://preview.artificialpoets.com/solutions/media-networks/) — "We own a dozen properties — why does our audience behave like it belongs to strangers?"
- [Get Them to the Second Episode](https://preview.artificialpoets.com/use-cases/video-recommendations/) — "Subscribers churn before a second episode — how do I get them watching in the first session?"
- [You Can’t Get the Traffic Back. You Can Get the Session Back.](https://preview.artificialpoets.com/resources/traffic-back/) — The economics of session depth, and the causal evidence behind it: +53% pageviews per user, 95% CI [+48%, +58%], with the modelling assumptions written down.

## See how this works on your titles

Thirty minutes on your analytics. We tell you what share of your sessions stop at the first page, and what this would realistically move first for a setup like yours. You leave with the annotated read. No deck.

[See it on my numbers](/request-a-demo/)

How these numbers are made: [the measurement method](/solutions/measurement/). Figures from a multi-title publisher network measured continuously on publisher web surfaces, August 2025 to May 2026. Last updated August 15, 2026.

## Questions this page answers

### Why is podcast discovery so hard?

Because discovery happens where you are not. Demand is at a record, with 58% of Americans listening monthly (Edison Research), but the top named sources of new shows are people they know at 56% and YouTube at 52% (Sounds Profitable). The surfaces you own are barely part of the discovery path.

### How would the Platform surface the right episode?

Every episode is vectorized through its transcript, so a show becomes searchable and matchable by what it is actually about, and one anonymous interest profile per listener carries across shows and formats. Stated plainly: our measured results are on publisher web pages, and this surface carries no measured claim yet.

### How much does it cost?

There is no public price list; pricing follows network size and formats. Because podcast discovery is unproven for us, a first deployment is scoped as a measured pilot: four-week baseline, frozen metrics, and a comparison group of shows held back so the result can be checked.

### Who is it for?

Publishers with a podcast network and an editorial site, where the articles and the shows cover the same subjects for the same audience but live in separate systems that never refer to each other.

### What are the alternatives?

Cross-promotion in episode reads reaches only existing listeners. Platform charts reward what is already large. Paid discovery buys attention without keeping it. The alternative here builds discovery on surfaces you own, which is the only place the referral is yours, and it is not yet measured.

### How do I get started with podcast discovery?

Transcripts and feeds are what the engine indexes, so the first work is making the catalogue readable to it. A first podcast deployment then gets the same measurement design as everything else: four-week baseline, frozen metrics, and a comparison group of shows held back.