crawlora-net/podcast-gold-rush

Podcast Gold Rush — data

The discoverable head of Apple Podcasts (US, June 2026): the top charts and a broad keyword-search sample, with each show's episode count and last-activity date. It is the dataset behind the Crawlora study The Podcast Gold Rush and the Graveyard You Can't See.

The point of the dataset is a methodological one: a discovery surface (charts, search) is a survivor filter. It ranks by popularity/relevance, so it shows shows that kept going — the millions that quit after two episodes never appear. So these numbers describe the discoverable head, and the gap between it and the full podcast universe is the story.

Files

file	rows	description
data/winners.csv	500	episode-sampled shows from the charts (top ~100 across 19 genres → 1,800 unique; 500 sampled for episode feeds)
data/graveyard.csv	2,096	the keyword-search sample (~24 generic terms) — episode count + last-activity per show
data/summary.json	—	computed aggregates (medians, podfade, episode-survival bands, Top-100 genre mix, Apple-vs-Spotify Top-100 overlap) + methodology + caveats

CSV columns

id (Apple show id) · name · genre · episode_count (current feed track_count) · survival_band (<3 / 3-9 / 10-49 / 50+) · last_episode_date · days_since_last · podfaded (no episode in 365 days). Winners also carry launch_year only when the full archive is returned (see caveats) and launch_estimated.

Headline findings (Apple Podcasts US, June 2026)

• Discovery surfaces survivors. Only 2.2% of findable shows have fewer than 3 episodes — vs 47% of all podcasts that have published 3 or fewer (Amplifi Media, on Podcast Index data). 79.6% of findable shows have 50+ episodes; the median findable show has ~250 episodes, the median charting show ~330.
• The graveyard is invisible. Only ~9% of findable shows have gone quiet (no episode in a year); across all podcasts roughly 84–89% are inactive (no episode in 90 days; Listen Notes / Podcast Index / Apple). Charts and search can't see the dead.
• The two surfaces disagree. Apple and Spotify share only 44 of their top 100 shows — discovery is platform-specific even at the very top (see summary.json → crossPlatform).
• What's on top: of the overall Top 100, true crime (19), comedy (12) and entrepreneurship (10) dominate, with daily news, news commentary and self-improvement behind.

Method (brief)

Apple Podcasts top-charts for each of 19 genres (100 each → 1,800 unique "winners") and a keyword-search sample (24 generic terms → 2,096 shows), pulled via Crawlora's Apple Podcasts API for the US in June 2026. Per show we read the current episode count (track_count) and last-activity date (release_date).

Caveats

• Both samples are survivorship-biased — charts rank by popularity, search by relevance — so neither reaches the abandoned long tail. Every figure describes the discoverable head.
• Founding dates are omitted unless a show's full archive is returned: the episodes endpoint caps at 200, so older/larger shows' launch years aren't reliably recoverable, and we do not estimate them.
• track_count is the current feed count, so shows that prune old episodes read smaller than lifetime.
• Macro totals (~4.7M podcasts, ~44% with <3 episodes, 50–70% podfade) are cited from full-RSS-index trackers (Podnews / Listen Notes / Podcast Index), not measured here — a discovery API can't see them.
• Point-in-time US snapshot.

How to cite

This repo ships a CITATION.cff (GitHub shows a "Cite this repository" button). Plain text:

Crawlora (2026). Crawlora Podcast Gold Rush Index: the discoverable head of Apple Podcasts (v1.0.0) [Data set]. https://github.com/Crawlora-org/podcast-gold-rush-data

Data collection & ethics

Each record is public catalog metadata (a show's public episode count and last-publish date — the same a listener sees). No content is republished, no authentication bypassed, no personal data — show names, genres, episode counts, and dates only.

License

Data is licensed CC BY 4.0 — free to use, share and adapt with attribution to Crawlora (https://crawlora.net). See LICENSE. Generated with Crawlora's Apple Podcasts API.