Data and coverage

What SponsorHound's data is, where it comes from, and what it can't tell you.

Where it comes from

What's in it

BrandsAbout 17,800 brands seen in sponsor reads, about 12,700 of them active in the last 90 days.
ShowsAbout 10,200 podcasts and newsletters with at least one sponsor read, mostly in English.
EvidenceFor each sponsor read: the brand, the show, the episode or post, its date, and a quote of at most 240 characters with a link to the source.
FreshnessRefreshed about weekly. Every answer carries data_version and as_of (the date of the newest episode in the data).

Figures are approximate and change with each refresh.

How sponsor reads are found

Sponsor mentions are extracted from feed text by rules ("brought to you by", promo codes, sponsor sections and similar patterns). Brand identities are grouped by web domain or name. On an AI-labelled test set the extraction is right about 91% of the time; a human-checked sample is in progress. Expect some misses and some things that aren't brands.

Shows are matched to yours by the meaning of their titles and descriptions, so a good one-to-three sentence description of your topic, audience and language gives the best results.

What it can't tell you

An ad read shows that a brand advertised on a show on a date. It does not prove that the brand paid, what it paid, its budget, whether your audience fits, or that it is open to new deals.

Corrections and removal

If a result is wrong (not a brand, the wrong brand, not a sponsor, stale, duplicate), tell us at [email protected] with the data_version and the brand_id or observation_id. Brands, shows and publishers can ask to be removed: see opt-out and removal.