A search, discovery and measurement
tool for podcasts recently launched from Tree Goat Media using artificial intelligence (AI) and machine learning to identify moments in content.
Marbyl, the platform,
serves up a few seconds of audio based on the search query. Company executives believe the technology will support content creators by filling a void — access to valuable data and audience
metrics that can help them create more interesting content and better understand and market to a specific audience. The working prototype is available in Apple App Store and in Google
Play.
"We realized when forming the company that podcasting was more of an art form then a real business," said Michael Kakoyiannis, co-founder and CEO at Tree Goat Media, who has a long history in terrestrial and digital audio. "We wanted to increase the overall audience and incremental revenue. There are about 120 million people listening, but they are not dedicated and have difficulty finding content."
The challenge for creators is getting content heard and monetized -- as well as digging through millions of episodes available across the web, but only 1% of podcasts average about 5,000 listens per episode.
Title descriptions and metadata in podcasts have historically determined what serves up in query results, Mills said.
Tree Goat Media estimates the total available market for podcast listeners at 116 million in the United States. It also estimates there are 70 million podcast episodes, of which 69 million
are not heard.
The generation that is most likely to listen to podcasts is the Gen Z market, at 29 million. The target market in this generation is about 21 million, Kakoyiannis said.
Kakoyiannis founded the company with Sherry Mills, who serves as Tree Goat Media's chief visionary officer.
The service is free today for beta partners, but the company is developing a subscription model that will run between $30 to $100 monthly, depending on the number of shows and frequency in which the episodes are published.
The company also is looking into a revenue-generating model for podcasters -- monetizing the podcasts by adding Marbyls advertisements through ad systems like Google Ads. An
API also is in the works to integrate with other platforms.
Marbyls -- defined as a transcript of the text and audio snippet -- are created and serve up in the app’s
search results. Listeners can even create their own Marbyls from the podcasts they listen to and share those Marbyls with friends via many social-media platforms, email, and
text.
Today, there are only about 400 podcast shows, some have many episodes, available in the system, as the company looks to sign additional partner
agreements.
Text-based search is available today, but the company is looking to add audio search, which Kakoyiannis calls the “holy grail.” If generative
chat technology adds the ability for audio, that also could become a possibility.