Guest
Jacob Varghese
Jacob Varghese is Founder at Noctil, as introduced on the show.
Episodes (1)
Noctil founder Jacob Varghese explains how machine learning matches royalty data and why black box money sits unpaid when rights ownership cannot be confirmed.
Questions these episodes answer
What is black box money in the music industry?
Black box money is revenue a collecting society or licensing company has received but cannot pay out, because it cannot determine which rights holder actually owns a track or how ownership splits between multiple writers. The money sits unpaid, sometimes for years, until the territory's own rules decide what happens to it.
From: Transforming Music Metadata with AI – A New Chapter with Noctil
How does Noctil use machine learning to fix royalty matching?
Noctil trains a supervised machine learning model for each client using that client's own previously confirmed matches, then keeps learning from what the client accepts or rejects afterward. Varghese says this lifted matching accuracy by 10 to 15 percent compared with the rule-based systems the industry traditionally used, meaning more money actually reaches rights holders.
From: Transforming Music Metadata with AI – A New Chapter with Noctil
Why is matching such a core problem in music royalties?
The same song's metadata often arrives at a licensing agency many times, once from each co-writer's different publishing company, and usage reports from DSPs, radio and TV also need matching against the correct sound recording or musical work. Entity resolution across these fragmented sources decides whether royalties actually get paid out.
From: Transforming Music Metadata with AI – A New Chapter with Noctil
Why did Noctil stay bootstrapped instead of raising investment?
Noctil's first product came out of a company's request for proposal, with the MVP started in 2016, so the company was bootstrapped from the start instead of raising money on an idea. The team reinvested revenue from each new client instead, deliberately building product-market fit and an enterprise-grade platform before ever pursuing outside funding.
From: Transforming Music Metadata with AI – A New Chapter with Noctil
Why is reported streaming and performance data often unreliable for royalty matching?
Varghese says data quality varies heavily depending on where it comes from. Platforms like Apple Music and Spotify report cleanly, but reporting from YouTube, radio stations and public performance venues is often missing identifiers or complete rights information, leaving gaps he describes as a lack of data, incorrect data and insufficient identification data.
From: Transforming Music Metadata with AI – A New Chapter with Noctil