The Matching Conundrum

Matching recordings to their underlying compositions is an intricately complex music industry puzzle. While it's something that inherently seems like it should be fairly straight forward, this process poses significant challenges.

The ability to match is only as good as the data collected to facilitate.

If you're a new subscriber, I encourage you to go back and read some of my prior articles to set the scene for what I'm addressing here. In fact, it's taken me far longer than I anticipated to write this article, as it's hard not to get too far into the weeds!

The importance of matching

A song has rights owners for both the recording (artist) and the composition (composer/lyricist). In order to ensure that artists, writers, and rights holders receive proper credit and compensation, the two components of the song must be married together.

While it would be ideal for all of the metadata to be captured at the inception of the song, the recording may be released into the market without confirmed publishing ownership. This leads to the need for post-release matching, usually once royalties have already been accrued.

This process isn't straight-forward due to various factors:

• a composition may have been recorded many times
• a recording may contain more than one composition
• many songs have similar titles or composer names
• rights owners may have different standards for data capture
• misspellings, inconsistencies, or poor data quality often go unchecked
• there may not be enough information to confidently make a match at all

The matching process

The matching is (usually) done by some combination of steps:

1. Reference an existing database of matches to find assets that already exist
2. Conduct (proprietary) fuzzy matching to find matches that are close enough with a reasonable level of confidence
3. Review manually when there isn't enough metadata to match automatically

Ideally, successful matches contribute to an evolving database that refines future matching efforts. This cyclical process, when supported by learning models, aims to continuously improve accuracy.

The development of music recognition technologies, enhanced matching algorithms, and standardized metadata systems have allowed for improvements in this area. However, achieving a perfect match between sound recordings and musical works remains a complex task that, all too often, requires manual manpower to review.

What happens when matches aren't made?

Industry-wide, there are millions of lines of unmatched recordings where the underlying composition hasn't been confirmed. If these are royalty bearing, those royalties go into suspense or escrow accounts pending identification of rights owners in order to issue payment.

Here in the U.S., record labels and publishers work to clear unmatched items, but much of the work is done by the Mechanical Licensing Collective (The MLC) as a central entity for the industry under the Music Modernization Act. The MLC holds unmatched/unclaimed royalties for a minimum of 3 years while attempting to find the right owners. Once those efforts are exhausted, royalties are paid "to music publishers and self-administered songwriters using a market share/activity-based formula."

Three years may seem like a long time to confirm and resolve ownership, but there are constant additions to the unmatched pool. To reduce the scale of the problem, the unmatched and unclaimed must be addressed faster than new assets are added. To do this requires a proactive effort to capture and confirm accurate composition information as close to release as possible, thus preventing recordings from ever getting to the unmatched state.

Looking forward: Technology & Collaboration

With advancements in technology, and music companies increasingly taking a "data-forward" approach, matching efforts are improving. Delivery of complete and high-quality data reduces the need for manual review, streamlining the process and reducing overhead.

The more data points that are available to match against, the higher chance of confidently identifying the correct combination. Consider unique identifiers above and beyond title, artist, and composer to enhance accuracy:

• recording codes (ISRC)
• work codes (ISWC, society Work codes)
• person identifiers (ISNI and IPI)
• product codes (Selection Number or UPC)

Collaborative data sharing efforts, such as the Verifi Rights Data Alliance (VRDA) or the MLC Supplemental Matching Network further aid with improved identification and resolution of discrepancies.

The rapidly growing number of new recordings being released into the market place tell us that these collaborative efforts and advanced technology will be vital for accurate and efficient matches.

Whether you're an established artist or a developing writer, paying attention to these critical data elements could be the difference between getting paid on time and never getting paid at all.

Natalie Jacobs is the founder of Equalizer Consulting, specializing in helping music companies improve their operations:

• Optimized workflow solutions
• Best practices in music data management
• Data reconciliation, harmonization, and documentation
• Comprehensive training and education programs

For a free 30-minute consultation, please use this link to schedule.