When music metadata meets blockchain governance
Mid-stream reflections on what web3 means for the music industry, and why the music industry matters for web3.
“[I]t’s a bit of a kick in the pants for me, having this discussion with you, because obviously metadata is being thought of in this quite professional, formal way […] as if it’s an NFT” — musician to interviewer Merida Sussex.
The quote above is an “aha” moment in which a music artist grasps the importance of music metadata. She likens it to a non-fungible token (NFT) — a blockchain token standard that represents a one-of-a-kind or one-of-a-series token, which can be owned and transferred between digital wallets. An NFT contains metadata which, thanks to blockchain technology, is an agreed-upon fact.
The quote encapsulates a fundamental shift in thinking about metadata that is occurring at the edges of the music industry. The shift is not to turn all music into NFTs, but to find more “formal” and coordinated ways to empower artists and those who support them to manage metadata.
Music industry metadata is data about tracks and their creators. As the UK’s Inquiry into streaming services found, there is consensus across the music industry “that issues with the metadata are a significant challenge to efficient and correct rightsholder remuneration” (House of Commons 2021, 50).
The music industry is complex and discrepancies and gaps in metadata are common. When these occur, unattributable royalties accrue in the accounts of collecting societies and digital service providers (DSPs), which are eventually paid out in ways that favour the major players, with some artists receiving nothing while others (or their labels) get paid twice. These deals and the pricing structures of streaming services are not my focus here (see Hesmondhalgh, Osborne, Sun & Barr 2021).
The research I am conducting on this problem is best described as an ethnography of entrepreneurial discovery; I am an ethnographer observing the efforts of an entrepreneur who has been attempting to understand and solve the music metadata problem. Entrepreneurial discovery is difficult because it aims to find signal and value where there is noise and confusion — knowledge amidst uncertainty — typically of a market process.
Merida Sussex has been involved in the music industry as a successful musician (Paradise Motel), the founder of an independent record label (Stolen Recordings) and a board member of an international industry association (Merlin). Merida and her chief collaborator Peter Harris initially set out to create a single source of truth, “a blockchain solution to registering and communicating music rights and metadata”, which they named Envoke. As part of her entrepreneurial process, Merida is interviewing people who create music metadata, including musicians and those working for labels, publishers, distributors and collecting societies (22 to date). I have been attending these interviews as well as some meetings between Merida and her collaborators. My study is following the experts, as Latour suggests, in order to understand how coordination problems are exacerbated or resolved by technology platforms, with a particular focus on blockchain.
The discovery process that I am observing involves not a market process but a new type of governance process. It has consequences for web3 approaches to music metadata failures. I will return to this point after describing how we got to it.
The problems with music metadata
The output of the music industry is music but how payments flow — who gets paid and when — is determined by metadata. Here’s what Merida and I derived from the interviews:
1. Metadata is jointly produced and changes over time.
This is the unavoidable reality of metadata. Metadata generation begins in the studio through negotiation (for instance over splits) and multiple people within a record label may be involved in producing it (from 2–15 in our group). Existing works require different processes. The historical separation of performance/recording and publishing means different entities can have responsibility for different segments of metadata.
2. Even for those actors with good intentions (including independent record labels), multiple points of failure are occurring.
Disconnects and gaps can occur at the creation of a song (getting artists/producers/managers to identify contributors to a track, for instance), onboarding an existing catalogue, having multiple processes for communicating the same data (resulting in duplicates and mistakes), or how DSPs receive/use information and from whom. For instance, labels find workarounds to bump an artist’s track on a platform, which can result in problems if that metadata is then transported onto another platform. We heard that Discogs, whereby fans contribute information about songs, can be the most accurate and accessible sources of metadata available to musicians and their representatives.
3. The systems used to store metadata don’t talk to each other: Registries containing ISRC numbers and accompanying attributes are not publicly accessible. Independent labels produce or procure bespoke systems and spreadsheets to manage metadata, but often still have to recreate metadata for different actors in the system (such as collecting societies and DSPs). Some choose to outsource metadata creation to aggregators and distributors who do not always give that metadata back when the interaction or contract ends.
4. Knowledge of how metadata should best be handled resides with labels, technology companies, DSPs etc. and not with artists. This creates opportunities for exploitation.
We learnt that there are multiple parties involved in the creation of metadata (within and across companies) and no unified agreement on how to implement or work with the existing standards. A comprehensive response, such as the DDEX standards, can be too complex for an unsigned musician or smaller label to engage with. This favours larger companies with greater administrative capacity.
Knowledge about why metadata matters does not reach artists, and often neither does the metadata itself, which results in financial disempowerment. The conversations between Merida and artists typically ended up with her answering questions as they began to grasp the importance of metadata and how little they knew.
What can be done?
The problems outlined above are not intractable, but solutions are not occurring through market forces. Attempts to bring metadata into a unified global repository have failed because those who are profiting from current arrangements are not incentivised to cooperate.
The other option is to assist those who are disadvantaged by current processes to generate better metadata and, when problems arise in the metadata (as they inevitably will), to have a means to address them. Merida has identified various options that could improve metadata that do not require a single platform for all, including: systems that enable you to export data in whatever format the receiver requires it; inviting artists to add data where it’s stored rather than sending information in emails; tools that tell you how complete your metadata is; encouraging shared and open source platforms and use of APIs; providing receipts for metadata; algorithms that search existing databases for inconsistencies (but do not privilege a large label’s rights claim over a smaller label or privilege new information over old). Some of these are already being created in-house by independent labels who hired good IT talent. The next level challenge is to get everyone doing it, which is where blockchains come in.
What does this tell us about Web3 entrepreneurship?
Blockchain music platforms are, in theory, designed to make it easier for artists to receive royalties by reducing the need for intermediaries and automating payments. These platforms tend to see their main problem as adoption. Getting people to use these platforms is hard partly because of all the administrative complexities outlined above. As one interviewee noted, when they use Audius they enter metadata again and in the form that Audius requires. The same interviewee expressed concern that Audius was only rewarding early investors and early adopters. As this interviewee astutely observed, blockchain music systems don’t work with the existing order but attempt to impose an external order on a complex industry. They can also create new problematic power dynamics and inequities.
Envoke commenced not from the idea of a royalties payments platform but from the notion of a publicly accessible ledger that could be collectively maintained. After talking to music industry insiders about their metadata practices, governance, not adoption, became the focus.
Governance here is not a political concept that involves voting or power blocks, but governance in the corporate governance sense of minimising opportunism. Economists call this an agency problem, whereby governance is used by outsiders to constrain the power of insiders.
Corporate governance is also necessary for renegotiation. In this instance, the renegotiation of metadata can be seen as the ingredients of contracts; contracts that need to change when new information becomes available. One example of this is when an artist’s contract with a label expires but the metadata is not updated and the label continues to be paid (opportunism). Or it might be due to conflicts in metadata requiring dispute resolution (maladaptation).
Web3 entrepreneurism grapples with corporate governance problems frequently, typically seeking to constrain the problematic accumulation of power in a decentralised system. We see this in the creation of decentralised autonomous organisations (DAOs) to ensure that teams administering decentralised financial applications (DeFi) do not pursue their own interests, or efforts to constrain the plutocratic forces in one-token-one-vote systems. As Davidson and Potts (2022) argue, these are agency problems for which corporate governance is appropriate, rather than political governance mechanisms, which exists to achieve legitimacy for rule/coercive acts.
Envoke’s discovery process has led them away from seeing blockchain as “eliminating the intermediaries” to creating new tools for overcoming coordination problems. The major labels have found ways to extract value from system failures (intentional or otherwise) which make it hard for musicians and smaller players to get metadata right. Once we diagnose the problem as a metadata and system coordination failure that is exploitable with enough capital and power, then creating blockchain platforms that favour new forms of capital and power is not the path forward. The correct response is governance.
Governance tools may include algorithms for recording and rewarding metadata contributions, dispute resolution “courts” to correct inconsistencies in metadata, or DAOs for artist collectives in which tasks can be allocated and paid. These tools are experimental and have their own problems but the music industry has real reasons to work on them and get them right. As the quote at the start of this article demonstrates, NFTs in particular are changing how artists perceive ownership, rights and payments. They turn metadata into a tangible object that artists feel a direct relationship with from the point of creation. If a novel and highly technical token standard (NFTs) can do that, it’s safe to assume there’s more to come.
What will the music blockchain future look like? If this one entrepreneur is right then we can predict that music blockchain projects that ignore the realities and power dynamics of metadata will continue to struggle with adoption issues. The blockchain platforms that succeed will transport existing metadata (instead of making new admin work), provide a means for artists and those who support them to align on facts, and incentivise keeping metadata up-to-date. Removing the capacity for opportunism means making the messiness of metadata visible and resolvable. In other words, making it governable.
This work-in-progress article was written for a presentation at the International Communications Association conference in Paris, May 2022. I am a Professor at RMIT, working across the RMIT Blockchain Innovation Hub and the Digital Ethnography Research Centre. I am also an Associate Investigator (AI 🤖) of the ARC Centre of Excellence for Automated Decision-Making and Society. I acknowledge the support of the Australian Research Council, FT19010372. Thanks to Merida Sussex for allowing me to tag along on her own research journey.