Most discussions on NFT architectures tend to focus on bringing efficiency, cost-effectiveness and scalability to how the NFT tokens are minted and exchanged. And these discussions are essential for driving NFT practicality. But eventually, more discussions and analyses of metadata architecture are required to expand NFT utility.
NFTs are on the verge of escaping from the confines of digital baseball cards and play-to-earn games to become mainstream enterprise technology. By 2024, Gartner predicts that 50% of publicly listed companies will have some sort of NFT underpinning their brand or digital ecosystem presence. NFTs will become a powerful marketing tool to underpin digital ecosystems’ effects and accelerate enterprise valuations.
Gartner goes on to describe NFTs as becoming a way to “leverage hyper-tokenization to grow business models exponentially.” This is a bit of a mouthful, but it points to the importance of thinking about trading tokens more efficiently and connecting to other things in the real world that ordinary people care about. In Gartner’s case, they are referring to an enterprise’s “tokenization strategy to capture their customers’ digital requirements and to secure future revenue streams.”
An expanded way of thinking about NFT gamification will propel an enterprise into the top ten highest valued companies by 2026, Gartner predicts. But for this to occur, developers will need to consider the importance of tools and architectures for NFT metadata management as a forethought rather than an afterthought.
In some respects, the whole definition of metadata is reversed when developing NFT applications. This has some value, which we will get to in a minute. But as a result, it also glosses over how the utility of NFTs can be expanded for more mainstream use.
In the NFT world, the primary “data” is considered the token whose ownership is attributed to various addresses on the blockchain. This ownership can be transferred by various processes characterized in standards like EIP-721 in the Ethereum universe. Other standards like EIP-1155 describe ways of bundling and trading collections of NFTs more efficiently. There is even work like EIP-2981 for sharing royalty data across NFT exchanges, although it leaves the actual payment processes to be sorted out by the respective exchanges.
The metadata in an NFT is the actual content and the description about the content described in the contract stored on the blockchain. This is kind of backward to the way people usually think about metadata. In an MP3, the music is the main event and information, like the artist, band, album, and beats per minute, is an appendix. In some DRM systems, ownership information is also encoded into the metadata.
In an NFT, the metadata is pointed to by a cryptographically uniform resource identifier (URI) referenced in the token. This URI could point to a website maintained by the NFT creator or a third party, an IPFS address or other data location that the NFT app can resolve. In almost all cases, the metadata lives outside of the actual NFT. The rare exceptions are a few ASCII artworks small enough to encode directly in a blockchain.
The Value of Technical Efficiency
As with many things in life, there can be tremendous value in looking at things differently. Focusing on the ownership and tradability of NFTs as the “data” can help bring efficiency to NFT marketplace implementations. Efforts to sort out the tradeoffs in various approaches for minting and trading NFTs will lead to technical scalability.
For example, Ethereum still leads the pack when it comes to NFT market share. A bevy of so-called Ethereum killers like Cardano, Tron, and Solano promise improved performance, which need to be weighed against differences in consensus mechanisms and the size of the community. Then there are the new Level-2 networks that take advantage of new trading features being implemented into blockchain main nets for rolling up multiple transactions using zero-knowledge proofs.
Others are exploring how sidechains for migrating contracts back and forth across connected blockchains. This is great in theory, but then messy details sometimes emerge in how these work in practice. The recent $600 million breach of the Poly Network is a reminder that sidechains can introduce security vulnerabilities missed by even the best teams.
The competition between these various approaches will help NFTs to become more practical and easier to implement for a much wider audience
Reframing the Main Event
Developers will also need to consider NFT metadata as the main event in some cases to expand the utility of NFTs to support more use cases. The first use of metadata-like things dates to the Library of Alexandria in Egypt around 280 B.C. Scroll tags indicated the title, author, and content of scrolls, making it easier to file and find specific scrolls. These were eventually formalized into card catalogs and the Dewey Decimal System for organizing books by category.
In the 1960s, academics first began using the term metadata for describing different approaches for cross-referencing data about other data stored in computer systems. This early research led to standards for tagging pictures, documents and eventually all types of enterprise data. Nowadays, metadata management (MDM) tools and systems play a crucial role in helping enterprises make sense of the tsunami of data arising from enterprise applications, the Internet of Things, and third-party sources.
In 2004, The National Information Standards Organization (NISO) formalized these ideas into three main categories:
- Descriptive metadata describes a resource for discovery and identification, such as the title, author and contents.
- Structural metadata describes how multiple elements can be put together, such as how multiple tracks can be organized into a song.
- Administrative metadata helps manage the resource, such as when it was created, who owns it and how it can be accessed.
The NISO paper argued that “metadata is key to ensuring that resources will survive and continue to be accessible into the future.” It is a bit dated, but some of the core ideas are still relevant, particularly as developers mull over ways to extend the utility of NFT applications.
Expanding NFT Utility
Some NFT projects never planned for survival, much less thriving. And creators and buyers of NFTs are starting to discover that NFTs are not as solid as imagined. For example, one artist going by the name Neitherconfirm deliberately changed the images pointed to after the fact to point to a critical limitation in how these systems are built today.
Here are four considerations for implementing NFT metadata that could extend the utility of NFTs in the long run:
Is the metadata provable? This relates to thinking through the safeguards to ensure that what the NFT URI points to is what everyone agrees upon when it changes hands. At the minimum, this includes generating a hash that points to an address that can be trusted, such as a reputable organization’s website or IPFS address with a well-established chain of trust. A second level of provability is to consider not just a hash of the URI but also the content. That way, if the content at the address somehow changes, then everyone would know.
Is the metadata retrievable? This makes it easier to find the metadata relating to a given NFT across a range of marketplaces and search engines. The NFT community could take advantage of early efforts to improve search. A collaboration by Google, Microsoft and Yahoo led to the creation of Schema.org to promote schemas for structured data on the internet. This could make it easier to find and reuse NFTs and their derivatives across multiple marketplaces and dApps.
Is the metadata traceably mutable? The NFT token itself is immutable by design, but many applications can benefit when the metadata can evolve with guardrails. For example, an NFT representing an in-game item like a sword might accrue additional powers with successful quests spelled out in the game. Today, this logic is often custom developed by each game developer. Down the road, decentralized databases that use blockchains and other distributed ledger technologies could include database primitives that simplify this development by storing the complete history of updates in conjunction with additional metadata.
Is NFT metadata combinable? Today each NFT is created and sold as a one-off item. It is also possible to organize and sell them as collections to simplify trades using standards like EIP-2981. Down the road, more nuanced ways of representing the metadata could make it easier to combine multiple NFTs into new assets with licensing and royalty management logic. For example, an artist could create a mashup of NFTs representing lyrics, guitar riffs, and baselines to create a new derivative NFT song that rewards all the artists with royalties.
The NFTs on the marketplace today are in about the same place as static web pages were in the early days of the internet. When early leaders like Netscape focused on building a better app, the emerging Web 2.0 leaders like Google focused on surfacing and contextualizing the information about how sites linked to each other. Down the road, we may find that the new leaders in the NFT space similarly discover new ways to contextualize, reuse and combine the metadata buried within the NFT ecosystem.