Abstract

SBT, proposed in Decentralized Society: Finding Web3’s Soulby Vitalik inspired by the game World of Warcraft, is a publicly visible, non-transferable, but revocable type of token. The accounts bound to these SBTs are souls, which correspond to individuals, institutions, organizations, etc., and a soul can own multiple SBTs. SBTs tokenize user’s characteristics and achievements on chain with verifiable information. SBTs are able to construct native digital identities of users in Web3, ultimately fulfilling the vision of a decentralized society. SBTs are somewhat similar to DID conceptually: by functionality, they all aim to portray a digital identity with description of user characteristics by credentials or identifiable symbols. SBT is a bottom-up solution for decentralized identity.

Main types of projects based on SBT are as: certificate, identification and semi-credit-lending. Certificate type of projects issue certificates in the form of SBTs for on-chain behavior of users, representing experiences or qualifications; other projects could utilize the certificates for more accurate targeting or filtering of users in order to conduct other activities or businesses. Identification type of projects can be deemed as an advanced type that aggregates multiple credentials issued by itself: SBTs are directly treated as credentials for identification, and they contain other credentials, such as KYC. Semi-credit-lending measures credit risk of users by SBTs that describe credit risk of users and provides more capital efficient services to qualified users with favorable credits.

Current SBT segment is still in early stage, as a unified technical standard is absent so that there is still a dilemma in privacy protection, validity, and solving loopholes in some application scenarios. There may be some new application scenarios when the whole segment progresses, such as cross-platform SBT as a solution for more flexible user operations, refined SBT for more accurate marketing, and SBT-based credit lending.

This summer, Vitalik came up with a novel term “soul-bound tokens” and elaborated potential scenarios in the future. Inspired by him, a number of conceptual SBT projects sprang up. In October, Vitalik minted an NFT that could redeem a digital identity of the Republic of Palau (an SBT), and recently Huobi Global announced that digital ID of the Republic of Palau can be recognized for KYC authentication, and the topic of SBT hence became intriguing again in the crypto market yet being a questionable existence: is it just an impractical whim of a genius, an eye-catching concept, or a spark of wisdom that could light up the crypto industry to usher in a new era? This article aims to provide deeper knowledge of SBT by exploring the composition, status quo and intrinsic value of SBT.

1. The Origin and the Diffusion Curve of SBT

1.1 Etymology and definition

The term Soulbound originated from World of Warcraft, being an underlying mechanism that guarantees ownership of some powerful game props in the game: special props will be awarded to players when certain level is cleared or assigned task is completed in the game; once received, it is tied with the user and cannot be transferred or resold. The unique design of this mechanism has created unparalleled gaming experience of users, which maintained the prevalence of World of Warcraft for years.

As the NFT market is heating, NFT is widely applied on declaration of ownership on scarce digital assets for the nature of uniqueness and indivisibility, which largely matches the characteristics of rare game props in form. In addition, some other functions are embedded, such as eligibility, credential, or social status. However, because of the transferability of NFT, which especially facilitated the wealth creations of NFT in the past two years, focusing the attention of NFT more on the property of trade. In this case, once cannot tell if the holder of an NFT possesses the credentials or simply because of affordability. Therefore, Vitalik was inspired by the soul-binding mechanism and attempted to invent some new application scenarios combining the two.

In Decentralized Society: Finding Web3’s Soul, Vitalik elaborates on his vision and introduces a clear definition of a soulbound token (SBT), a publicly visible, non-transferable, but possibly revocable-by-the-issuer NFT that represents an action or experience of users on chain, which is naturally traceable as they are on blockchain. In addition, SBTs are commonly adopted to constitute commitments, credentials or affiliations of social relationship in Web3, functioning as an extensible resume by labeling characteristics and achievements of corresponding souls in the form of tokens on the blockchain with verifiable information. In this case, SBTs are able to construct a native digital identity in Web3 as a projection of actual identity from the brick-and-mortar world, fulfilling the vision of decentralized society by repeating the establishment and expansion of such social network.

1.2 New solutions for decentralized identity

Doubts may be imposed by prudent audiences: SBT seems to be highly identical to the existing DID, and SBT might just be a replica of DID. Admittedly, the two segments are highly related, and they overlap with each other in many ways. Therefore, it is necessary to clarify differences between the two on their respective value and development path.

DID is short for Decentralized Identifier, which emphasizes decentralization that aims to endow users with full control over their identity rather than being controlled by centralized platforms in the web2 world; it is also the core narrative of previous DID projects.

The standard concept of DID was proposed and elaborated by W3C in First Public Working Draft of DID in 2019. Later in 2021, Proposed Recommendation of Decentralized Identifiers (DIDs) v1.0 was published, explaining DIDs in detail: “A DID refers to any subject (e.g., a person, organization, thing, data model, abstract entity, etc.) as determined by the controller of the DID.” Evidently, DID is a symbolization of a subject, which is not specified in the text.

The emergence of DID is to outmaneuver the status quo that digital identities are controlled by centralized platforms. In other words, it envisions a digital identity that can be universally applied while returning the ownership of digital identities to users; similar to an SSN, being the unique ID that cannot be transferred, and any credentials can be affiliated. Therefore, some people prefer to interpret DID as Decentralized Identity, a global identity that contains various characteristics of a user.

Architecturally, a DID system includes DID subject, DID document, DID resolver, verifiable credentials (VCs), verifiable data registry and other components.

In the set of recommended standards of W3C, DID uses verifiable credential (VC) for verification, which is generated by a third party (Issuer) to prove certain attributes of a user. For privacy protection, information regarding user identity is stored off-chain, and only the Issuer has access to this data; ZKP can be generated by the Issuer afterwards and uploaded on chain, which enables the identity verification without breaching privacy. This allows proving the user’s identity without compromising privacy. VCs can be stored in wallets for repeat usage similar to other crypto assets.

Thus, VC and SBT are different despite the fact that they are all solutions of DID. The characteristics are summarized in the table below:

It is clarified at this point that the confusion of DID starts from the pun of DID: when an identifier is engaged to represent normal credentials in a DID project, especially when it is non-transferrable, it serves the same function as an SBT, whereas it is the soul bound by tokens when representing identification.

To sum it up, the main differences between SBT and DID are concluded: (1) SBT is a bottom-up solution for identity with more emphasis on the process: what kind of certificate to issue and what kind of capability and qualification to incorporate in the form of non-transferable NFT; while DID is the result of on-chain identity that it sets foot on aggregating multiple labels, addresses, even multi-chain addresses to synthesize an icon for global recognition. (2) SBT aims to construct a solution based on addresses, while DID, to some extent, is dedicated to build a solution based on aggregation of addresses.

In the future, the ideal state may be that: each user corresponds to a unique decentralized identity, and many non-transferable credentials that represent the user’s experience, skills and qualifications in different domains reside under the identity; it may be a single address or a collection of multiple addresses.

Thus, the competitive landscape of the credential and identity sector is naturally clear. Any organization can issue credentials to users, which is a low-barrier business, but it takes several large, influential organizations to aggregate credentials into an identity that can be recognized in various applications or even multiple networks, which is an extremely valuable business with a relatively high threshold.

2. Typical SBT-based applications

Vitalik’s paper lists eight SBT-based application scenarios, such as reputation provisioning, SBT community recovery, soul airdropping, soul governance, and soul lending. Therefore, without further elaboration, only a categorization of his ideas is made here:

1. SBT as a certificate to establish identity.

2. Enabling more advanced functions with identities built on SBT, such as credit lending, more democratic governance, recovery of community keys, etc.

3. SBT as a vehicle for special assets that are not transferable.

The first 2 types of applications are already on the market; SBT and the associated functions of identity will be viewed from representative projects.

2.1 SBT as certificate

Due to the non-transferable nature of SBT, it is suitable to demonstrate qualifications and credentials received by the holder rather than the asset. Currently, certificate type of projects are the most commonly seen SBT-based projects, including POAP, Project Galaxy and RabbitHole, issuing certificates to on-chain or off-chain user behaviors, which represent experience or qualifications of users. Projects that have partnership with them are able to conduct other forms of activities or businesses by more accurate targeting or filtering of users. An elaboration will be given taking RabbitHole as an example in order to clarify the business model of this type of projects. (Although certificates in some projects are transferable NFTs, the design is all the same.)

RabbitHole is a class for beginners with rewards: it engages users by posting tasks with bounties; the bounties are collectible after completion of study or use of a certain Web3 application, including customized NFTs or tokens issued by relevant parties.

In order to be eligible to participate in the activities with bounties, users first have to complete the basic course in RabbitHole’s Bootcamp to obtain a certificate in the form of SBT. For example, in order to obtain the Intro to L2 certificate, users must complete 3 small tasks, which are transferring ETH to the Optimism network by the cross-chain bridge of Hop, exchanging ETH to OP on Uniswap, and minting an NFT on Quix at the end.

Once a certificate is obtained, it serves as a prerequisite that users could participate in more rewarding tasks. For instance, Uniswap has once partnered with RabbitHole to release Quest of Learn L2s with Uniswap, which requires a Bright ID (real person certification) and an Intro to DeFi certificate issued by RabbitHole itself (there was no Intro to L2 certificate at the time). There were 4 small tasks, which were to provide liquidity and trade on Uniswap deployed on Arbitrum and Optimism, respectively. After completing these 4 tasks, users were rewarded with 2 UNI tokens and an NFT.

Certificate type of projects are somewhat similar to data service providers on chain that collaborative projects can directly query the certificates instead of accessing the user interaction records themselves. It saves time consumption and economic cost of processing data for the partner projects. The profit model is also to charge a service fee to the partner project.

RabbitHole, for example, will work with the partner program to run the entire campaign, including setting the agenda of the campaign, task setup, rewards, estimated number of participants, community and media outreach, tutorials, etc. RabbitHole charges the partner 20% of the campaign reward amount as a service fee, and locks for 1 year. In this case, the service fee is more of an operational cost with partially being the cost to cover data usage.

The key point of certificate type of projects relies on recognitions on certificates from the partners and proceeding activities afterwards. On the one hand, the income is heavily dependent on the overall activity of the market. When the market is active and all kinds of projects have budgets in user acquisition, the income is somehow guaranteed. But when the market is cold, the income might pale. On the other hand, the revenue relies on the degree of influence on the market. If a program has been running for a long time with a large and active community, the actual influence would be more materialized and realistic as more attention are drawn from users, therefore, the certificates might be more recognizable.

2.2 SBT as identification

In this regime, SBT functions as same as DID. Representative projects are BAB and digital identity of the Republic of Palau.

BAB is the first soul-bound token based on BNB Chain launched by Binance in September 2022, which is the concrete landing of the SBT concept in the identity sector. BAB is bundled with Binance account following ERC-721 standard; only users who have passed KYC authentication on this trading platform are eligible to apply. Moreover, only 1 can be minted associate with a certain account; the cost for minting is 1 BUSD. After minting, it is not transferable and can only be revoked.

Although BAB has been launched for a short time, some opportunities are foreseeable after the bona-fide implementation.

(1) For the holder, although the non-transferable nature of BAB does not promise direct financial value, it represents a verifiable on-chain credential corresponding to the address. As a holder per se, it is already a proof that the user has successfully passed the KYC process of Binance so that the holder address will not miss any airdrops due to the reason of being labeled as a bot.

(2) For Binance, BAB cannot only generate considerable on-chain revenue by minting, but more verifications can be conducted to testify the minting time of BAB: not only can loyalty of users be confirmed, more customized service can also be provided by thorough analysis of on-chain behaviors of BAB holders.

(3) For ecological projects, whether by airdrop or other incentive methods, a bona-fide growth in community building or the overall development can be achieved if the real demand of BAB holders is captured by integration of BAB with the ecosystem of the project. For this reason, 14 ecological projects of BNC Chain have announced incentive plans on BAB holders in less than a week after the launch of BAB.

The digital ID of the Republic of Palau is also an identity type of SBT. The application of rental requires an annual fee of $248, and photo of legal ID, portrait, height and weight must be uploaded. Designated application scenarios include address verification, KYC authentication, access to digital banks, legal on-chain signatures, digital company registration services, and other potential products such as asset ownership and insurance. Officials also specifically state that it is legitimate in any application that requires a government-issued ID.

The identity type of SBTs can be treated as an advanced form that aggregates multiple certificates issued by themselves. BAB per se is an identity, which contains a certificate of KYC on Binance, and it may contain more information in the future, such as interactions and positions on BNB Chain. A certificate of KYC is the portal that an actual user in the real world could enter cyber world, which average blockchain projects cannot afford to issue for their users, only large institutions, such as exchanges and custodians, are capable to do so. Well-known exchanges have relatively large traffic, which serves as a privilege in issuing SBTs. In the future, exchanges may be seen partnering with L1 chains that are closely related and issuing on-chain identifications respectively, which eliminates the interruptions of bots on crypto applications and ensure a more comfortable environment and better user experience. If one of the on-chain identifications as SBTs issued by an exchange could receive mass consensus, even across chains, incremental user base and capital influx will be reciprocated by the SBTs.

2.3 Semi-credit-lending SBTs

Since the DeFi Summer in 2020, credit lending has seen a boom, but the vast majority of them are collateralized lending programs. Collateralized lending is consistent with the no-permit and no-threshold nature of the crypto world, but it is inherently disadvantageous in the utilization rate of capital funds. Credit lending projects, such as TrueFi, Goldfinch, and Maple, have started to become outstanding by providing credit loans to crypto companies that have completed off-chain KYC authentication, expanded the market size of credit lending and improved the efficiency of capital utilization for institutional traders and crypto companies. But a loan on credit for average users is still infeasible, for one of the reasons being the lack of authentication for on-chain credit and identification impedes lenders from measuring the credit risk of borrowers. Vitalik proposed the concept of “soul lending” in his paper as a response to the issue with the intention that loans can be issued without collateral but solely on credit based on SBTs as proof of reputation and credit record.

Although no projects so far have provided loans to average users completely based on SBTs, semi-credit-lending programs have been spotted that a user’s loan-to-value (LTV) is adjusted according to the level of credit risk measured by corresponding on-chain behaviors. This type of project is classified as the second major application scenario based on SBT.A typical example of ARCx is introduced in detail in the next to elaborate how SBT could facilitate the operation of this type of project.

First of all, it should be clarified that the word “SBT” does not appear explicitly in the official documentation of ARCx, but the practice is similar to SBT, and it is considered equivalent of SBT in this article.

ARCx has introduced a credit score similar to SBT as DeFi Credit Score, which rates users based on their behaviors on ARCx and other lending platforms, ranging from 0 to 999. ARCx has a 3-tier lending pool, and the higher the credit score, the higher the user’s maximum LTV, i.e., the more money they can lend with the same value of assets in staking.

ARCx’s credit score is calculated based on the record of interaction between lending protocols and the on-chain address (Polygon and Ethereum), and the formula is fully disclosed to the public. The credit score consists of 3 major components: Daily Score Reward, Survival Score Reward and Liquidation Penalty. The Daily Score Reward evaluates borrow usage (i.e., current LTV as a percentage of max LTV) on ARCx Credit vaults over the prior 120 days relative to a “responsible” borrower archetype and rewards points according to a “Rewards Curve” on a daily basis. ARCx found, based on research with Compound, that most of the skilled users have a utilization rate of borrowings around 60%, and an optimal utilization rate on capital can be achieved on the basis of risk management. As a result, when users have a utilization rate of 60% on borrowings, more points will be accumulated; any deviations from this percentage, either higher or lower, will result in a lower rate of growth on score. Both the Survival Score Reward and the Liquidation Penalty measure the risk of being liquidated, which assess all behaviors across various chains. If a user is immune from a large scale of liquidations, the user is implied to have higher risk management skills, which deserves more scores; otherwise, if the market is smooth in whole, most users will not be exposed to be liquidated, therefore no exemplar could be outstanding to receive a higher score. The Liquidation Penalty is simpler: if a user has been liquidated within 120 days, 250 points will be deducted directly. As the total score is only 999 points, one recent liquidation would be enough to reject a user out of the highest LTV pool of ARCx.

The seemingly semi-credit-lending programs like ARCx actually still operate within the security line of collateralized lending: descriptive SBTs are in use to measure the credit risk of users, and more capital efficient service can be provided to users with good credit; it is line with the status quo of the crypto industry, and it is an effective complement to traditional over-collateralized lending.

3. Challenges in the SBT Sector

Although various SBT projects have emerged after Vitalik’s paper, difficulties lie on the development path that prevent SBT from further market recognition.

3.1 No unified technical standard

Issuing tokens on Ethereum is essentially achieved by creating a new smart contract ledger under a certain address, on which all the transactions of users interacting with it are recorded. This method of issuance has resulted in several unified standards, such as ERC-20 and ERC-721, and developers could construct the corresponding type of new token simply following the rules step by step.

SBT presents a new token type to Web3, but most of the existing SBTs are slightly modified from ERC-721, and the non-tradable, revocable and verifiable features do not yet have a dedicated technical interface in existing APIs. The absence of a formally defined interface not only increases the security audit burden before deployment of the contract code, but may also trigger unexpected attacks at runtime. On the other hand, SBTs with different token standards create additional workload for other projects to integrate, which hinders mass adoption and further recognition. The key to solve this issue is relatively simple: to establish a unified technical standard with universal consensus.

In addition, SBT is also programmable that it is promising to combine it with dynamic NFT to introduce oracles for dynamic updates on properties, which should be included in the unified technical standard.

3.2 The dilemma between privacy protection and verifiability

For SBTs, privacy protection and verifiability are difficult to balance and reconcile. On the one hand, SBT is generated based on on-chain behavior of users, and it is inherently verifiable with credibility. On the other hand, SBT is a summary of scattered user behaviors with labels; it aggravates the risk of privacy exposure that on-chain analyzing tools could easily grasp behaviors and preferences of users. The dilemma is most burdensome for those projects that require KYC authentications: SBTs will more or less lack credibility and limit the scope of application scenarios if the KYC cannot be verified by third parties; however, if KYC information of users were exposed to others for verification purpose, it will inevitably lead to information leakage, including information of users in real life, and unnecessary trouble can be triggered, which will also undermine the brand image of the project and the affiliated SBTs. Therefore, it remains a challenge for every SBT project to obtain as much real user data as possible to build a proof with mass consensus while ensuring user privacy. Technologies, such as zero-knowledge proof and multi-party secure computing (MPC), may play significant roles in balancing privacy and verifiability.

3.3 Loopholes in some application scenarios

When SBTs are functioning in more democratic governance, current voting mechanism by tokens, on which DAOs are highly relied, often encounters sybil attacks and vote manipulation. The emergence of SBT could screen out bot accounts and ineligible voters that do not belong to the community, alleviating the problem of unfair voting when it comes to DAO governance. However, problems of voting that exist in reality also live in the blockchain world, such as bribery. With SBT, souls with high voting weight will be more exposed to those with evil intent, and bribery clearly contradicts the original intent of SBT for fair community governance; this subject matter goes beyond computer science to the research area of sociology.

4. An Outlook: Potential Application Scenarios of SBT

SBT and on-chain identity are currently in an early phase, and many are still concentrated on how to issue SBTs. Assuming a more uniform technical standard that leverages information while privacy is guaranteed, the role of SBTs will not be limited to soul airdrops and governance.

4.1 More flexible user operations with cross-chain SBTs

Current Internet applications can enable cross-application login based on Google, Facebook, WeChat and other accounts, but users will be asked to fill out the same information for registration or go through a process for skill measurement when accessing similar types of applications. From this perspective, SBT is generated on the blockchain, and the information is openly traceable, which is conducive to cross-application logins and sharing across platforms. However, the scope of SBT in application is currently limited to a certain platform; user experience will be vastly enhanced if SBTs can cross. For example, a player of FPS is in the market to choose a game to play for long term, an SBT from previously played game A could directly send the player to advanced level battlefield and eliminate the process of leveling up and accumulate points for ranking. In addition, game A could receive a fee from game B as user acquisition cost to make up the loss if the player chooses game A over game B eventually.

4.2 More accurate marketing by refined SBT

Current SBTs are rather shallow that they measure user behaviors from merely several dimensions, including interactions with certain applications, or used to receive airdrops from certain projects, etc. Although some projects have recorded and analyzed on a deeper layer of user behaviors, SBT must be continuously refined to reflect more aspects of user behaviors, including but not limited to the amount in user operations, risk aversion, active time period, length of use. Especially when combined with information from exchanges, such as trading frequency and preferences of strategy, news feed of users in medias, etc., SBT will be capable of portraying a more vivid appearance of users. Based on more refined user profiling, more accurate marketing campaigns could be conducted and the ideal potential users could be located at a lower cost.

4.3 Facilitated credit lending under mass adoption of SBTs

Credit lending in reality is not successful without strict enforcement of those who default. If a borrower defaults, the credit report will be ruined that many privileges will be forfeited. For example, a court order may impose restrictions on consumptions of defaulters that they cannot travel on airplanes, soft sleepers on trains, second class or higher cabins on ships, high spending in star hotels, restaurants, night clubs, golf courses, or purchase real estate or build, expand or renovate high-grade houses, or rent offices in luxury buildings, hotels and apartments, etc. Relevant government departments, financial and regulatory institutions, industry associations will be notified, more restrictions will be added on economic activities to implement joint credit discipline in accordance with laws, regulations and relevant provisions, including bans in the following types of restrictions: government procurement, administrative approval, bidding, government aid, financing and borrowing on credit, qualification approval, market entry, corporate governance, issuance of debt or securities, employment as directors and supervisors as senior management and other aspects.

It is not impossible to conduct credit lending if defaulters could be labeled and restricted on on-chain activities as what happens in the real life. If a widely recognized SBTs can be issued by a league formed by major credit lending protocols, with the joint cooperation of other protocols and exchanges that restrictions on defaulters are collaboratively enforced, such as refusal of service, the cost for default will be catastrophic. Furthermore, if the on-chain defaults are also synced to financial institutions in real life and enforced as strict as that for traditional financial activities, users will weigh meticulously before even thinking of default, and complete credit lending could become feasible.