Protocol Analysis #2
This is the second article in our Protocol Analysis series. In this series, we use our investment template outlined in How to Analyse a Web3 Protocol to fundamentally and critically analyse a variety of crypto protocols, aiming to separate the ‘real’ from the hype. Let’s dive in!
Protocol: Ocean Protocol, a decentralised data exchange aiming to bring together data providers, consumers, and service providers into a data marketplace.
Overview of Ocean Protocol🗒️
Ocean Protocol connects providers and consumers of data in a marketplace, while also providing the tools and infrastructure for developers to build value-added services and other marketplaces on top of the data assets created on the protocol. Given this unique blend of services, Ocean Protocol occupies a distinctive place within the Web3 value stack, being both an Infrastructure play and an Exchange.
It is built on Ethereum and any EVM-compatible network. Many Layer 1 blockchains are EVM-compatible, including Moonbeam (EVM environment on Polkadot), Binance Smart Chain, Aurora (on NEAR Protocol), Evmos (on Cosmos), Fantom Opera, and Avalanche. Ocean Protocol, therefore, can be deployed on some of the largest L1 blockchains today, increasing its total addressable market (TAM). Ocean also plans to be multi-chain, meaning that its TAM could be even higher in the future. This may constitute a risk, however, in terms of the liquidity potentially getting fragmented across chains, and interoperability between chains will need to be a lot smoother in the future to accommodate true multi-chain exchanges.
To date, Ocean Protocol has raised a total of $30.83 million over its seed, pre-launch, network launch, and Initial Exchange Offering (IEO). Its investors include Digital Currency Group, Synapse Capital, Merkle Tree Capital, 1kx, BlockAsset, Scytale Ventures, Amino Capital, Berlin Innovation Ventures, 3.0 Capital, Fabric Ventures, Zeroth.AI, Outlier Ventures, Signum Capital and Tokenstack Partners. Many of these are reputable VCs, which grants some level of comfort, but definitely does not mean that the project is guaranteed to succeed.
How Does Ocean Protocol Work? 🤷
Ocean Protocol has two main types of tokens:
A Data NFT, which is a non-fungible ERC-721 token that represents copyright and ownership of the data;
A Datatoken, which is a fungible ERC-20 token that grants data consumers access to that data.
Each data service provided has its own Data NFT and has zero or more datatokens associated with that Data NFT. Having 1 datatoken for a data service grants access to that service. Note: if a Data NFT has no datatokens associated with it, then the dataset is free to access.
Ocean performs two key tasks: it creates the data assets and it facilitates the consumption of the data assets. In the middle, Ocean provides the tools for anyone to build ERC-20 based applications on top of that data, such as Ocean-based data marketplaces. The below diagram may help illustrate this further:
Flexibility Provided by Ocean Protocol
Ocean offers a lot of flexibility for data publishers and consumers.
Firstly, Ocean supports both static and dynamic data from multiple avenues:
Web2 cloud services like AWS; non-permanent Web3 storage services like Filecoin, or Web3 permanent storage services like Arweave for static data;
Web2 streaming APIs, Web3 public data oracles like Chainlink, and Web3 private data oracles like DECO for dynamic data.
Secondly, Ocean allows for the sharing of private data through its ‘Compute-to-Data’ service, which gives data consumers compute access on privately held data. This data never leaves the publisher’s premises, and Ocean allows third parties to run specific compute jobs on that data to obtain inputs into their AI models. Essentially, specific access to private data is granted, rather than that data being directly shared.
Thirdly, Ocean allows publishers freedom in choosing where to host their data. They can hold their dataset in Web2 cloud services like Google Drive, Dropbox and AWS, or locally on their phones or servers. Or instead of having a publicly hosted file, they can have the ‘Compute-to-Data’ service mentioned above for private data.
Lastly, Ocean enables publishers to implement fine-grained access controls for the data services they provide. The level of access is configurable; it can be perpetual (allowing a datatoken holder to indefinitely access the data service), time-bound (e.g., access it only for an hour, a day, within a specific time period, etc.), or one-time (after the service is accessed, the token is burned). Ocean also lets publishers create allow or deny lists, e.g., a consumer can only access the data if they have a datatoken and are on the allow list, or even if a consumer holds a datatoken, they cannot access the data if they are on the deny list.
Pricing of Data: Fixed Pricing & the AMM Model
Ocean uses two methods of pricing data: fixed price and Automated Market Makers (AMMs). Fixed pricing is pretty straightforward to understand: data publishers set a price (in $OCEAN) and consumers buy access to the dataset using $OCEAN and receive datatokens in return.
The AMM model is intriguing. Ocean Market uses Balancer pools, each of which contains a datatoken paired with a base asset like OCEAN, ETH, or DAI. OCEAN is the default base asset, which is intended to drive usage of and demand for the OCEAN token. Balancer was chosen as the AMM model for 2 reasons:
Balancer enables the addition of liquidity through only a single token, which most AMMs don’t allow. When a new datatoken pool is created, only the publisher has datatokens. This single-sided addition of liquidity enables liquidity providers (LPs) to add only OCEAN tokens to the pool. The LPs are incentivised to provide liquidity since they are essentially making a bet on the datasets that are likeliest to be used the most and can consequently accrue the highest fees.
Balancer allows non-equal weights among tokens in a pool. For example, a pool can have 90% weight towards the datatoken and 10% weight towards OCEAN, resulting in the data publisher only needing to provide 1/5th of the OCEAN liquidity for the initial datatoken price. This enables them to bootstrap liquidity to their token faster by reducing the upfront capital requirement. The dynamic weighting also provides a way to mitigate impermanent loss.
The AMM model theoretically allows for organic price discovery. LPs stake OCEAN by adding liquidity to a datatoken-OCEAN pool and earn a cut of transaction fees from the pool in proportion to their stake. The very act of staking OCEAN in a particular pool is a pointer towards the quality of that dataset, since it logically flows that the highest quality datasets would be accessed or traded the most, earning LPs higher fees for staking towards that dataset.
Ocean also recently released V4 of their protocol, which claims to solve the rug-pull problem that previous iterations of the protocol faced. The solution has meant that data publishers don’t get their datatokens all at once; rather, the pool auto-controls datatoken supply. Datatokens are therefore only minted when someone stakes OCEAN into a pool. The amount of datatokens minted is just enough to keep the datatoken price stable.
Ocean Market Purgatory Process
Ocean also has created a solution to account for IP violations and sensitive data. In the Ocean Market Purgatory Process, a data asset enters purgatory if a formal request for takedown is submitted by the alleged actual data owner or aggrieved party due to sensitive data, IP violation, or other reasons. The asset stays in purgatory indefinitely unless it meets specific conditions that lead it to reverting back to default.
Potential Fragility of Ocean’s AMM Model
While Ocean’s AMM model innovatively attempts to solve the pricing problem for data, aiming to help publishers evade the difficult question of how much their data is actually worth, in practical terms it seems to be infeasible, especially at scale. Imagine a future in which there are millions of datasets on Ocean Market. Deep liquidity would be needed for all of them to accurately price the datatokens and not suffer from high slippage. Even though the price discovery for fixed pricing is not as authentic as that in an AMM model, it may be more intuitive for a data consumer to connect to Ocean Market, view the price of a dataset, and choose whether to buy it without needing to worry about slippage or whether the trade will execute due to low liquidity. A bid/auction model may also be a potential middle ground in terms of ease of understanding/usage and price discovery. A potential solution could be to merge the auction and fixed price models. This is how it could work in practice:
Have an auction for a dataset the first time it is introduced;
The price that it has been bought at remains the fixed price of the dataset for a specific time period, say 3 months;
After 3 months, another auction is held to reflect changing demand for the dataset and update the price. This cycle can continue indefinitely.
If the AMM model for Ocean is not successful, it will bring up a critical question: how will OCEAN generate fees and, in turn, how will the OCEAN token accrue value?
The Trends Supporting Ocean Protocol’s Emergence 📈
The main trend behind the need for Ocean Protocol is the exponential rise of AI globally. AI models need reams of data to constantly improve and learn, and good-quality, clean datasets are not easy to come across. The big data market was estimated to be worth $138.9 billion in 2020 and is expected to top $229.4 billion by 2025. This highlights the space in the market for players like Ocean Protocol that provide liquid access to a wide range of datasets while simultaneously allowing data owners to monetise their data.
Businesses also require quality, varied data to be able to trust AI. We’ve seen how bias can infiltrate AI algorithms due to the nature of human biases being reflected in training data used for AI. Using as much training data from as varied datasets as possible could help in smoothing this problem out, consequently increasing trust in AI algorithms, while synthetic data created on the basis of datasets from Ocean Protocol could provide high velocity of data to expeditiously train AI algorithms.
Launch, Traction, Use Cases, and Future Roadmap 🚀
Ocean Protocol’s network launch took place in Q1 2019, with its seed and pre-launch stages attracting 3,500 contributors across 100 countries. Some key statistics around Ocean Protocol’s traction to date are:
Holders: 34,488, an increase of 161.5% since September 2020. However, the significant increase in holders may be masked by plateauing growth since September 2021. The growth in token holders from September 2020 to September 2021 was 130%, while the growth between September 2021 and now has only been 14%. This reflects the general market conditions.
Market Cap: While this doesn’t really reflect traction, Ocean’s market cap is ~$111 million, a drop of ~85% from its peak of ~$786 million in November 2021.
Liquidity: OCEAN’s liquidity is high, with the token being listed on a number of exchanges (both centralised exchanges and DEXs), such as Binance, Kraken, Bittrex, Uniswap, Balancer, gate.io, Poloniex, SushiSwap, Cream Finance, Crypto.com, Huobi, and Coinlist.
A positive sign is how Ocean is working with reputable companies across the board. Firstly, Ocean is working with companies like PwC, KPMG, Holland & Knight, GSR, Messari, and Certik to comply with national laws, regulations, and guidelines, being proactive in ensuring full compliance and increasing user trust. Secondly, Ocean has implemented projects with entities like ConnectedLife, NextBillion, and the Government of Singapore. Stakeholders in these projects include some of the largest companies in the world, such as Unilever, Aviva, Roche Diagnostics, Johnson & Johnson, and Microsoft. The projects demonstrate some of the use cases of Ocean’s product:
ConnectedLife is developing a data marketplace on Ocean Protocol that enables anonymised, non-personal data to be shared freely for open research. Building on top of Ocean Protocol allows patients full control over their data, enables them to securely transfer it, and gives them full auditability over how their data is being used and by whom.
NextBillion has access to large amounts of data about the cleanliness of public toilets in Thailand that is used to train AI models to dispatch maintenance and cleaning personnel faster to clean toilets. NextBillion is working with Ocean Protocol to create a data sharing marketplace where this data can be shared for free for public research or sold for market research.
The Government of Singapore’s Info-communications Media Development Authority (IMDA) is aiming to create a Trusted Data Framework alongside Ocean Protocol and PwC to build a standard to govern data sharing and enable safe and trusted access to data. The aim is to build a generalised data exchange protocol to be used across jurisdictions.
Ocean’s future roadmap is very ambitious (and broad):
Data as an Asset Class for DeFi, where data assets are securitised, used as collateral in stablecoins and loans, and bought and sold in DEXs and CEXs.
Composable Data Assets, where data assets can be composed into a variety of bundles, for example data indexes, where the top 100 data assets are tracked and investible in a single token.
The team will have to be careful to focus on perfecting the core product before focusing too heavily on its long-term roadmap. Both of the above are also possible to be created as services on top of the data assets by external parties. If the team is dedicating too many resources to building out ancillary features before the core product is fully developed, it’s a danger sign.
Ocean Protocol is co-founded by Bruce Pon, a veteran in the blockchain space since 2013. Bruce and his team have created several blockchain applications, all revolving around data. These include BigchainDB, a blockchain database for data-driven use cases which pioneered the Proof of Authority consensus algorithm; ascribe.io, a service that enables creators to register and track intellectual property on the blockchain; and the Interplanetary Database, a global blockchain database network. Prior to working on blockchain applications, Bruce was a founder at Avantalion International Consulting, a company that aimed to build banks for the unbanked. He also worked at Daimler Financial Services for 5 years and Accenture for 6 years.
The second co-founder of Ocean Protocol is Trent McConaghy, who is extremely experienced within the AI space. He started his career by doing AI research for national defence in the 1990s, explored the relationship between human creativity and AI-based creativity in his first startup, ADA, and created the Solido software used for most modern chip designs. He has also advised the German, Dutch, Estonian, and other governments about their blockchain and AI strategies.
Advisors to the Ocean Protocol include crypto personalities and builders like Meltem Demirors, the Chief Strategy Officer at CoinShares, and Ryan Selkis, the founder of Messari.
On the Ocean Protocol website’s team page, a few of the team members link to broken LinkedIn pages, while some key team members all seem to be related. While on the face of it this may seem to be a bit of a red flag, it isn’t clear what this means in practice.
All in all, it seems as if Ocean has a strong, experienced team and set of advisors that have significant relevant experience in the blockchain and AI fields.
Thoughts from Using the Protocol 🖱️
It isn’t especially difficult to use Ocean Market, but this can be caveated by the fact that, for the purposes of this article, we went into quite a lot of detail around how Ocean Protocol exactly works. Each data publisher and consumer may not want to go through a multitude of documents to understand how to use the service.
A key issue with the UX is the lack of standardisation. The datatokens have confusing names for no immediately apparent reason and ought to be standardised to reduce misunderstanding. Moreover, stating the price of the datatoken in ‘mOCEAN' terms is also unnecessary, especially because it isn’t clear what mOCEAN refers to.
The rationale for using OCEAN as the base price asset is understandable since it accrues value to the OCEAN token, but it would be easier to price the dataset in stablecoins to increase user adoption, reduce friction, and increase the ease of convertibility. A good feature is that when you click on the dataset itself, the price converts into the currency of your choosing, but this happens too late in the process. The price of the dataset should also be listed on the primary marketplace page, rather than having to find it by clicking on the dataset.
Therefore, it is obvious that a lot of work needs to go into improving the UI/UX of the protocol. The UI/UX faces the same problems that most crypto protocols face, in that they are built by Web3-native developers for Web3-native users. In order to gain true adoption, a lot of the Web3-native nuances should be ironed out and the focus should be on complete ease of use for users.
The OCEAN token is primarily used on the Ocean Protocol network as a means of value exchange and a mechanism to incentivise ‘keepers’ on the network who validate and verify services and store the history of transactions on the blockchain. It isn’t especially clear what keepers validate and why they’re needed to store transaction history, since Ocean is built on Ethereum and EVM-compatible chains which have their own validators / miners and store transaction history on the blockchain.
OCEAN’s value accrual ‘loop’ is:
The tools and services built on top of Ocean’s data assets are intended to incentivise increased usage of the OCEAN token and a higher volume of transactions from which fees can be collected. This network revenue is directed in two ways: one, burning OCEAN tokens, which reduces OCEAN token supply and in turn increases the value of the OCEAN token. 5% of all network revenue is burned. Two, to OceanDAO, which directs funding received from network revenue and rewards towards ‘workers’ or application builders who create value-added services using the unique properties of Ocean’s data assets.
Ocean is also introducing a veToken (vote-escrowed token) model in Q3 2022. In this model, users lock up their OCEAN tokens for a pre-determined period where it is impossible to unlock their tokens. The tokens vest linearly over this period, and in return for illiquidity, proportional to their lockup period, users receive ‘ve’ tokens, which are special governance tokens. The longer the lockup period, the higher the governance power the veTokens have. This applies to rewards as well – users get higher boosts to their rewards the longer they lock up their tokens. veOCEAN holders will vote for OceanDAO grants in this new model. veOCEAN holders gain value from community fees, where 50% of transaction fees will go to veOCEAN holders and the rest will go to OceanDAO grants, etc.
A veToken model can be a risky proposition in the volatile crypto market, where no one even knows what will happen in a few months. Crypto winter has been known to swallow the market caps of scores of non-blue chip (and many times, even blue chip) tokens. If users do not have the ability to exit and the future of the protocol is in question, then their financial stake is at high risk. So, if stakers want to exit and cannot due to the lockup period, there is a good chance they become unmotivated and therefore do not participate effectively in governance, having the opposite effect to what was intended.
Ocean Protocol raised $27 million over its initial seed and pre-launch rounds. For the Network Launch, Ocean released and exchanged 32 million tokens with SAFTE holders (i.e., initial seed and pre-launch investors) at a price of €0.22 per token, and released 191.7 million tokens in total, representing 13.6% of the supply cap of 1.41 billion tokens.
A significant proportion of tokens will be released as network rewards, representing 51% of the overall OCEAN token supply. The network reward is intended to be programmatically released over 20 years. Given the strong focus on token incentives to reward data publishers and network validators (keepers), the OCEAN token is inflationary. However, given that 5% of all network rewards and any remaining funding from OceanDAO are burned, it can be seen that Ocean is attempting to put some controls on inflation.
Token inflation is intended to follow the below schedule:
The first year of the network (2019-20) saw inflation of 104% from token acquirors, the Ocean Foundation, and the founding team receiving their token allocations. 2020-22 saw inflation ranging between 17-21%, fuelled by the founding team getting their remaining allocations and the network reward beginning to be disbursed. Inflation is intended to be reduced to 6.8% in 2024 and 5.9% in 2025 as the protocol transitions to a decentralised and permissionless chain, and the network reward becomes the only form of inflation.
The initial high token inflation in the first few years aims to bootstrap usage of the protocol, and inflation decreases year on year. If incentives decrease and there isn’t a high amount of usage (i.e., network effects) on the platform by 2025, when inflation decreases to between 5-6%, then Ocean Protocol will have to find another way to get past the Cold Start Problem and incentivise users to keep using the platform.
Token Distribution & Vesting Schedule
The initial supply of OCEAN tokens was set at 690.9 million with a supply cap of 1.41 billion. The tokens are intended to be distributed in the following manner:
70.5 million (5%) for the Ocean Protocol Foundation to build the project's community and ecosystem (i.e., OceanDAO);
282 million (20%) for Newton Circus (DEX) & BigchainDB (i.e., team and founders);
338.4 million (24%) for Acquirers supporting development and token liquidity;
719.1 million (51%) for Ocean rewards.
The vesting periods are as follows:
Founders have a vesting period of 5 years from the project start;
Seed contributors were locked in for 2.5 years (1.5 years prior to network launch and 1 year linear vesting after network launch in Q1 2019);
Pre-launch contributors were locked in for 1.5 years (1 year prior to network launch and 6 month linear vesting);
OceanDAO’s initial treasury was funded from 51% of the unminted OCEAN, some of which it converted into ETH, DAI, and other tokens by selling OCEAN bonds. OceanDAO’s treasury is meant to accrue value from network rewards and active treasury management, while funds from the treasury are spent on network rewards and grants.
Comparing OCEAN’s distribution to the optimal token distribution (where the team and investors obtain 17.5% each, the treasury gets 50%, airdrops account for 5%, and ecosystem incentives are 10%), it’s immediately apparent that ecosystem incentives (i.e., network rewards) are disproportionately high. This shows that the project relies significantly on token inflation, especially in the early stages. However, Ocean released only 13.6% of tokens at network launch, leaving a lot of scope for gradual token release.
Ocean’s treasury allocation is not very clear. While documentation says that OceanDAO’s treasury is allocated from the 51% of unminted OCEAN, it would be better if this was made clearer, including how much ETH, DAI, or other tokens they have in their treasury from selling OCEAN bonds.
In terms of token unlocks onto the market, all tokens allocated to seed and pre-launch contributors (i.e., acquirers) have already been released since Network Launch was in Q1 2019 and their vesting period ended in Q3/4 2021. The main impending token release is from the tokens allocated to the founding team, which should unlock around Q1 2024.
Ocean Protocol is supported by OceanDAO, whose governance procedures are about ‘directing the flow of resources towards the teams and projects that offer the best chance for growth.’ Essentially, OceanDAO implements a one-token-one-vote model where OCEAN holders vote (on Snapshot) for the most promising projects to fund from OceanDAO’s grants and treasury. If there aren’t enough value-added projects available to fund, the remaining OCEAN from the funds pot gets burned. One wallet address can only vote on one proposal, meaning that if an OCEAN holder wants to vote on multiple proposals, they need to fund multiple wallet addresses. This seems like an obvious UI/UX design flaw and may not incentivise accurate voting on proposals.
Moreover, the one-token-one-vote model has several flaws; since the OCEAN tokens are transferable, votes can be bought by whales, who can potentially take control of the voting for the grants. It can also limit scalability since high community participation is required to pass proposals, significant time and energy is utilised in campaigning, and it also becomes extremely time-consuming for average community members to vote on every proposal to meet quorum. The veToken model that is being implemented may also bias voting towards whales who can afford to lock up liquidity for years on end.
The governance structure and procedures of Ocean Protocol and OceanDAO seem to be not very well developed, are rudimentary and definitely are areas for improvement.
OCEAN’s price has dropped by 83% since its high in November 2021, reflecting the broader market downturn. Weekly active users have decreased in terms of old and new users using Ocean Protocol as well.
The drop in weekly active users can be seen as a concern since the usage of datasets to train AI models shouldn’t depend on market conditions. Therefore, using the OCEAN token for payment and all over the Ocean ecosystem may actually end up being detrimental compared to using stablecoins as a means of payment and transaction. This is because the OCEAN token brings in a lot of speculative activity that is unrelated to the actual utility of the protocol. Also, the reduction in weekly active users may also indicate that network effects or user stickiness have not been achieved to any sort of level yet.
Gaps in our Analysis 🕳️
While we have tried to gather as much data as possible to ensure complete, holistic analysis, there are a few gaps. We have not been able to source the protocol revenue, token incentives distributed, and other metrics that can usually be obtained from Token Terminal. These would have been immensely helpful in comprehensively understanding the success of Ocean’s token incentive program and its stickiness as a protocol, but we have used the Weekly Active Users metric as a proxy for this.
Secondly, we have not been able to find the total amount staked and any changes in this number, which would be useful to ascertain growth in use of the platform.
Lastly, the composition of Ocean’s treasury is not readily available, which would have been useful to analyse to understand the effectiveness of their treasury management and in terms of transparency.
Suggestions & Closing Thoughts ⌛
Ocean Protocol is an extremely innovative project trying to ingeniously solve one of the most pressing problems with computing and AI today: the lack of high-quality datasets and the lack of ability that data owners have to monetise their data. If Ocean manages to scale up and become the ubiquitous infrastructure layer for data assets, it could mean a step change in the quality and effectiveness of AI algorithms across all industries.
However, in order to achieve its end-goal of making data a liquid, monetizable asset, Ocean needs to improve on several areas.
Firstly, the UI/UX of the marketplace needs to be much simpler to use, perhaps by using stablecoins or a more widely used unit of account across the marketplace, which will help in engendering more usage. Prioritising the use of stables like USDC and pricing datasets in that stablecoin may make it easier for non Web3-native users to use the platform. A lot needs to be done to attract these users, since they are the ones who will both provide and consume datasets. The marketing and user acquisition and retention strategies have not been best in class since they have been too geared, like most Web3 projects, towards Web3-native users. Users need to feel like they are just using a marketplace and not get too deep in the details about the Web3-related aspect of it.
Secondly, the AMM model may also be too complicated for most datasets and scalability is not guaranteed. Setting fixed prices and enabling the ability for providers to have auctions may be the better option. Once there are enough users on the platform to provide liquidity for datasets, introducing the AMM model in the future may be a good choice.
Lastly, the tokenomics of OCEAN need to be simplified. Value being accrued in OCEAN is not absolutely necessary to attracting users and keeping them on the platform. The focus needs to be on attracting and keeping users sustainably, so while token inflation and rewards may make it easier to attract users, the focus needs to be on keeping them via improved UI/UX and ease of usage.
To join this journey into the worlds of tech, business, and Web3, subscribe below!
This is a personal blog. Any views or opinions represented in this blog are personal and belong solely to the article authors and do not represent those of people, institutions or organizations that those authors may or may not be associated with in professional or personal capacity, unless explicitly stated. All content provided on this blog is for informational purposes only. The owner of this blog makes no representations as to the accuracy or completeness of any information on this site or found by following any link on this site. The owner will not be liable for any errors or omissions in this information nor for the availability of this information. The owner will not be liable for any losses, injuries, or damages from the display or use of this information. Any views or opinions are not intended to malign any religion, ethnic group, club, organization, company, or individual.
👇🏽 please hit the ♥️ button below if you enjoyed this post.