Research Summary: A Short Survey on Business Models of Decentralized Finance (DeFi) Protocols


  • This paper explains the revenue stream of DeFi protocols to help investors and users understand the growth, sustainability, and security of different protocols.
  • The authors discuss the business model of three DeFi protocol domains, abstract the general business model and identify key DeFi actors and actions.
  • The authors conclude by asking research questions on subjects such as the valuation of similar governance tokens, the regulatory sustainability of DeFi business models, and more.

Core Research Question

What are the common elements in the business models of DeFi platforms, and how do the protocols make a profit?


Xu, T. and Xu, J., 2022. A Short Survey on Business Models of Decentralized Finance (DeFi) Protocols. [2202.07742] A Short Survey on Business Models of Decentralized Finance (DeFi) Protocols


  • DeFi Protocol: Provides open, non-custodial, permissionless, and composable variations of classical financial services in exchange for fees on asset movements such as borrowing, swapping or investing.
  • Composable: Participants can arbitrarily create novel and complex financial services by combining functions and services from multiple existing DeFi protocols.
  • Non-custodial: Participants are the custodians of their own assets.
  • Permissionless: Participants can interact with services without requiring authorization.
  • Total Value Locked (TVL): the general value of all crypto assets in a DeFi protocol or DeFi protocols. It covers assets for transactions, lending, and borrowing. TVL shows a protocol’s current usability based on the number of active users.
  • Liquidity Pools: Liquidity pools are a mechanism by which users can pool their assets in a DEX’s smart contracts to provide asset liquidity for traders to swap between currencies.
  • Investor: Lenders, liquidity providers, or vault users that undertake the protocol’s underlying risks, such as protocol misbehavior, impermanent loss, or rug-pulls, by depositing his assets for passive income.
  • User: Borrowers, buyers, traders, or other active actors using the protocol on the go through borrowing or swapping and pays fees for the movement of assets.
  • DeFi Financial Service: Lending pools, liquidity pools, vaults, or other protocols that use smart contracts to let investors and users interact indirectly; it locks investor assets, satisfies user requests, and prevents protocol misuse.


  • Since 2021, DeFi TVL has notably surged. Prospective investors are interested in identifying the business models with steady and constant revenue streams before investing in the underlying project.

  • The authors group DeFi protocols into three domains: Protocols for Loanable Funds (PLFs), Decentralized Exchanges (DEXs), and Yield Aggregators.

  • Protocols for Loanable Funds / Lending Protocol

    • Protocols like Maker, Aave, and Compound let users lend and borrow digital assets. Lenders deposit assets in the pool and get interest from fees paid by borrowers. Borrowers also deposit collateral before taking the loans; the loans are over-collateralized to encourage payback.
    • Business Model: The PLF’s cash flows depend on the interest rate model, the underlying demand and supply, and the total amount borrowed. The interest rate model can be linear, non-linear, or kinked. The interest is split proportionally among lenders, and the PLF takes a percentage. Some PLFs offer flash loans with fixed or no interest rates to generate more revenues for the protocol.|572.3876210105553x303
  • DEXs / Liquidity Pools / AMMs

    • These services allow users to trade or swap cryptos through smart contracts called automated market makers (AMMs). The AMM pools funds from investors called Liquidity Providers (LP), who in turn get LP tokens representing their share of ownership of the pool.
    • Business Model: DEXs generally adopt the maker-taker model where fees paid by buyers are split proportionally between liquidity providers, and a percentage goes to the protocol’s treasury. This percentage is the primary income source for DEXs. The Swap fees may be dynamic and set by the users or fixed.|572.3876210105553x303
  • Yield Aggregators /Yield Optimization Protocols

    • These protocols generate returns measured in annual percentage yields (APY) for investors through strategies that may be too expensive or complex for individual investors. Such strategies include finding the best lending interest rates, borrowing assets, and leveraging another position by exploiting different protocols like Compound and Uniswap.
    • Business Model: The strategy’s performance dictates the protocol’s cash flow. The protocol earns commission fees from the strategy’s profit. YearnFinance v2, for example, applies 20% as a performance fee and an additional 2% as a management fee. |572.3876210105553x303


  • The researchers surveyed, analyzed, and compared the business models of three major DeFi domains. They identified the common actors and services that access the domains.
  • This was then abstracted to create a generalized framework for DeFi business models based on shared attributes, services, and actors in the DeFi domains.



  • The authors found that the general models employed in DeFi is similar to the concept of “Two-Sided Markets” in classical finance. In “Two-Sided Markets” the investor provides liquidity to the financial service that peer users can use. The protocol and investors then receive income from fees paid by protocol users.

  • Finally, the authors provide a table grouping actors and services in DeFi and showing the differences in naming conventions.

    DeFi Protocol Smart Contract Investor User Financial Service
    PLFs Lending Pool Lender Borrower Loan
    DEXs Liquidity Pool Liquidity Provider Buyer/Trader Exchange
    Yield Aggregators Vault Vault User - Asset Management

Discussion and Key Takeaways

  • The authors provide a general abstract of DeFi business models; this could in later research be used to develop a formal approach to valuing tokens.
  • The authors also mention some strategies for generating increased revenue utilized by different protocols. It will be great to see formal research on how specific strategies like facilitating flash loans affect the entire DeFi ecosystem.
  • DeFi TVL since the release of the paper has risen and fallen. Investors and builders will highly benefit from research comparing the performance of protocols by the business model adopted.

Implications and Follow-ups

  • Value Investing: Does adopting a value investment strategy, which pays off often in classical finance, generate similar or higher returns in crypto-assets?
  • Voting Power: Why are similar DAO tokens having the same business models and providing the same services priced differently?
  • Regulatory Issues: What DeFi business models are sustainable and economically secure given the different regulatory regimes worldwide?


  • The study creates a generalized framework of business models adopted by DeFi protocols. This generalized framework encourages further research on the sustainability of different DeFi platforms in divergent economic, technical, and regulatory contexts.
  • This research breaks down DeFi’s service business model and helps investors and users understand which protocols have sound fundamentals and reliable cash flows.

Thank you @tomideadeoye for the summary and animations you’re carefully put together.

This is an interesting paper that surveys the business logics of DeFi services we’ve seen so far. Do you think they are fundamentally different from traditional finance (I see PLF, DEX, and Yield Aggregators each loosely corresponding to the iconic images of Banks, Stock Brokers, and Mutual Funds)? What nuances in terms of the contents of the flow chat will be different for traditional finance?

I’m also curious about another thing - for yield aggregators, their job is to invest the locked assets from their clients for a higher profit. Given that they can only invest in DeFi products (otherwise it defeats the purpose of DeFi), they must be turning to the other two types of businesses for opportunities.

This makes it important to scrutinize how the other two business logics operate at first principles. For PLFs, maximizing interest margin is at the core of their profit model. However, in traditional finance, banks lend money to businesses that sell or produce consumable* products and/or services. Do similar businesses like that exist at scale in the DeFi world? If not, how do the users of PLFs justify the interest margin in borrowing?

*consumable as, not for investing


@tomideadeoye Thanks for this writeup. The animations are great and clearly show the flow of funds in the different protocols.

Following up on your follow up:

Voting Power: Why are similar DAO tokens having the same business models and providing the same services priced differently?

This bring to my mind Uniswap: despite a protocol fee of 0, the UNI token has value greater than 0. This begs the question: what exactly are people paying for when they purchase UNI? I’m guessing that the answer to this question will shed light upon the voting power follow up.

This is also closely related to the Value Investing follow up question which is predicated on the calculation of the “business” value of a DeFi protocol. There has been some work on this topic (some material here) but it is still in its infancy.


However, in traditional finance, banks lend money to businesses that sell or produce consumable* products and/or services. Do similar businesses like that exist at scale in the DeFi world?

This is a really interesting question and I wonder how something like that would function in a permissionless setting. In other words, if I’m building a DeFi protocol and I need funds, I might want to take a loan. But in the PLF’s, I need collateral whose value exceeds the value of the loan I’m taking so I’d be better off just using my own money.

If I have a small business and the bank gives me a loan, the collateral is my business itself coupled with the (perceived) strength of my business plan. I wonder if there is some way to create contracts which “incrementally” collateralize a loan, or give the lender some sort of recourse without KYC to recover funds from an undercollateralized position that fails.

1 Like

Regulatory Issues: What DeFi business models are sustainable and economically secure given the different regulatory regimes worldwide

DeFi is here to stay and it certainly has the capacity to finance real-world assets. So far, however, DeFi has little to show in terms of doing any real-world good. Over-collateralized loan pools typically seen in DeFi do not help solve challenges such as financial inclusion, affordable public services and environmental sustainability.

It is not surprising, though, that DeFi has not yet made its mark in much of the real world. The field is still in its infancy, and the ease of use remains a challenge for non-crypto folks. Using DeFi carries significant risk for an individual who is unfamiliar with how it works. And even for those who understand the risks, it can still be a tricky affair given the volatility of many tokens. If you borrow tokens from a loan pool, for instance, regular monitoring of the market is needed to ensure that there is always sufficient collateral against the loan. Failure can result in the loan pool being liquidated and the borrower losing their collateral.

1 Like

If you read Perez book(s), the take-away is that new technology waves are not (initially) financed by TradFi because, frankly they are too crazy. Hence crypto-anarchists and degens migrated to the most unregulated space with biggest arbitrage opportunity … hence we had the ICOnU boom-bust of 2018 (which generously speaking funded a bunch of public R&D from idiots seeking alpha), the NFTcraze of 2021 (which will be cracked down sooner or later once regulators get their A out of G) and probably a more reasoned ABC - STO in next 5 years as (un)stable coins flee to safe harbors.

So far, however, DeFi has little to show in terms of doing any real-world good.

Well, it’s certainly pushing a lot of R&D into AMMs, out of which we have motivated the black hat sector and counter-reactionary red hat response. If you think of it in evolutionary terms, we’re seeing the rise of parasitic flashbot viruses and emergence of protocol antibodies which are needed to build up more resilient Internet of Value. The hype cycle is slow but empirical evidence is that the progress is real (if erratic). The J-curve below is a myth but the aggregate effect does allow for non-linear take-off effects. However the value accretion is not obvious. For example, there were hundreds of car manufacturers in Detroit but a smart investor would have bet on suburban real-estate as white-flight from inner city slums moved value to city fringes.

1 Like

Hey @Twan , the basic service pursued by the user is DeFi is still recognizable and basic TradFi business models could also look like the abstracted DeFi business model. I do not think they are fundamentally different other than the general differences operating a service on a Blockchain creates. However, since DeFi is built around crypto assets and also because innovation in DeFi has not been highly constricted by regulations, there are some difference introduced by concepts like automatic liquidations, keepers, composability and more which can’t exist in traditional banking systems.

How do PLF users justify the margin in borrowing?

That’s a very interesting question. I think a business owner producing consumables could as well deposit crypto assets as collateral and draw up stablecoins for real world investment - if he doesn’t want to liquidate his Eth. The borrower might not need to use the funds for another on-chain transaction. I guess it’d be interesting to know what the average PLF user does with borrowed funds.

Thanks @notthatintodefi. Could credit protocols be what you are looking for?

1 Like

I enjoyed the article and comments. There are a number of DeFi business models that aren’t easily comparable to the traditional finance institutional models, and I personally find those the most interesting.

Consider the Terra ecosystem, which has crashed but was the largest decentralized algorithmic stablecoin ecosystem to date. In theory, one USTerra is pegged to one USD. When demand for USTerra grows, its’ met as an equivalent number of Luna is burned. So Luna operates as a governance token, but also to back the peg. The peg was maintained at higher resolutions by arbitrage. The stablecoin ecosystem crashed but the model is replicated by dozens of other protocols and its’ too early to proclaim the death of the (undercollateralized) decentralized algorithmic stablecoin, IMHO. This is a novel application of DeFi tools that can’t be replicated by non-nation-state actors in a tradfi sense, the product of Terra was a better stablecoin. See also MakerDAO and Celo, both of which issue a collateralized stablecoin which is meant to be “backed” by natural capital assets.

Interestingly, Terra had a high-functioning auto-lending protocol, Anchor Protocol. Anchor allowed the user to bond Luna, collateralize their bLuna and borrow UST up to a portion of the total value of collateral. For extensive time periods, Anchor paid out negative interest as it was rewarding borrowers with allocation of governance tokens - this artificial subsidy is one means that PLF users justify the margin, and @tomideadeoye already mentioned case where post Eth as collateral, for stablecoins to inject into IRL business. Back in Terra Classic, Kujira built a highly efficient liquidation market called Orca. My experience was highly positive, the protocols functioned quickly, cheaply and as-advertised with uncertain long term sustainability. That is, until the stablecoin de-pegged…

Then we have blockchain networks that work, and provide energy, communication and/or computing services in a decentralized fashion, as infrastructure. Perhaps this isn’t pure DeFi, and I’d characterize these types of blockchains as decentralized infrastructure, or perhaps defi for infrastructure. Consider Helium which offers an incentive mechanism to grow its’ IOT mesh network, and 5G mobile telecom network. Imperfect but uniquely web3. Consider Akash, which offers cloud computing, or Render, graphical processors for rendering. These models, generally built on proof of stake governance, provide value beyond the maintenance of the network (proof of work like BitCoin). Arguably, Regen could be considered within this category, as enabling the formation of markets that aren’t achievable without utilizing blockchain, smart contracts and crypto.

@WaterLily poses an interesting question, that is which of the models described in the article, or elsewhere in the forum, are sustainable and economically secure given different regulatory regimes. There is certainly a great deal of risk related to regulatory compliance, and a great amount of uncertainty. I ascribe relatively high levels of risk to the experiential sovereign stablecoins, like Terra, as being a likely target of legislation and agencies. Lesser risk attaches to those blockchain networks that work, Helium or Akash – at least by securities regulators like the SEC, although they may be targeted by monopolistic industry.

Of the DeFi models described in the original post, the lender pool is likely to be regulated, in the US at least. And the auto-aggregators might or might not resemble a security, depending on the utility of the gov token and the courts, among other things.

As a final note, its’ worth noting that the value of a gov token is unlikely to correspond to the success of the protocol’s business model, unless paired with appropriate token economics. This was mentioned by @notthatintodefi. The ability to have a vote and receive staking and LP income may have some value, but the gov tokens are often disconnected partially or fully from the overall success of the business model. This was an attraction of Terra Classic, in that increased demand for the stablecoin product put significant depreciation pressure on Luna’s supply, and drove incredible value appreciation – again, until depegged and broke. Luna built value as demand for the stablecoin grew, and lost value when diminished, and had value beyond governance, to pay fees and provide liquidity (assuming success of its’ stablecoin).