How do DeFi apps engineer sustained growth in an industry characterized by little or no user loyalty?
The relatively low cost of switching between decentralized applications, or “dApps,” coupled with the constant availability of “switching incentives” in the form of yield farming rewards, has made it infinitely harder for any dApp to engineer user growth strategies. Additionally, the open source nature of the blockchain has eroded product differentiation and made it challenging for dApps to sustain a favorable value proposition over competition.
To answer the question, I’ve examined the impressive growth of Curve, a popular constant function market maker (CFMM), to understand what has prompted its fast growth and suggest a framework that other dApps can follow to engineer sustainable growth.
The Impressive Growth of Curve
There is no doubt that automated market makers (AMMs) have recently experienced unprecedented growth, in part because of the substantial value they deliver to a growing user base.
Curve has managed to garner a substantial market share in terms of daily trading volumes, becoming the main destination for many daily DeFi users. Between June 2020 and February 2021, The monthly total value locked (TVL), trading volume and fees as you can see in this dashboard, or take a look at the charts below.
Deconstructing Curve’s Growth Strategy
Curve’s success can be attributed to three main reasons:
Well-crafted value proposition.
Sound growth strategy.
Let’s examine each area briefly.
Value Proposition That Improves on Existing AMMs
Curve’s market design creates significant value by reducing inefficiencies that happen when traders and liquidity providers don’t reach the best possible outcome as a result of a trade. Those inefficiencies are commonly known as Slippage and Impermanent Loss.
To reduce inefficiencies, Curve cleverly:
Offers pools of similarly priced tokens with stable ratio to each other, resulting in less market volatility when a trade takes place.
Rebalances a pool’s ratio of tokens to restore an ideal state. The stability achieves high liquidity utilization compared to other AMMs.
Constant function market makers (CFMMs) thrive when they are highly capitalized. Large liquidity enables high liquidity utilization in the form of large trading volumes.
This makes liquidity providers the primary user group Curve seeks, followed by traders. It might appear on the surface that Curve targets all types of traders and liquidity providers but it is hardly that simple.
While a complete analysis of Curve’s users and their derived benefits is beyond the scope of this post, we can broadly assume that Curve’s overall value derived from reducing inefficiencies is divided equally among three overlapping user groups:
Liquidity providers: Users with working capital, seeking a competitive yield without losing the value of their investment (Value: APY + minimized impermanent loss).
Curve DAO members: Among liquidity providers exists a high-value user group that has a vested interest in the future of the DAO. These users likely participate in governance, create proposals, advocate for the platform, etc. (Value: Boosted APY + appreciating token value + minimized impermanent loss)
Traders: Traders seeking the maximum value for a trade (Value: Low trading fee + low slippage). Traders include those who indirectly use the CUrve via DEX aggregators.
Curve’s economic design primarily aligns the interest of all user groups using trading fees. Fees are divided between liquidity providers and DAO members (veCRV holders). The current trading fee is priced competitively in an attempt to generate large trading volumes.
Focused Growth Strategy
With a clearly defined value proposition and user groups, Curve sets to formulate its growth strategy to address two fundamental questions: What is growth, and what drives it?
There are many ways to define growth. Some apps look at it in terms of overall revenue generated, while others use usage metrics, such as the number of users.
Personally, I believe growth should be defined and measured in a way that reflects whether the application delivers maximized value to its user groups.
In Curve’s context, its market design delivers maximized value when both of these are achieved:
A liquidity provider earns competitive APY at zero or significantly low impermanent loss.
A trader obtains the maximum value (lowest slippage) for a trade for a low trading fee.
Therefore, for a value-maximized trade to take place, Curve must keep large amounts of liquidity locked in any given pool.
Based on the above economic logic we might conclude that:
a- Liquidity drives trading events.
b- Curve’s growth can be tracked and verified by looking at two main metrics:
Total liquidity locked in its pools.
Total trading volume.
We will call the above North Star metrics.
When attempting to achieve a consistent lift in growth metrics, growth engineers face two challenges:
There is more than one driver of growth.
Growth drivers will impact growth in varying degrees.
One way to overcome those challenges is to group call growth drivers in a growth formula and then test each driver to determine its correlation with the success metric and nature of that correlation (direct, inverse).
In Curve’s example, there are a number of growth drivers that can be pulled to produce a lift in its presumed North Star metrics (TVL/trading volumes). While generally one growth formula is enough, we will need two, given that there are two user groups in Curve’s ecosystem, traders and liquidity providers.
Formula 1: Increasing total value locked
Total value locked = Base APY x Lending APY x Yield Farming APY x Number of Liquidity Providers x Average Value Locked x Lock Duration x Number of Pools x Average Pool Quality X Impermanent Loss x Network Fee
Formula 2: Increasing trading volume
Trading volume = Number of Traders x Number of Trades x Avg. Trade Value x Number of Pools x Slippage x Trading Fee x Network Fee
Pulling Growth Levers
With North Star metrics and growth levers defined, let’s look at where Curve has actually been focusing its efforts.
Spread of Pools
One of the primary drivers of Curve’s growth in general is its selection of pools. Take the 3pool, for example. It delivers optimal value to stablecoin traders at a time that has seen an explosion in the use of stablecoins. But it’s strategy for market-making extends beyond stablecoins. Pools like wBTC-rentBTC and hBTC-wBTC, among others, provide liquidity markets for popular tokenized versions of Bitcoin on the Ethereum blockchain.
But Curve’s cleverest move has arguably been piggybacking the growing lending markets by creating “lending” pools such as Aave and Compound. Users with interest-bearing tokens such as aDai or cDai can add them to Aave and Compound pools, effectively making a combined APY from both lending and yield farming.
For those who don’t hold interest-bearing tokens, they can still earn lending APY by contributing stablecoins to those pools. As a result, the lending pools have become markets that enable multiple use cases such as lending and lending-APY optimization via exchanging interest-bearing tokens.
Finally, an initiative like the Pool Factory, where any project can deploy a metapool (think gusd/3pool, usdn/3pool, etc.) has led to an increase in the number of pools available for both liquidity providers and traders. Metapools are permissionless and are listed on Curve.fi UI when they meet certain criteria. With more high-quality pools available on Curve, Curve has been able to attract both liquidity and trading volumes.
In addition to offering a broad spread of high-quality pools, Curve charges a significantly lower trading fee compared to those of other AMMs, such as Uniswap.
Charging a low trading fee is a pricing choice that likely signals a belief among Curve’s team that a low trading fee drives larger trading volumes and fees.
While tapping into the “pricing” growth lever has likely had a positive impact on trading volumes, strategically, Curve’s trading fee creates a serious misalignment between the objectives of the traders on one side and liquidity providers and the DAO.
On one hand, traders seek to pay the lowest trading fee, and on the other, liquidity providers and veCRV holders seek to make the highest APY generated from trading fees. I will get back to this point later in the post.
Since its launch in August 2020, CRV rewards have served as a major incentive to attract liquidity providers. In mid-January 2021, the CRV rewards APY increased dramatically following the appreciation in CRV tokens’ value, springing up unprecedented growth in total liquidity locked and resulting in the AMM becoming the leader in terms of total locked value.
At the heart of Curve’s yield farming strategy is a unique design mechanism: the boost. Liquidity providers can choose to vote-lock CRV tokens into the DAO for a period of time to receive a boost in CRV rewards and a portion of trading fees. The boost keeps Curve DAO’s interest close to heart and achieves two goals:
It encourages participating in the governance of the DAO. This incentivized behavior creates a class of high-value user group with a vested interest in the welfare of the DAO.
It creates an incentive to hold the token, effectively driving the token’s value up and the CRV rewards APY.
As of today, the total CRV locked accounts for ~47% of total circulating tokens, with an average lock time of 3.63 years.
Curve’s yield farming has led to a significant increase in total value locked and, consequently, trading volumes and fees.
Delivering Maximized Value Is the Heart of Growth Engineering
The growth formulas mentioned earlier list many potential areas that Curve can tap into to drive its success metrics up.
Here are a few ideas that map onto specific growth levers.
Reduce Network Fees (Growth Lever: Transaction Fees)
Scalability is the hottest topic in the blockchain space right now. Rising transaction fees have rendered many DeFi applications less usable, eliminated many users from accessing DeFi services, and profoundly changed the usage behavior of many others.
This makes scalability the biggest opportunity to engineer value to both small and large net-worth users.
For Curve, lower transaction fees will likely have a major positive impact on its North Star metrics. For example, with lower fees, new liquidity providers, especially those with smaller “working capital,” can afford to join in. Also, the number of governance participants increases, existing liquidity providers manage their stake with freedom, and traders will find it much cheaper to trade, increasing trading volumes. The list of benefits is endless.
Increase Trading Fees (Growth Lever: Base APY)
Curve charges a 0.04% trading fee, and half of it (0.02%) goes to liquidity providers and vested users. Fees generated from trading, combined with CRV rewards, make up the main incentive to lock liquidity and participate in governance. (It’s worth mentioning that other pools, such as lending pools, offer additional APY derived from lending.)
When Curve’s CRV-based rewards drop significantly, either because the token value is devalued due to market conditions or when reward emission eventually depletes,trading-based APY will likely plummet, potentially leading to a liquidity flight.
To illustrate the point in numbers, liquidity providers in a Uniswap stablecoin pool like USDC-USDT earn a 0.3% trading fee, which translates to 10.4% APY annualized given a modest daily trading volume of $5.4M. In comparison, Curve’s trading fee APY for 3pool on a day of $61M trading volume translates to around 1% APY. The difference is striking.
Unless Curve expects its trading volumes to grow exponentially by the time its CRV rewards deplete, raising the trading fee is an eventuality, in my opinion, and Curve DAO should not be deterred from considering it. I base my argument on the possibility that a low trading fee is only one of many reasons a trader, or DEX aggregator, chooses Curve over the competition.
Curve’s broad offering of low-slippage pools likely shapes the main incentive to use the platform. But to confirm this hypothesis, Curve DAO can commission a market research and economic analysis to determine the minimum rate required to provide sufficient incentive to align the interest of traders, liquidity providers, and the DAO.
In a market where 0.3% seems to be the standard in exchange for instant liquidity, Curve might be able to raise the trading fee to that of its rivals and still maintain advantage over other options.
Subsidized Insurance (Growth Levers: Number of Liquidity Providers, Average Locked Value)
Assuming having a Curve cover policy is a driver of liquidity, Curve can become an insurance policy provider on one of the insurance protocols where it deposits CRV tokens as collateral to mint claim tokens that can be awarded to liquidity providers as an additional incentive.
One way to do that is by redesigning its tokenization model to allocate a portion of the emitted CRV rewards to create a base cover — effectively lowering the price of existing cover sold on the market — for those who are deterred from adding liquidity or locking bigger amounts because of security concerns. This will serve as an additional incentive for users to add liquidity, improve Curve’s reputation, and attract more liquidity providers to the platform.
The Growth Engineering Model
Many DeFi applications use yield farming as the main tool to seed user traction and engineer growth to the extent that yield farming has become the main value proposition and main value delivery mechanism. Such dApps can never sustain growth and likely face demise as soon as users discover the shallow value such projects deliver.
Yield farming is not equivalent to market design, so dApps should follow a clear framework through which they can consistently engineer value and deliver it to their user groups in order to grow and sustain a market position.
So here is a glimpse of a model that I suggest dApps can follow. I call it the Growth Engineering Model.
Vision: Articulate your product vision and what goals you aspire to achieve.
Value proposition: Define your dApp’s value proposition precisely, detailing its complete use cases, intended markets, users, and the business strategy you will devise to deliver the value. A well-defined value proposition resolves any conflicting goals between user groups and helps you prioritize your dApp’s goals.
User definition: Conduct a complete benefit analysis for each of your target user groups. The analysis should output their needs, problems, and preferences on how they will be fulfilled. This can serve as a basis for adding value in the future.
Success metrics: Define your growth and how you’re going to measure it with North Star metrics. Success metrics are derived from your dApp’s value proposition and target user groups. Ideally, you should focus on one success metric, but if you have two distinct user bases, such as the case with constant function AMMs, you might want to define two success metrics.
Growth formulas: With success metrics defined, hypothesize about different areas of focus (growth levers) that when tapped into produce a lift in your success metrics. Each growth lever is a hypothesis that should be tested and validated. Growth formulas are living organisms and should be examined and updated regularly.
Experimentation: Founding teams that develop a taste for experiments likely win. Approach each business initiative or build with skepticism and try to limit its scale in favor of faster shipping. Then set to validate the hypothesis that it is supposed to validate through limited-scale experiments. When experiments are not possible, given the nature of smart contract immutability, teams can conduct qualitative market research to validate certain assumptions to the best of their knowledge. In fact, market research is an important step that should precede any experiment when the latter is complex to build — the more complex the experiment, the more important it is to conduct limited market research.