Written by Patrick Fitzgerald. Twitter: @patfitzgerald01
What is the internet native go-to-market / growth strategy?
The most misunderstood thing about crypto by the general public is that crypto is simply a bunch of gambling tokens, not a new digital asset class that unlocks a broader ownership economy, a new set of stakeholders outside the company that can share in and are incentivized to contribute to your success. But in fairness, who could be blamed for having this opinion if you looked at the manic musical chairs of crypto investment from afar? It feels like the mean lifespan of a web3 startup is closer to weeks than years. The average shelf life of open source financial software is surely the result of many protocols having little ambition for longevity, conceived to enrich the founders rather than serve users with existing pain points. However, it is also true that many early failures are a result of protocols lacking sustainable internet native go to market and growth strategies.
Token incentives are an essential piece of Web3 growth strategy. Tokens can attract users and employees to work on the product. Token rewards can be used to incentivize native usage or composability with other protocols. Despite these other use cases, the majority of token incentives are spent simply heavily incentivizing someone to hold your token and stake it on an exchange to create a liquid market.
The fact that this core question of how internet native companies should go to market has not been resolved suggests that there is much to be figured out and discovered. Furthermore, it invites the question of how much suppressed future innovation might be unleashed the second the system begins to function properly when the go to market model is fixed? Looking at the flipside, however, as we are able to point out the system’s current flaws, how much of the current decentralized financial system is imminent perishable, built on a shoddy foundation? We should be reminded that there is so much value in fine tuning the core elements of a system because ultimately that is the most powerful lever to move the world and change behavior. The system is what the system does.
Admittedly, I find myself becoming cynical and disillusioned with the current state of decentralized finance and most acutely when I compare it to the original mission that attracted me to the industry in the first place. The legitimate technological achievements of crypto are increasingly lost and discounted in the noise of greed and speculation. In a sense, it is a near deterministic outcome when you connect the global human mind to a web of open information and now to a web of value. The stories encoded in that information become a powerful wrapper on the movement of value that can amplify or dampen its success. Financial propaganda masquerading as harmless memes, deep tribal affinity and the resulting angry toxicity, and a mix of powerful storytelling, outright fraud and slight of hand tokenomics to lower the cost of the capital are the natural consequences of the system as we have currently built it. Slightly tweaking the process at a structural level might be the lever we can pull to reduce volatility of tokens, complexity for users as well as fraud and general malfeasance that grows in the shadow cast by all this complexity.
There are really only a few essential building blocks of DeFi, and of finance more broadly. You need exchanges to buy and sell assets, money markets to borrow and lend, and a variety of mechanisms to transfer risk. In this essay we will mostly focus on decentralized exchanges and risk transfer mechanisms.
While many think that Uniswap was the first decentralized exchange, it was just the first to go mainstream. However, Bancor actually created the first automated market maker or AMM. So what is a market maker and then what is an automated market maker? A market maker is an entity that is willing to take on the risk exposure of holding an inventory of stocks or tokens, in a way not dissimilar to how a retailer takes on the inventory risk of physical items, in the hopes of selling for a profit. The market marker in its essence plays the role of providing enough liquidity in trading pools so that natural trading activity does not lead to rapid price swings resulting from low pool depth. When there is not enough liquidity in a market, slippage occurs, which is the difference between the current price of token(s) A and the value that you get in exchange in token(s) B during a trade. A traditional market maker is willing to supply these assets, in this case tokens, in exchange for a spread between what someone will sell for and what someone will buy it for.
In 2017, Bancor introduced the concept of automated market making, but it’s more understandable for users to think of it as democratized market making. Unlike the traditional system where a monolithic centralized institution like Citadel earns all the profits from trading activity, any token holder can be a market maker themselves. After purchasing a token or earning it natively within an application, users could now earn additional interest on that token by supplying it to trading pools to earn fees. There was one issue. If a user wanted to stake their token(s), they had to own equal amounts of Bancor’s native token, $BNT, and the token they originally sought to earn interest on. Users saw little value at the time in holding $BNT as it introduced a capital concentration/productivity tradeoff. Imagine someone only had $1000 to invest in cryptocurrencies. In order to supply any of them to Bancor, they would need to own $500 of $BNT and have only $500 to allocate to the other token(s). What if they wanted to construct a diversified portfolio with the full $1000 or concentrate the full $1000 in a single asset they had deep conviction in? It was impossible to contribute the full amount of those tokens to a market maker for additional interest at the time. More specifically, there was a capital concentration / capital productivity (efficiency) frontier that was created by this tradeoff. A token holder could either have the portfolio construction of their choosing with no forced exposure to the common pair token of the market maker (in this case $BNT), or they could achieve maximum productivity by getting interest on the full $1000, but have to take on half exposure to $500 of $BNT.
Later on, the efficient frontier shifted as new lending protocols like Compound or Aave enabled users to supply all of their desired token(s) for yield. The attractiveness of that opportunity would be short lived as yield farming was soon popularized as a go to market strategy. This once again shifted the efficient frontier to favor AMMs and the combined native and incentivized pool yield it gave users exposure to.
This line of inquiry exposes another core tradeoff. The market maker who chooses productivity over desired concentration hopes the prices of the two tokens do not diverge too much while hoping there is high trading velocity in the pool to produce more fees. Without going too deep into the details of arbitrage, when two assets diverge in price greatly, the pool leaks value as arbitrage bots are able to exploit temporary price discrepancies between exchanges and buy a token on one exchange and sell it on another exchange for more until these autonomous pools once again reach an equilibrium. These arbitrageurs are not necessarily leeches or bad actors for exploiting this. Just like users who provide tokens (liquidity providers or LPs) provide a valuable service to a democratized and permissionless AMM, arbitrageurs are a natural ecosystem participant that add value by bringing market rates to equilibrium without manual manipulation (centralized) or oracle information feeds (exploitable).
Unfortunately, this is where the tradeoff arises. Volatile prices are often the catalyst for high trading volume. The liquidity provider who has chosen capital efficiency/productivity (interest) over capital concentration (no forced token exposure) faces the unfortunate dilemma that often at the times where the most revenue is generated, they often profit the least as the pool leaks the most value. The industry as a whole is still working on terminology to communicate this crypto native risk still, but this is commonly known as “impermanent loss” or “divergence loss”. At a high level, at times of high volatility, as the two tokens in a pool rapidly diverge in price and open up arbitrage opportunities, the resulting impermanent loss leads to an outcome where the loss becomes permanent and the token staker would have been better off just holding the token and not providing liquidity. Furthermore, since this investor chose productivity (interest) at the expense of capital concentration, their loss is amplified to the upside because they had to take on 50% exposure to the common pair token on whatever exchange they are using. They would have been better off holding each of the 50% of token A and B without supplying to the AMM and they would have been even better off to just have 100% in the original token of their choosing which outperformed the exchange token.
Given that volatility often preempts fees, but impermanent loss often builds up faster than the volatility-catalyzed fees can compensate for during these times, why do users still supply tokens to these AMMs? Users across decentralized finance appear to be consistently losing money, but the problem is particularly insidious because many users have no way of knowing and their exchanges certainly don’t want them to be any the wiser. Imagine I supply $1000 of token A and $1000 of the common pair token to an AMM, Token B. After 1 year, my $LP tokens are worth $3000 (including price appreciation and swap fees). Great, I just made 50%! But what if in reality token A went up 75% and Token B went up 60%. Held off the exchange, they would have made more - $350 dollars more to be exact in this hypothetical situation as the lost principal to arbitrage was not counteracted with enough fees during that time period. Furthermore, had the user just concentrated the $2000 in the token they originally sought out exposure to with none to the pair token, he would have had $3500, a difference of $500 to the current value of his / her LP tokens and a return of 75% or 50% better than the 50% return of pool 2. This information is a hidden risk to the user and yet they may never know, especially when exchanges prioritize traders' needs over LPs in service of amplifying trade volume, user numbers and “revenue” but care little about educating LPs about the dangers of IL. We are all joyful little fools burning our hands on the stove, but lacking any pain receptors.
With that aside complete, let’s talk about how the AMM space has evolved at a high level. With a grant from the Ethereum Foundation at hand, Uniswap cleverly forked Bancor’s concept and solved their chicken and egg problem. Few users held $BNT in their wallets and the users that might have supplied tokens did not see why they should acquire $BNT to do so. So Uniswap decided to replace $BNT with $ETH as the common pair (or reserve asset) because it was the most held token in DeFi. If users already had $ETH in their wallets, there would be less of a percieved tradeoff with their capital and less friction to liquidity provisioning because it was already in their wallets. For me, that choice invites evaluation as to whether ETH or L1 tokens generally are the most performant pair token or were simply the most expedient and available token at the time.
However, there was another problem. The liquidity that was supplied was distributed over a very wide price range. Since the pool was cooperatively owned, everyone earned proportional yield for any trade volume going through the pool, even though some of the tokens were not being used. This was another form of “capital inefficiency”. I am breezing by much technical complexity, but with the introduction of “concentrated liquidity”, users could supply their tokens to a more narrow range of price that a token buyer was most likely to purchase within to minimize any tokens sitting idle. This allows the pool to act as if there were more tokens for trading than there actually are. Concentrated liquidity is most explicitly useful in the case of the trading of stablecoins - tokens backed by or explicitly pegged to fiat currencies. As they are supposed to trade at or around $1, the further down or up the curve tokens are supplied (ex. $.95 or $1.05) the less likely the tokens are to be used for trading and earning yield unless the peg were to break. Now with the introduction of Uniswap V3, liquidity providers could concentrate their liquidity around any price range, both for like kind assets like two stablecoins or for any other asset (often paired with ETH).
While an impressive technical achievement, this has had several unintended consequences. First, the increased sophistication and associated UX seemed to directly contradict the intended purpose of the automated market maker and even defies the cooperative ownership principles of DeFi more broadly. The AMM as a tool was supposed to not only enable us to trade a diverse set of currencies with one another (originally intended to empower local community currencies), but also to bring about a world where anyone could participate as a market maker by supplying their tokens and be able to do so without complexity. With UNI V3, users now need to have technical expertise. Users can no longer supply liquidity without persistent oversight. With V3, you must make the decision of what range you want to supply your tokens to. The more concentrated the liquidity, the more fees you possibly get, but you run a higher risk that the price will go out of range and have to incur cost to re-stake your liquidity into a different range in addition to earning no fees when out of range. If liquidity provisioning is effectively going short on two tokens, this is like going short with 20x leverage. This is clearly a professional’s realm. Even if it worked to perfection, what have we done, but create a decentralized exchange in governance power only with centralization of market making and profiting? The UX, both in cost and complexity, is at odds with the values of democratized market making.
Recent studies confirm the old truth that risk can never be eliminated, but it can be transformed. So concentrated market making is not a free lunch, it simply amplifies IL. If LPing is effectively going short on your tokens, concentrated liquidity is shorting with leverage. The Topaze Bue study named “Impermanent Loss in Uniswap v3”, analyzes both passive liquidity providers and more active liquidity providers who seek to manage impermanent loss. Regardless of the time they provided liquidity, 53% of passive and active liquidity providers were losing money relative to holding value. It further found that the only group of users that is consistently making money (outpacing IL) are those that are providing just in time (JIT) liquidity, which effectively means all other user groups are getting pick pocketed - passive and active alike. Might the ability to remove liquidity with no cool down period prove to be a bug not a feature? On a long enough time horizon, can any exchange that consistently acts against the interests of their system’s laborers (the LPs), re-centralizes market making and defies the ownership economy thesis have a case for survival? Furthermore, if the exchange’s native token itself is not used in functioning of the system, does not backstop any risk and can be forked by anyone, how legitimate and defensible is their claim to a sizable and defensible profitable position?
The next logical lego of the story is a brief explanation of the popularity of yield farming / liquidity mining. Every network and associated network effect has a 0-1 moment, a jumpstart. DeFi protocols face a similar fundamental chicken and egg problem - how do you get users before you provide any real value and how do you provide value before you have users? In crypto, where the main “job to be done” is making crypto assets productive, the native growth strategy requires attracting a lot of liquidity to make applications more performant. While I do not agree with the means they use to achieve this end, I am intrigued by the analogy of liquidity to bandwidth that the Tokemak team makes. The higher the liquidity (bandwidth), the more users and institutions can participate in these open networks.
With an understanding that to attract users you need liquidity, the crypto native growth hack is liquidity mining. You can mine for liquidity by tapping the future value of your protocol’s token and offering it to early users of your product as a yield multiplier on specified platform actions. Liquidity mining and yield farming are essentially two sides of the same coin. The protocol mines for liquidity by offering token rewards and users “farm” that yield by depositing liquidity. For example lending USDC might give you x # of Compound $comp tokens to add to native yield or as a reward for lending, but Compound might also give $comp tokens to the $comp / eth liquidity pool on an exchange to incentivize pool depth. An analogy I would use to describe liquidity mining is the Uber of 2011 used rider discounts and a combined supply-side strategy of cash driver bonuses / higher initial take rate to incentivize early usage. 2022 Uber would have just given away native $Uber tokens directly to bootstrap a network effect on both sides before the organic fees for drivers were sustainable and before the supply of drivers was sufficient to drop wait times and prices, making the service valuable enough to users.
For the purposes of this piece, I want to specifically focus on the use of a token’s native treasury to incentivize liquidity pools on decentralized exchanges. If exchanges like Uniswap and Sushiswap are the Coinbases of decentralized finance, then liquidity mining has been the cost of staying listed as well as cultivating a loyal and motivated community by rewarding early users with ownership. Such a growth initiative is only valuable if that liquidity actually sticks around when artificial incentives run out. To illustrate a point and avoid maligning any individual protocol, I have chosen to create the $patrick token to represent a new cross-chain derivative exchange liquid staking yield optimizing AI enhanced bridge in the Metaverse. Super bullish.
P.S. Some fundamentals here: [insert influencer + shamelessly shitty VC] are in, Binance and Coinbase listings soon so the team is very excited. 100% confirmed soon moon ser. Must be nothing. Stay poor Anon. But also WAGMI fren.
*If that meant nothing to you, you’re lucky*
So in order to compensate people for buying and holding $Patrick and taking on exposure risk + IL risk to supply initial liquidity to an AMM, I have chosen to give a predefined number of $Patrick from my cap table to $Patrick / $ETH stakers on an AMM. All of the sudden the 5% they were making in native yield from swap fees annually (pre IL calculation) is upwards of 50%, no 1000%, no 150,000%! I even saw a team try infinity %. This is an internet protocol’s equivalent of a car dealership inflatable or sky light - a flash in the pan, fireworks. Pure attention mining. So now the majority of the return is coming in the form of a native token, $patrick. From the perspective of a protocol thoughtful (and honest) about treasury management, this token allocation should be purpose oriented to accelerate user acquisition, hiring, and deepening exchange liquidity etc. However, most token incentive programs are poorly designed and lead to a destruction of short and long term equity value.
So as we said before, in order to access the pools on exchanges where liquidity mining rewards are distributed, users are required to stake 50% Token A and whatever the paired token is (often ETH or another L1 token). Imagine I stake $1000 - $500 in $Patrick and $500 in ETH, but my returns for swaps are $5 in organic ETH fees after a week, but a whole $100 of $Patrick earned, a protocol which is spitting off a faucet of 10% per week, 50 gazillion APY per year. At this point I can choose to “compound” or I can sell. If I sell, I am dumping the full amount of 100 $Patrick into the pool creating downward pressure on the price of the token, but if I choose to compound and re-restake, I still have to sell half my $Patrick to pair with $ETH in order to create the desired 50/50 pool. Both create sell pressure for the protocol token and buying pressure to the native L1 token or exchange token that it is paired with and frankly, it’s just extremely annoying UX that creates multiple unnecessary and sometimes very costly transactions. Furthermore, as I sell off rewards to re-stake, I forfeit the ability to have concentrated compounding exposure to the token of my choice.
As protocols jockey with one another for users who have infinite opportunities to supply their capital to other undifferentiated commoditized software protocols, they raise APRs as a marketing expenditure and quickly find themselves in a mimetic arms race to infinity with other protocols. If we assume that either re-staking or wholesale selling of rewards creates immense sell pressure on the native token, you can start to see the negative feedback loop emerge. Stake 50/50 $Patrick/ETH. I (a loyal narcissistic hodler) sell half my $Patrick to combine with $ETH to re-stake. Another user, chooses to dump all their rewards putting 2x the sell pressure on the pool that I did. Both the long term staker and mercenary reward seller alike are contributing to a vicious feedback loop. It looks something like this; the competitive attentional commons require massive APYs to stand out, which creates massive sell pressure regardless of yield farming action (re-stake or sell) which leads to lost value for shareholders (token holders) from price pressure, but impermanent loss as well (lost principle) from the divergence of the tokens. With these new risks in play, the protocol then gives a higher APY to compensate which just accelerates the negative feedback loop. More importantly, this leads to a game theoretic outcome that regardless of what the protocol is, what APY they are using or which exchange, the situation eventually leads hodling liquidity providers to become more and more mercenary, farming everything like a band of locusts and outpacing IL if and only they were first in first. The behavior of the holder was altered by the behavior of the more mercenary participant who set the tit for tat rules. Furthermore, the protocol’s only chance of curtailing this behavior, if there is no application native use for their token in the protocol, is to create an additional single side staking mechanism for users to “compound” their rewards into and ensure it is not sold for an immediately realized gain, leasing to more token emissions. As tokens in the single sided contract are not able to simultaneously be used as liquidity, the only main solution for stopping pool 2 yield farming sell pressure further increases emissions and potentially undermines their original liquidity goals.
Once everyone had been fully conditioned in the FIFO (first in first out) liquidity provisioning method deterministically caused by the construction of the DEXes upon which their token rewards programs were administered, there naturally arose the phenomenon of perpetual vampire attacks. As yield farmers devour the rewards like locusts, rewards soon deplete and liquidity on exchanges immediately dries up. Even high quality projects can wind up dead on arrival. A peculiar thing unique to the overlapping worlds of open source software and crypto economic systems is that the story often does not stop there. As soon as the well dries up, developers would simply “fork” or copy the code switching up either the blockchain the protocol was launched on or the tokenomics and slapping on a new brand to the door of the establishment. X Protocol with Y Tokenomics on Z Chain - change one and you have a 20m Series A.
There are seemingly endless forks of just about every original innovation and while exceptions exist, forks are often in search of new narratives to artificially stimulate demand rather than solving existing user needs. VCs as the main institutional capital in the space desperate for new asymmetric return profile skipped bidding on the DeFi blue chips and went for a much played out L1 foundation + ecosystem fund strategy on rinse and repeat. This creates a reflexive narrative in the process as the first core protocols on the new chain launch. Immense buying pressure for the L1 token is created by newfound demand for use as the reserve asset on a new ecosystem’s native exchange, to lend or borrow in money markets and to borrow against to mint new magical stablecoins. In addition, the construction of Pool 2 yield farms themselves on exchanges inevitably leads to buy pressure for the paired L1 token as rewards are sold to buy more L1 tokens to restake. The initial pump of the L1 token as the first Dapps launch increases the absolute value of both the L1’s own ecosystem fund and the VCs’ furthering a reflexive loop for them to increase capital ownership.
To close the loop here, it is worth meditating on the idea that a certain percentage of the perpetual vampire attacks were almost certainly directly caused by greedy influencers and their tribe of retail investors as well as VC-driven narrative building driven by fund economics, concentric pump circles etc etc, but perhaps the largest cause (and most culpable party) is the system itself, the Pool 2 construction that leads to a system that inevitably moves deterministically towards these ends. Maybe we’ve had the chain of causality wrong all along? Maybe Pool 2 amplified greed and perishability of protocols much more than would have otherwise occurred. And who benefited? Dexes taking a risk free standing by while their LPs got pickpocketed and L1 tokens, whose valuations surged well beyond reasonable fundamentals driven multiples driven by DEX pairing demand and buy pressure caused by pool 2 farming.
If it feels like things are moving too fast, it’s not because there is too much quality. It’s because there is too much noise. The structure of the system led to a repeatable process by which protocols would be milked of all their native tokens, but a few founders, insiders and fast twitch traders were able to benefit from being first in first out. Incentives like this led to the creation of thousands of protocols where the sole purpose was to copy an existing codebase, layer on a farming program (maybe on a different chain to at least feign differentiation) and benefit from a 20% ownership in a flash in the pan that was simply a venus flytrap for greedy speculators.
Sigh. Ok so maybe then we can establish in hindsight that the largest embedded flaw of current decentralized exchanges is that they require a single user to seed both sides of the pool.
The current system’s construction of exchanges and associated yield farming programs lead to a few unintended and undesirable outcomes.
User Experience Friction
a. Acquiring both assets
b. Getting them onto desired chain
c. Balancing token quantities with farming sales to restake and associated multi-transaction costs
d. Involuntary Token Concentration - Must take on Linear and Equal exposure to Layer 1 tokens to get access to the higher AMM yield (swap revenue + farming rewards)
e. Multiple transactions cause unnecessary active management + costly transactions
Token Reward Program Induced Sell Pressure
a. Any growth programs built to incentivize liquidity with Pool 2s ultimately undermine themselves. You need to have liquidity on decentralized exchanges. The incentives for many protocols are competing with a number of other identical or commodity products, so the rewards must be high to get mind share and TVL (eyeball hours and hit rates of the 2020s). Liquidity mining / yield farming becomes a drug because you don’t want to stop, but the two sided staking model necessitates sell pressure on the token’s price regardless of whether someone compounds or sells. Any % selling ultimately trains everyone else to dump as well as the game theory becomes well known. To curtail the dump, the main solution is a separate single sided staking contract, so rewards for providing liquidity cannot in turn be used to generate additional liquidity. Now we are back to square one. A bunch of tokens spent and no long term committed community members present other than the loyal LPs who experienced a ton of impermanent loss like dopes.
Capital Productivity / Concentration tradeoff
a. Concentration / Productivity Frontier - The user has to choose between being 100% in the token allocation it wants to or supplying that full amount to an exchange
Impermanent loss
a. Amplified by sell pressure from farming
b. Barrier to entry - Technical skills required to mitigate IL at odds with democratization of market making.
Summary DeFi 1.0 : Many protocols are back to square one - few long term token holders and community members, depleted treasury from high cost of capital, still no sticky liquidity for your native token on DEXes.
Thanks for reading! If you want to support or contribute to our mission to reduce noise and create thoughtful DeFi research, please follow us on twitter @analystdao and join us in Discord @ https://discord.gg/eAkrN4XN.
Follow @patfitzgerald01 to get notified when Part 2 drops.