Lifetime Value for Crypto Businesses
How to think about Lifetime Value for Crypto Businesses
Around this time last year I started writing about crypto businesses and how their unit economics were broken in a way that rendered most of the industry a joke. It wasn’t until the collapse of FTX that people really started paying attention. You should give these posts a read to get a better idea of where my headspace was back then:
Since then I’ve spent a lot of time not just thinking, but building the required components to solve these problems. We’re still a few months away but I’ll be sharing some increasingly tangible things as we crack these problems once and for all. Until then I’ll share some extra thinking that I’ve done on conceptual frameworks that help think through these problems.
Lifetime Value (LTV)
Arguably one of the most important metrics for any business — how much is one customer worth in terms of fees earned over their lifetime. Now the nuance here is that each business earns fees in a different way so their mechanics need to be understood in nuance to understand where value is created and subsequently captured. Below is a rough conceptual framework for how to think about the problem for different categories.
These are all on a per-user basis so of course it’s common sense that more users will net more revenue but the key component here is that the quality of these users matter since you’re dealing with averages at the end of the day.
Chains
Even though networks ≠ companies in their purest sense, knowing how much in fees each user brings on average can tell you a lot about the health of the chain. This poses a challenge for high throughput L1s and L2s that have much lower fees than a network compared to Ethereum. They need to either show much higher activity volumes or add complimentary services that earn majority of their revenue (think movie cinemas where the movie is at cost but the real money is generated at the snack bar).
For chains/general purpose compute platforms the equation is:
Fees Paid = Complexity of Transaction (gas) * Price of Computation (gas price)
Meaning the two dimensions compute platforms (chains and L1s) should be thinking about is:
How can I increase the complexity of the transactions happening? DeFi is one example, on-chain games would be another example (as long as the game play is happening on-chain).
How can I increase the average price of compute (gas prices) consistently. When there’s extreme price volatility Ethereum rakes it in with fees. Outside of that it’s earnings start to reduce. Having consistent activity is another dimension to increase fees paid per user.
This isn’t a exhaustive framework but rather levers that should be studied.
DEXs
DEXs have a simpler framework for understanding their customer lifetime value since it really comes down to:
Fees Paid = Size of Transaction * Fee Applied (%)
The two dimensions that DEXs need to optimise on is:
How do I increase the average size of transactions happening on my DEX? This can be by specialising in a particular vertical such as stablecoins (Curve’s strategy) or long tail speculator coins (Uniswap). Although both are trying to each other’s share.
How do I increase the average fee I can charge per trade? This is where NFT marketplaces have shot themselves hard by competing in a race to the bottom business where not only are fees 0% but unit economics are negative due to incentives being paid out. We’ll cover CAC (cost to acquire customers) in a later post.
Lending
Any lending/borrowing protocol has the same unit economics and how they solve these challenges is up to them, but critically speaking their equation looks like this:
Fees Paid = Profit earned by lenders (%) * Fee Captured of profits (%)
With these two dimensions, you need to be doing serious scale in order to have a business that makes sense:
How do I increase the profit that lenders can earn? This is probably the most under-optimised variable since borrowers are relatively price insensitive for markets where the underlying asset can 10x in a year (not true for stablecoins). However, very few borrowers are willing to pay north of double digits which means the best case is 10% (and that’s on the higher side).
How do I capture a % of the profits earned by lenders? This makes things tricky because margins are already so tiny and then capturing a percentage of the margins makes things even more difficult. For example, Aave takes 10% of the profit earned by lenders but assuming that lenders earn 5% for stables, earning 50 basis points is the pure earnings for Aave and the value of that customer.
Stablecoins
These are the businesses that everyone loves because it means you get a money printer machine. More often that not though, to get these businesses off the ground you need to spend a lot of money in acquiring customers through liquidity incentives. However to demonstrate why these businesses are lucrative you only need to look at their fee structure:
Fees Paid = Interest rate charged * Dollar value of asset being borrowed
That’s it. You set the interest rate and you encourage borrowers to borrow the most they can. When you think about your two levers you simply need to understand:
How do I ensure I can charge the highest possible interest rate while still being competitive? This is largely dependent on what the competition is doing (money markets) and what my costs are for liquidity.
How do I get borrowers to borrow the most they can? This is challenging because the quality of asset will determine the risk of the platform and your ability to stay solvent.
Any money you earn is your profit. The only issue is, your costs are going into ensuring your stablecoin stays at peg or has solid demand sinks baked in.
Yield Aggregators
This basically covers any service that says “give me your money and I’ll make you the most money”. These were great businesses when they first came out but the biggest challenge here is defensibility and negative network effects. The fee structure is similar to lending businesses but there’s a difference in their underlying business mechanics.
Fees Paid = Size of deposit * Profit earned by depositors * Performance fee (%)
The reason why I included size of deposit here is because the more money a yield aggregator has, the less yield it produces for its other users. It actually has negative network effects!
As a yield aggregator, you obviously want lots of $ so you can earn your depositors more profit and then capture your performance fees. The only challenge is that as you don’t want too much money because you’ll struggle to make increasing profits for all your depositors which drops your performance fees.
In order do increase the profit earned by your depositors, the largest challenge you have that any strategy you deploy on-chain can be copied by anyone else — and very quickly. You get squeezed on both sides unfortunately.
Performance fee. This is also a really hard vector to optimise around since any % you charge your depositors, technically they can cut you out and go straight to the source. It’s like LPs investing directly into startups that their VC funds find.
Closing
As you can see, each of these crypto primitives offer very similar to businesses you see in the real world but have slightly different mechanics due to the nature of the environment they live in. If this post gets enough traction, I’ll release a second part around the costs — since that’s the part that everyone forgets about when factoring in their equations. Also the relation between costs and profits is a delicate one where an element of “incentive elasticity” comes in. Anyways, I’m running ahead of myself here. I’d love to hear your thoughts on this piece!