There are two mechanisms we know of where a business becomes more effective with greater scale. The first is network effects, and the second is economies of scale.
A social network is the classic example of a network effect in action. If I am the only person in the network, it is not very useful; if 10 of my friends are in the network, it is more useful; and if all of my current and potential friends are in the network, it is very useful.
Network effects are powerful because each incremental user makes the network marginally more effective in acquiring and retaining users. This means that the cost to acquire and retain one more user decreases as the network grows [1]. Becoming more effective with more scale is the ultimate form of defensibility.
Most crypto projects we know strive to achieve network effects, primarily in the form of liquidity network effects. AMMs, lending protocols, and NFT marketplaces all chase it. The theory is that more liquidity begets better pricing; better pricing begets more users; more users beget more fees; and more fees beget more liquidity.
Unfortunately, network effects are rare, and most businesses we know don’t have one. Instead, most businesses we know exhibit economies of scale, the other mechanism in which a business becomes more effective with greater scale.
Amazon’s fulfillment centers are a great example of economies of scale in action. If I’m the only Amazon customer in my town, the cost of delivering my package is significantly higher than if everyone in my town is an Amazon user. This is because Amazon can spread the fixed cost of delivery – warehouse facilities, delivery vans, and labor – across a greater number of users to drive down the delivery cost per customer [2].
Like network effects, economies of scale allows the product to become marginally more effective with each incremental user. Unlike network effects, economies of scale exist all around us.
Unfortunately, blockchains typically exhibit dis-economies of scale, where incremental usage makes the network marginally more expensive to use. Because blockchains have competitive fee markets for blockspace, more usage drives up the cost of blockspace.
One hypothesis I have is that we need to introduce economies of scale into blockchains to crack open a diversity of crypto applications. It’s hard to know why most crypto applications are variations of marketplaces that chase network effects. But only having access to network effects and not having access to economies of scale severely limits the types of defensible businesses that can be built onchain.
The big question is how we will achieve economies of scale. It will take creativity and tradeoffs between onchain and offchain components. For example, we created rollups to amortize blockspace across more transactions, but we still need to post all the data onchain. Is there a design where all data does not need to be onchain? To make tradeoffs, we will need to be thoughtful about what needs to be posted onchain, with what levels of verifiability, and with what levels of guarantees. Financial use cases, the current showcases of crypto applications, may not make sense to take these tradeoffs. But for web3 applications in social, gaming, and other verticals, economies of scale might be what we need to unleash their potential.
—
[1] Here’s an illustrative example of how this happens:
If a startup social network wants to grow from 1 user to 10 users (10x), it needs to invest in marketing to generate awareness and educate potential users about the benefits of joining.
Let’s assume the cost of marketing to acquire a user is $10. To grow by 9 users, the social network needs to spend $90, and the total customer acquisition cost per user is $10.
Now let’s assume that over the next few years, the social network has successfully grown larger. It now wants to grow from 1m users to 10m users (10x).
Let’s assume the cost of marketing to acquire a user remains $10. This time, let’s also assume that the 1m users manage to convince one other friend to join the social network so they can be on it together.
Because 1m users were acquired for free via word of mouth, the social network only needs to market to 8m users and spend $80m. The total customer acquisition cost per user is $8.89 ($80m marketing spend / 9m user growth).
Customer acquisition cost decreased from $10 / user to $8.89 / user as the network grew.
[2] Illustrative example #2 for economies of scale:
Let’s assume Amazon buys 1 warehouse, 1 delivery van, and employs 1 delivery person to service 100 package deliveries per day in a town.
Let’s also assume each user has 1 package delivered per day, and that Amazon operates at breakeven (i.e. zero profit).
If there is one Amazon user in this town, that user bears the entire cost of the 1 warehouse, 1 delivery van, and 1 delivery person.
If there are 100 Amazon users in this town, the cost of the 1 warehouse, 1 delivery van, and 1 delivery person are spread across 100 users, and the cost per user is 1/100.