(Amazon) Prime time in web3
The benefits of tokenized membership bundles in building great consumer products
A few weeks ago, major crypto exchanges met with congress to talk about regulation of a movement that now has over 3 trillion dollars in its bank account. And while it’s easy to laugh at the 70 year old politicians trying to understand NFTs, we as builders in web3 haven’t exactly made it easy for others to imagine the future as we see it — the one worth fighting through the chaos, scams, and regulatory risk to create.
Consumer products are one of the most exciting but difficult to imagine aspects of that future. Crypto theoretically gives us the thread to weave a rich tapestry of community owned platforms, incentivized for user value rather than advertising dollars. But nobody has really figured out how these companies create defensible and sustainable businesses that can fund the huge effort it takes to build at that scale and complexity. And perhaps most importantly, we haven’t fully described what those businesses will feel like to us as consumers - why should we be excited?
Clear language is a huge catalyst in the process of innovation, so I’d like to try and offer a basic mental model: I believe some of the leading consumer product business models in web3 will feel like Amazon Prime — membership bundles mediated via tokens. This is in a sense building on my arguments around tokenized SaaS in general and for media companies specifically. If those were about using tokens to build initial traction, this is about how that traction can further encourage valuable products or protocols under the same subscription.
Fat protocols, sticky applications
To rewind a bit, the reason the topic of how consumer applications build businesses is up for debate is because value in crypto has primarily been captured so far on the protocol layer, essentially in transactions that happen behind the scenes of the apps we use. Joel Monegro at USV coined what is now a seminal concept in web3 on this point, the “Fat Protocol Thesis:”
“The market cap of the protocol always grows faster than the combined value of the applications built on top, since the success of the application layer drives further speculation at the protocol layer. And again, increasing value at the protocol layer attracts and incentivises competition at the application layer.”
I don’t disagree with this theory, but I do find the language somewhat unsatisfying in terms of how to think about the incentives for building valuable applications. It goes on to suggest that the primary motivation for people to build applications is to increase the value of the ownership they already have in the protocol layer:
“Then some of these early adopters, perhaps financed in part by the profits of getting in at the start, build products and services around the protocol, recognizing that its success would further increase the value of their tokens.”
This would seem to mean that if you’re an app developer who wants to jump into web3 and build a valuable app on top of ethereum, you’re pretty much out of luck. Obtaining enough ownership in such a massive protocol to then benefit as its value increases would take a prohibitive amount of capital, not to mention that a single valuable app is unlikely to do much to change the value of Eth in any noticeable way.
My other issue with being too protocol focused is that if we can’t capture much value on the application layer, it will simply lead to shitty apps. As anyone who has tried to figure out how to navigate the world of web3 knows, we need to strive for vastly better and easier to use consumer applications than we have right now. Product builders who optimize ruthlessly and primarily for the user experience rather than an underlying protocol should and I believe will be able to thrive in web3.
All that said, I actually agree that a lot of what we have so far should be optimized for value capture at the protocol layer. Most of the widely adopted products right now are extremely transactional - apps that help you exchange tokens for example. As Jesse Walden correctly points out, these products should commoditize their applications in order to drive more demand for their truly differentiated product, the underlying exchange.
So the meta point here is that if web3 really is the future, I believe it will look remarkably less transactional than it is right now. We will need applications built on crypto rails to connect with friends, find restaurants, share photos, and so on. Those products will still feed the protocols underneath them, but they will also be expensive to build and only possible if they can capture significant value on the application layer.
Enter membership moats.
Ready for Prime time
The Amazon Prime story is legendary (book rec). To reduce it to one sentence, after noticing how much more they sold during free-shipping promotions, Bezos defied the finance team and the fancy models saying it was a horrible idea and demanded a roll out of Prime across the US in a matter of weeks.
It’s difficult to estimate the share of Amazon’s success that’s attributable to Prime, but it’s safe to say the analysts’ concern that they would lose too much money on small purchases has been disproven. There are over 200 million prime members worldwide, generating north of $25 billion in revenue. The real story isn’t the revenue from Prime though; it’s what Prime does for the rest of Amazon’s product. Recent studies showed that Prime members spend roughly $2K on Amazon a year, 4X more than non-Prime members. There’s no way to prove causality given the numbers they supply, but from my personal experience it felt like Prime really catalyzed the movement to where people buy pretty much everything they need on Amazon, from toothpaste to TVs to movie rentals. It created an incredibly sticky set of products.
There are a few elements of Prime that are very important and I believe transferable to web3 consumer applications:
Prime membership simultaneously drives adoption and usage of transactional products (eg buying TVs on Amazon) as well as other subscription products (eg Prime Video) and infrastructure/protocols (eg Amazon warehousing)
Because Prime’s main purpose is to drive behavior on other services rather than revenue on its own, Amazon can ensure the benefits do not scale linearly with cost. In other words, the perceived value of Prime increases exponentially as Amazon adds more products and services underneath it.
There are two main differences to call out in how Prime type memberships might work in web3. First is that rather than simply having members, we have owners. Many DAOs such as Friends With Benefits are already operating in this way - holding a certain amount of token grants you access to the suite of products and services that the DAO offers. Early adopters can benefit from future demand, thereby encouraging early adoption.
The second difference is that in web3 it would be far easier to create a Prime type bundle using products you don’t actually own. Rather than having to authenticate into your Amazon account on another product, the interoperability of wallets means that any product could simply check to see if you hold the right tokens. A reasonable analogy would be this is a dramatically simpler to implement version of cardholder benefits for credit cards that give you access or discounts to things the credit card company doesn’t own.
I’m going to dive into the weeds a bit in a second, but first here are a few predictions you might make if you buy the viability of this model:
Aggregation, acquisitions, and partnerships will happen at the product layer as much or more than in web2, even given an interoperable and composable landscape. People will want fewer rather than more Prime-like memberships, so established organizations with already successful products will have an edge on building or partnering with complementary products they can slide under the same umbrella.
As consumers, our most valuable affinities will be to brands and communities, not specific products.
From application to protocol
I don’t think my argument here is actually counter to the fat protocol thesis. I’m simply arguing for a different route to that place, and for the importance of value capture at the application layer. In other words, you could consider Amazon’s global distribution system, recommendation engine, AWS, and warehousing all as protocols that capture huge value. But those were created because of the scale of Amazon’s applications, not the other way around.
I will note an obvious counterpoint which is that Amazon started with a highly transactional product, had revenue from day one, and might even look more like a Uniswap from the start if they were building in web3. Those things are all true, and I am actually not suggesting that web3 consumer applications will all follow Amazon’s exact path. I am simply arguing that leveraging a Prime-like membership to progressively move from great consumer UX to deeper protocol and infrastructure layers is a strategy that feels pulled from the Amazon playbook.
To make this even more concrete, imagine a specific case study. Let’s say I’m building the web3 version of Reddit via my company which I call The Prime Company. I need to build a sustainable business, so I consider a few options at both ends of the spectrum:
Value capture only at the protocol layer. Build and open up the protocol underneath the product and take a cut of the value that flows through it. The issue is that it’s not very clear in this case what that platform would be.
Value capture only at the application layer. Charge people for access to the product. The issue here is that it likely inhibits growth and adoption.
Membership is the middle ground. Let’s say this Prime-like membership is mediated through $PRIME tokens. It might look something like this:
The product is free for everyone to start. Perhaps $PRIME tokens are awarded to early adopters who help build the initial network.
At some point, $PRIME owners with enough tokens get some specific value only applicable to power users. For example, saving their favorite threads to a saved list.
After success in the first product, the team creates a related product which groups the threads from our Reddit product by DAO, allowing you to see all of the conversations relating to that community. Data flows between the two products, but there are specific features and data in each. $PRIME owners are again able to access some unique value, such as advanced analytics on DAO job openings.
The reason I think this is a middle layer is because as each new product is brought under the membership umbrella, the ability to create deeper value at the protocol layer increases. Over time $PRIME might build open protocols to handle threading across any product in web3, or a Karma system that can be implemented in any social product. But building those protocols doesn’t distract the team in the early days when they’re just focusing on providing value to users of their applications.
It’s still extremely early days for consumer products in web3, and I hope my thinking aloud here can help other builders who are tackling similar problems. Creating community-owned platforms and a healthier web is web3’s opportunity, not web3’s inevitability, and we all need to keep working to try and paint a picture of how the future will look and feel. My hope is that the magic of membership and of simplicity that people associate with Amazon Prime can help.
To summarize my takeaways and predictions from this article:
Consumer focused applications can move down the stack towards protocols and more transactional products, even when they don’t own a stake in those protocols from the start. Sticky user behavior is incredibly valuable, and that won’t change in web3.
People will want fewer memberships than more. Whether by acquisition, merger, or partnership, a few major membership tokens will emerge.
Those memberships will feel like Prime, where more and more of the products and services you need live under one umbrella. This will help make the UX of web3 feel less scattered.
Tokenized versions of Prime will enable a number of different ownership models that further incentivize early adopters. Rather than offering their products for free until protocols emerge, smart consumer applications can leverage these models to create sticky behavior and strong loyalty.
Thanks to Rapha Menezes for most of the thoughts in this post at one point or another, and to Jess Sloss for encouraging me to write it. I’m also building on thinking from Li Jin and David Phelps about moats in web3.