Chris Dixon recently tweeted this thread about how Web3 represents an opportunity for new projects to successfully compete with existing ones by charging lower “take rates”.
In my view, Chris’s thread is a perfect example of what’s wrong with today’s blockchain ecosystem. It demonstrates the kind of flawed reasoning that’s led to an explosion of funding and attention for worthless projects.
Let’s take Chris’s tweets one by one.
2/ Jeff Bezos famously said “your margin is my opportunity” referring to the way Amazon took market share by lowering prices and eating into competitor margins.
3/ What Amazon did in commerce is what the internet did more generally. Lowering prices and redistributing value back to users has been the internet’s core economic dynamic since the 90s.
4/ Craigslist did this with classifieds, Google and Facebook did this with media, TripAdvisor and Airbnb did this with travel, and so on.
So far, I agree.
5/ Today this trend continues as Web 3 startups begin to eat into the margins of Web 2 incumbents. The higher the take rate, the more vulnerable the incumbent.
6/ The video games industry does about $120B/year in sales, a significant portion of which is virtual goods. Most video games have 100% take rates.
7/ Web 3 (aka crypto) games reduce the take rate dramatically. For example, Axie Infinity has generated over $1B in gross sales in the past year, most of which has gone back to users.
8/ In most video games, some people pay to get ahead and other people work to get ahead. The difference in Web 3 is that the economy is peer to peer: players fund other players, not just the game developers.
I don’t think it’s relevant that “most video games have 100% take rates” when there are popular contemporary non-crypto video games like Roblox that have only 30% take rates. Axie’s take rate is only about 4.25%, according to this interview, which does indicate that there’s a trend toward lower transaction fees for companies hosting virtual economies.
My beef is that Chris seems to want to credit the difference in take rate to Axie’s use of blockchain technology. He’s saying, “look, Axie is a blockchain application and Axie has a low take rate”, but he’s not making any logical connection between a company’s ownership-ledger technology and its take rate.
A trend toward lower take rates is simply what we expect to see in all competitive industries. For example, Robinhood introduced commission-free trading in 2013, long before it had any crypto features. Over the years, most other brokerages have also started offering commission-free trades in order to stay competitive. That’s not surprising, and doesn’t require blockchain to explain. But I feel like Chris’s MO would be to make a tweet today like, “look, Robinhood allows you to trade Web3 assets like Bitcoin and Ethereum, and unlike traditional Web2 brokerages, its take rate is 0%. Welcome to Web3 where your take rate is my opportunity!”
Let’s turn to Chris’s next supporting example.
9/ Today there are over 8 million musicians on streaming services, yet less than 15,000 musicians (less that 0.2%) make more than $50K/year. That’s because the vast majority of the revenue is kept by the streaming services and music labels.
10/ With NFTs, musicians keep over 90% of sales. By cutting out layers of intermediaries, musicians can credibly support themselves with just a thousand true fans: https://cdixon.org/2021/02/27/NFTs-and-a-thousand-true-fans
I don’t understand how Chris’s vision of “cutting out layers of intermediaries” is meaningfully different from what already exists today.
Today, if you know which artist you want to listen to, it’s easy to “cut out the middleman”: you just go to that artist’s website and buy their songs, albums, merchandise, concert tickets, etc. The musician can keep “over 90% of sales”. No need for NFTs.
Chris seems to be implying that streaming services such as Spotify are capturing too much value by occupying a middleman position between users and their songs. Currently Spotify makes about $0/yr net income on $10B/yr revenue. How much less value does Chris think Spotify should be capturing?
Chris’s other assertion here is that the music labels are keeping too much value. But why do artists use music labels? The music labels offer them some kind of distribution advantage; otherwise, the artists would just handle their own distribution. How do artists get the level of distribution they’re used to in a world where all music lives on the blockchain as NFTs? Having your music sitting on the blockchain with no music-label fee might feel empowering, but it’s not clear how it’s any more empowering than the currently-available option of having your music sit on your website. Whether or not NFTs are in the picture, the hardest challenge for artists is how to compete in a market where the supply of songs far exceeds the supply of hours per listener per day.
Surprisingly, Chris makes zero mention of the fundamental distribution problem that musicians face in his tweetstorm. Maybe the answer will be elaborated in his linked blog post, NFTs and A Thousand True Fans? In the post, Chris lists “three important reasons why NFTs offer fundamentally better economics for creators”:
- Removing rent-seeking intermediaries
- Enabling granular price tiering
- Making users owners, thereby reducing customer acquisition costs to near zero
I’ve already addressed #1: I haven’t seen an argument that explains why Spotify’s “rent-seeking” is excessive. Spotify today just seems to be capturing a fair piece of the value it creates as a well-crafted software client built to serve the needs of music listeners.
#2 is not much of a value-add compared to the status quo: Artists today simply use Spotify as the lowest-price option to monetize the listening time of their not-so-true fans, and their (Web2) home page gives them plenty of flexibility to upsell a broad spectrum of higher-priced options to their true fans.
#3 is the alleged benefit that Chris says is “the most important”, but also the one that seems to me like the most illusory: “Making users owners, thereby reducing customer acquisition costs to near zero”.
Let me try to describe a specific example of what Chris is talking about. Say I’m listening to a Linkin Park song that I own an NFT of. In this scenario, I guess Chris would say that Linkin Park has “made me an owner” moreso than I am today — moreso than by purchasing their song on iTunes or by purchasing their album from their website. Hmm, I’m not sure I agree, but let’s finish thinking through Chris’s assertion. In addition to saying that I’m an owner of this Linkin Park song, Chris also seems to arrive at the conclusion that Linkin Park has paid near zero to acquire me as a customer. This seems both false and a non-sequitur?
Maybe instead of using Linkin Park as an example to understand Chris’s assertion #3, we should consider the case of a new musician who no one has ever heard of. Is there some way that NFTs can make this new musician have near-zero costs to acquire me as a listener? Chris answers this in the affirmative by citing… the success of Bitcoin and Ethereum?
Open any tech S-1 filing and you’ll see massive user/customer acquisition costs, usually going to online ads or sales staff. Crypto, by contrast, has grown to over a trillion dollars in aggregate market capitalization with almost no marketing spend. Bitcoin and Ethereum don’t have organizations behind them let alone marketing budgets, yet are used, owned, and loved by tens of millions of people.
Chris seems to be arguing that since Bitcoin and Ethereum became well-known without having a marketing budget, we can therefore expect an undiscovered artist to successfully compete with millions of other artists to get themselves on my playlist and monetize my attention, somehow sidestepping the fundamental laws of supply & demand that currently prevent the world’s million musicians from all getting free access to my listening-time.
Sorry, I don’t understand how this makes any logical sense.
Chris offers just one more piece of support for his assertion #3 about NFTs reducing customer acquisition costs:
The highest revenue NFT project to date, NBA Top Shot, has generated $200M in gross sales in just the past month while spending very little on marketing. It’s been able to grow so efficiently because users feel like owners — they have skin in the game. It’s true peer-to-peer marketing, fueled by community, excitement, and ownership.
Bitcoin, Ethereum, and NBA Top Shot had low marketing costs because you’re selecting the top projects, the ones that broke out and achieved viral growth. If we consider the whole ecosystem of crypto projects and NFTs, most projects are languishing with zero users. Just because 0.1% of blockchain projects and NFTs have gone viral, that doesn’t mean the blockchain has a magical marketing engine that millions of musicians can all simultaneously take advantage of.
Chris makes the same claims about Twitter, Instagram and TikTok that he made about streaming services like Spotify:
11/ Social media platforms like Twitter, Instagram, and TikTok have take rates of 100% — they don’t share any revenue at all with creators! That’s been great for them but bad for users.
12/ In contrast, Web 3 social platforms like Rally, Mirror, and BitClout have effective take rates well below 10%. Most of the value is sent back to users and creators.
Let’s unpack Chris’s claim about “Twitter’s take rate”. When a user tweets, a bunch of people get some value by reading it. Those readers give back value by focusing some of their attention on the ads in their tweetstream. Chris is pointing out that Twitter keeps 100% of the proceeds from selling this ad-attention to advertisers.
YouTube’s take rate is only about 40%, so I agree that 100% is probably *not* a defensible take rate for Twitter to maintain in the long term. But I disagree about two aspects of Chris’s claim:
- I think Twitter can sustain a take rate that’s 20%+; it doesn’t need to go down to 0–10%.
- I don’t see how building with NFTs gives would-be Twitter competitors any extra advantage on the lower-take-rate front.
Twitter deserves a 20%+ take rate because it adds a ton of value as a distribution channel, i.e. the same reason Twitter is currently able to sustain a 100% take rate. Today, most people who write popular tweets are already capturing plenty of the value of those tweets by building their own brand. There are already many indirect ways to monetize a Twitter audience: having a big enough Twitter audience enables you to rapidly make tons of money selling them books, courses, consulting, etc.
Furthermore, Twitter has already started lowering its take by launching Super Follows and Tips. I believe Twitter is following in YouTube’s path and will eventually stabilize its take rate at something fair.
Chris claims that a platform like BitClout, by virtue of being “Web 3”, has some kind of advantage that lets it threaten Twitter’s dominance by offering a take rate below 10%. This claim doesn’t have any empirical evidence (yet?) because its engagement metrics haven’t begun to challenge Twitter’s.
BitClout’s low take rate evidently hasn’t been sufficient to differentiate it from Web2 offerings. I see this as part of a consistent pattern that what Chris sees as a “Web3 opportunity” is not a real market opportunity, but just a mirage.
Here’s the final tweet in Chris’s thread:
13/ Web 2 platforms depend entirely on creators for content, yet give only scraps back. This is not sustainable. Web 2’s take rate is Web 3’s opportunity.
I would ask Chris: If you really think Twitter “isn’t sustainable”, and a Blockchain-based version of Twitter *is* sustainable, why not directly advise Jack Dorsey to move Twitter to the Blockchain? Why didn’t you use your thread to call out a specific Blockchain-based Twitter competitor that you’ve funded in accordance with your “Web 2’s take rate is Web 3’s opportunity” thesis?
Here at Bloated MVP, I specialize in pointing out when a company or investment is bad not because of its execution, but because it neglected to have a logically-valid value proposition at the idea stage. In the last few years, the crypto space has turned into Bloated MVP heaven. It’s full of smart people writing out chains of logic that contain gaping non-sequiturs, and using those as the justification to throw away $millions in dollars and effort. In this post, I believe I’ve documented an interesting and representative example of what a potent but logically-invalid pro-crypto argument looks like.