Tag Archives: Heterogeneity

The advantages and disadvantages of decentralized services

In an earlier post, I wrote about the hype around initial coin offerings (ICO’s) in general, and the specific danger of conflating the ICO investors of a decentralized app with its eventual users.

In this excellent post on cryptocurrencies, Adam Ludwin makes the same point:

“There is a pervasive narrative out there that supports entrepreneurs looking to create new crypto assets. The idea is that by selling assets to users before your network launches, you create “evangelists” who will be early users and promoters you wouldn’t otherwise have if there were no financial incentive to participate in your community.

The problem with this line of thinking is that it conflates early investors with early users. The overlap between people who buy your crypto asset and people who actually want to use the service you are building is likely very, very small, especially during market manias like this one. It creates a false sense of “product-market fit.” Yes, people are buying your crypto asset. But that’s because the “market” are people who want to get rich and the “product” you are selling is a “way to get rich.””

However, more important than this specific point, is Adam’s observation that:

“On almost every dimension, decentralized services are worse than their centralized counterparts:

They are slower

They are more expensive

They are less scalable

They have worse user experiences

They have volatile and uncertain governance

Except for one dimension.

Censorship resistance.

Censorship resistance means that access to decentralized applications is open and unfettered. Transactions on these services are unstoppable.”

In other words, Adam is stating that for most people, centralized services do just fine. As Adam states, “you cannot argue that for everyone Bitcoin is better than PayPal or Chase. Or that for everyone Filecoin is better than Dropbox or iCloud. Or that for everyone Ethereum is better than Amazon EC2 or Azure.”

But for those users who are optimizing for censhorship resistance, the tradeoffs of decentralized services are worth the censorship resistance feature that it enables.

In other words, there is heterogeneity in the homogeneity.

And that is why Adam concludes with the following statement, with which I fully agree:

“Don’t bet against crypto assets in the long-run: as we approach the 10 year anniversary of the Bitcoin paper it is clear that they aren’t going anywhere and that decentralized applications may very well find an important place alongside all the other forms of organization we have come to take for granted.”

The heterogeneity in the homogeneity

AngelList recently shared the performance results of its 2013 vintage of syndicate investments. Although the results are unrealized, they’re very impressive, with a 2.4X return multiple in less than 3 years and a 46% IRR. Keep in mind that this is the return from all AngelList syndicates without any insight applied to select the best ones.

This reminded me of the reaction of many VC’s when they first heard of AngelList. Many VC’s believed that AngelList wouldn’t be able to achieve great returns because the great investors who invest in great startups wouldn’t share these opportunities on AngelList. This would leave AngelList with mediocre startups.

Fast forward several years and AngelList’s aggregate unrealized returns from its 2013 vintage signal that this hypothesis is likely to turn out to be incorrect. Why were so many investors wrong?

One reason is that AngelList is a partial substitute to traditional VC money. The more money that the best startups raise from AngelList, the less there is left over for traditional VC. So many traditional VC’s see AngelList as a risk to their business and dismiss it out of a combination of self-interest and fear.

The second reason is that groups of people are not homogeneous. In this case, the group of great investors is not homogeneous.

Those investors who initially dismissed AngelList took this position because they believed that great investors wouldn’t share their opportunities to invest in great startups with others on AngelList. Instead, they would attempt to take as much of the round for themselves as possible.

For most investors, this is exactly what happened. Institutions like Sequoia Capital, Benchmark, and Accel don’t run AngelList syndicates. However, these aren’t the only great investors out there. A few great individual investors who have access to great startups did end up running AngelList syndicates. They did so because AngelList gave them the opportunity to invest more money than they could do so on their own while also earning a carry on this additional firepower.

And what you need to make something work is not everyone, but a few people. In fact, if you try to serve everyone, you’re unlikely to succeed. There are very few needs and wants that are universal.

Most needs and wants are shared by subsets of people. In other words, most groups are not homogenous. Specific individuals within what looks like a homogeneous group at the surface are actually heterogenous when viewed across a specific dimension. However, it takes time and effort to discover the sources of heterogeneity in what initially appears to be a homogeneous group of people. And therein lies your opportunity.

Most people don’t want to share the best startups they discover. But some do.

Most people don’t want to wear used clothes. But some do.

Most people don’t want to buy their homes or cars online. But some do.

Most people don’t want to host strangers in their home or let them in their car. But some do.