Invest in the ugly duckling — decentralization, product/market fit, and the law

By John Backus

Product/market fit is complicated for blockchain companies due to the fundamental relationship between decentralized technology and the law. The same was true in the 2000s when people went crazy building p2p file sharing technology. When Napster died, blood was in the water and hungry entrepreneurs flooded the market with decentralized file sharing apps. By mid 2001, Napster was basically dead and Limewire, Kazaa, eDonkey, and BitTorrent were all available for download. If we followed the file sharing market from 2001 to 2006 and treated the companies like traditional early stage software investments, we’d fail to predict that BitTorrent would win the file sharing wars in the end.

Some of the most interesting decentralized technologies flirt with the boundary of the law. Privacy coins, prediction markets, censorship-resistant file hosting, and fiat-pegged stable coins are a few examples. These spaces might play out in a similar fashion to the file sharing revolution where companies brushed up against the law. We may want to be cautious about products in these spaces that see strong immediate growth that lead their space in adoption. We should look for the ugly ducklings.

Tech investors pay close attention to user growth for early stage software companies. Usually the key performance indicators are:

  • How many people are using your product?
  • How frequently do they use it? (daily? weekly? monthly?)
  • How much money do you make per user?
  • How fast is your user base growing?
  • Based on similar markets, how many users could you have one day?

The most well known companies in Silicon Valley get incredible valuations because their user growth is incredible.

If we adjust for the amount of people online in 2000 and how frequently they used the internet, Napster’s user growth may have been even more impressive than Snapchat’s legendary user growth in 2011. When Napster died, decentralized competitors like Kazaa grew at similar breakneck speeds.

Kazaa’s protocol, “FastTrack”, quickly replacing Napster — The Economist (2001)

Comparing Napster’s and Kazaa’s growth to Snapchat is important: breakneck user growth like Snapchat’s is rare. When it happens, investors are more than willing to take a risk on the growth even if profitability isn’t anywhere in sight.

The general public likes to hoot and holler at Silicon Valley investors for valuing companies like Snapchat in the billions when they don’t make much money. Not many companies have the daily attention of 190 million people though. When companies have successfully monetized network effects of this scale, it has both created and destroyed some of the biggest businesses in the world.

Y Combinator’s advice “make something people want” is well known and compresses a lot of less obvious pieces of advice. When you have a vision for what your product should be, most founders then go all in on building the product and don’t regularly ask their users for feedback.

“Talk to your users to see if it serves their needs, and then take their feedback and iterate. These tasks should occupy almost all of your time/focus”
YC’s Essential Startup Advice

This ties into other advice which focuses a lot on customer service and making people who use your product happy.

“Offer surprisingly good customer service”
Startups in 13 Sentences by Paul Graham

I could showcase a half dozen other quotes, but the core idea is that startup founders should always be listening to customer needs and then improving the product to service those needs.

Breakneck user growth like we saw with Snapchat and Napster are the goal. The core idea behind Y Combinator’s user-obsessed advice is that you’re constantly listening and experimenting with the goal of discovering massive untapped desires that existing companies aren’t servicing. This process isn’t just for executing on totally new ideas like Uber, it is also about finding even better fit by discovering that people actually want UberX and UberPool. In pursuit of growth, the name of the game is constantly honing in untaped desires and doing so better than everyone you’re competing against.

Breakneck user growth is the name of the game for software startups and the way to get there is listening to customers and giving them what they want. What did file sharing users want? All the free music and movies they could get their hands on!

In 1999, the legality of file sharing was untested and Napster wasn’t really built to be legally resilient. Kazaa, Limewire, eDonkey, BitTorrent, and many other file sharing products were built with the law in mind though. Companies deployed clever uses of cryptography and decentralization in order to protect themselves from the radioactive central search that killed Napster.

An obvious part of making products people want is actually making sure potential users know that your product fulfills their needs. Startups scrutinize what is displayed “above the fold” on their website because, if a potential user visits, this content will likely determine whether they try the product out or not. Take a look at the 2004 landing pages for Limewire, eDonkey, Kazaa, and BitTorrent:

One of these landing pages is not like the others! Limewire, Kazaa, and eDonkey at least give some signs that the user is in the right place by using keywords like “p2p” and “file sharing.” Limewire shows an example of downloading audio files, eDonkey mentions that users can download mp3s, and Kazaa explicitly mentions “music.” BitTorrent seems intentionally unclear and doesn’t even try to convey that it can be used as an alternative to the other three companies. If you visited BitTorrent’s introduction page, you would be led to believe that BitTorrent isn’t even a consumer product!

Wait, I thought my friend said I could get free movies with this?

Maybe BitTorrent was making something people want, but if you went to their website in the middle of the p2p file sharing wars you would have no idea. Remember, BitTorrent’s client didn’t have a search bar for discovering content either so downloading the client wouldn’t clear things up. If you visited their FAQ, only two answers even talked about how to find content. They linked to a website for sharing legal concert records and Google. If you wanted to learn how to download music and movies with BitTorrent, you were going to have to go somewhere else.

Despite BitTorrent’s actively confusing website, the BitTorrent protocol was actually taking the lead as the most popular file sharing website in 2004 according to an ISP level traffic study.

By June 2004, BitTorrent was 53% of p2p traffic according to CacheLogic

In court, Hollywood would often point to conduct by file sharing services as an indication of bad faith.

“This is about the fact that the sources are linking to are blatantly pirate sites which they are aware of. They link to sites that say ‘Super Pirated MP3s.’
RIAA: No Hyperlinking Allowed from Wired

The Seventh Circuit Court of Appeals latched on to Aimster’s tutorial as an invitation to infringe.

“In explaining how to use the Aimster software, the tutorial gives as its only examples of file sharing the sharing of copyrighted music, including copyrighted music that the recording industry had notified Aimster was being infringed by Aimster’s users. The tutorial is the invitation to infringement
Re: Aimster Copyright Litigation United States Court of Appeals

A bunch of companies were also called out for having a feature which let users search for “Top 40” songs.

users of Aimster’s software can join for a fee and use to download the ‘top 40’ popular-music files more easily than by using the basic, free service”
Re: Aimster Copyright Litigation United States Court of Appeals
Morpheus in fact allowed users to search specifically for ‘Top 40’ songs … which were inevitably copyrighted.”

MGM v. Grokster

Even providing good support was dangerous! The courts pointed to specific emails where staff from file sharing companies seemed to help users infringe on copyrighted content.

“StreamCast gave assistance to infringing users. For example, when users notified StreamCast by email that they were unable to play downloaded movies like “Tomb Raider” and “Shrek”, StreamCast employees responded with technical advice suggesting how to fix the problem. When an individual complained about the lack of music from Elvis and various other artists, [StreamCast support] told the user that “maybe you should load some up.”
Code Wars: 10 Years of P2P Software Litigation by Rebecca Giblin

The Supreme Court decided to judge companies based on if they intentionally tried to attract users interested in copyright infringement. The Supreme Court eventually ruled that each of the companies in the case was clearly going after former Napster users, even pointing to Grokster being named after Napster.

“The first factor was that “each [defendant] company showed itself to be aiming to satisfy a known source of demand for copyright infringement, the market comprising former Napster users.” As well as advertising to Napster users, internal company documents made reference to Napster; Grokster appeared to have derived its name from Napster, and the defendants’ software functioned similarly to Napster. The Court concluded that the defendants’ “efforts to supply services to former Napster users, deprived of a mechanism to copy and distribute what were overwhelmingly infringing files, indicate a principal, if not exclusive, intent on the part of each to bring about infringement.”
Code Wars: 10 Years of P2P Software Litigation by Rebecca Giblin

As soon as the Supreme Court introduced the concept of “inducement” which killed off the companies built on top of Kazaa’s FastTrack, public policy experts turned to BitTorrent and groaned.

the Court did not offer the clarity we might have hoped for, opting instead for what Tim Wu has described as the Miss Manners rule, under which vendors must avoid showing an unseemly interest in infringing uses of their products. This would appear to protect vendors who are honestly uninterested in fostering infringement, as well as those who are very interested but manage to hide it.”
BitTorrent: The Next Main Event (2005)

Was BitTorrent’s terrible website a good call after all? What about the fact that their download software didn’t have a search bar? I’m sure that both of these decisions really hurt BitTorrent’s adoption from 2001 to 2005 since you couldn’t actually learn how to use BitTorrent by going to their website and you definitely wouldn’t know how to find free music and movies if that is what you wanted. Still, BitTorrent seems to have insulated itself from the issues that killed these other companies.

Limewire, Kazaa, eDonkey, and BitTorrent were all launched when Napster finally died, but BitTorrent didn’t start taking the lead until about 6 months before the rest were killed off by the Supreme Court’s ruling. If we rewound to 2001 and watched these companies closely, BitTorrent would seem like the worst of the bunch until the very end. BitTorrent’s adoption was much slower and, judging by their website, they didn’t seem interested in giving users what they wanted in the ways that Limewire, Kazaa, and eDonkey were.

The BitTorrent company didn’t know that it would all come down to conduct though. Similar to the crypto space today, no one really knew what was going to happen from a legal perspective.

While companies like Kazaa and eDonkey decentralized the search index that killed Napster, BitTorrent outsourced it entirely.

“[BitTorrent] incorporates no internal search functionality, and creates no large proprietary network like that of the Napster and Grokster and Sharman technologies. These design features combine to ensure that no BitTorrent software provider could be liable for its users’ infringements under the existing contributory or vicarious liability law.
Code Wars: 10 Years of P2P Software Litigation by Rebecca Giblin

An earlier court case against the company “Aimster” may have gone to the Supreme Court if Aimster hadn’t gone bankrupt. Aimster was built on top of the AIM network and applied cryptography in clever ways that made it impossible for them to supervise what was going on within their network. Although these tactics were similar to evasive measures that other file sharing companies took, the Seventh Circuit Court of Appeals ruled this constituted “willful blindness” on Aimster’s part.

It deemed [Aimster’s] utilization of encryption to constitute willful blindness. “[A] service provider that would otherwise be a contributory infringer does not obtain immunity by using encryption to shield itself from actual knowledge of the unlawful purposes for which the service is being used.” Accordingly, [the founder] could not use [encryption] “to prevent himself from learning what surely he strongly suspects to be the case: that the users of his service — maybe all the users of his service — are copyright infringers.””
Code Wars: 10 Years of P2P Software Litigation by Rebecca Giblin

By comparison, Bram Cohen claimed that BitTorrent‘s design was the exact opposite and made it easy for rights holders to sue infringers.

Using BitTorrent for illegal trading, [Cohen] added, is ‘patently stupid because it’s not anonymous, and it can’t be made anonymous because it’s fundamentally antithetical to the architecture.’”
File Sharing’s New Face (2004)

Separate from willful blindness, the law could have gone down a different path where file sharing companies were judged by how much of its usage was noninfringing even if the company couldn’t control how it was used. Unlike Kazaa, eDonkey, and Limewire, BitTorrent had many legitimate use cases it could point to in court

“the technology’s current legitimate uses are varied and growing, and include the distribution of the enormously popular World of Warcraft computer game, independent films, Linux operating systems, and data published by NASA
Code Wars: 10 Years of P2P Software Litigation by Rebecca Giblin

Of course, these use cases aren’t necessarily legal defenses, but it is still noteworthy given that most other file sharing companies did not have similar non-infringing use cases to point to. Someone working at the BitTorrent company probably did a lot of work to make some of these use cases happen.

If you visited the “introduction” section on BitTorrent’s website in 2005, you would find an odd essay on free speech and how it relates to the distribution of large files.

Even if BitTorrent is “free speech”, why is that part of the introduction to BitTorrent?

When BitTorrent showed up in 2001, they didn’t magically predict that the law would eventually look at the company’s conduct in order to determine its fate. Instead, it looks like the BitTorrent company tried to imagine as many legal outcomes as they possibly could and then positioned themselves from the start to have the best legal defense possible. Looking back in time, I can spot at least five clear differences between BitTorrent and the other front runners that were plausibly also legal defenses:

  • Unlike other protocols, BitTorrent’s technical design meant they had nothing to do with search
  • Unlike every other company, didn’t even mention “file sharing”, “p2p”, or “music”
  • Unlike Aimster, BitTorrent did not obscure the IP addresses or content in the protocol
  • Unlike other file sharing companies, BitTorrent had many strong non-infringing use cases
  • If all of those defenses weren’t enough, BitTorrent positioned itself early on to argue that its technology is an act of free speech

BitTorrent was legitimately harder to use than many of the competitors in 2004. You could just open up Kazaa or eDonkey and search for the content you want. With BitTorrent, you had to download this client and then separately go fetch a .torrent file for it from some third party website. BitTorrent’s website basically stonewalled you with a mixture of missing information and misleading information.

If you weren’t watching closely, BitTorrent may have seemed incompetent and the fact that they won the p2p wars might have seemed lucky. If you look closer though, it seems obvious that BitTorrent had a half dozen legal strategies up its sleeve from the start and it really paid off. Of course, their technology was also legitimately innovative and we shouldn’t discount that. Still, Gnutella also had swarming downloads and eDonkey was using distributed hash tables before BitTorrent. Technology mattered, but legal strategy and targeted decentralization killed off the majority of the market in the long run.

If we put on our usual venture capitalist hats and looked at the file sharing market in 2001 or 2002, we would have seen Limewire, Kazaa, and eDonkey exploding with Snapchat like popularity. Their websites made it clear that they were tools for sharing music and movies, and their products included useful features like “theaters” which let users listen to music and watch videos while they downloaded. Executives at some file sharing companies even kept tabs on whether popular artists were available for their users.

“StreamCast’s executives monitored the number of songs by certain commercial artists available on their networks”
MGM v. Grokster

If companies building decentralized technology are flirting with the boundary of the law, maybe immediate breakneck user growth and explicitly giving users what they want is not the best heuristic. Maybe it is better to review the competitors in the space carefully and search for Straussian readings that may reveal which companies are strategically handicapped so that they can survive an eventual legal correction.

As I’ve already written, I think the decentralized technologies that manage to become mainstream successful will be the ones that flirt with the boundary of the law. In 2018, a few examples might be

  • Privacy coins
  • Prediction markets
  • Censorship-resistant file storage
  • Decentralized corporations
  • Stable currencies acting as proxies for government fiat currencies

I’m not talking about any specific company, coin, or token. I’m talking about the spaces. For each of these, we might benefit by asking ourselves “Who is Kazaa? Who is BitTorrent?” If we haven’t seen any sort of legal attention yet, we should be cautious in how we interpret growth if we’re investing in the long run. Weird legal structures and strange public relations decisions might actually be worth analyzing deeply in search of hidden meaning. Spotting the ugly duckling may be the name of the game when investing in decentralized technology.

If you liked this post, follow me on twitter. I have several other posts about p2p and blockchain and I’ll be writing more.