How to Efficiently Distribute Network Tokens

“for an entrepreneur building a network, her first 100 users IS her product, every bit as much as the tech itself”

I’ve spent the past few months chatting with hardcore buidlers, hunkering down to code during the pricing frenzy these past few months. A focal point of many of these conversations has been token distribution, the crypto-native version of market making and, more importantly, community building. During the ICO mania, this concept was often overlooked, partially due to ideals of equal access for all investors but also partially due to greed. In retrospect this has been detrimental to some because for an entrepreneur building a network, her first 100 users IS her product, every bit as much as the tech itself. Even today, we oversimplify the process by using the term “airdrop”, as if tokens are literally meant to fall from the sky. However more and more so, I see quality projects spending time on hiring for the right Head of Community role, which has more of the responsibilities of a traditional Head of Growth role in a centralized company.

If we look to the most successful companies who have built digital networks in the past, they were able to bootstrap user-generated virality starting with success with an early adopter group who then is able to spread the excitement to the mass user. The challenge is that the early adopter has very different needs than the mass user; thus, the sale to the former is very different than to the latter — even more complex when you need the former to sell to the latter.

While tokens are a powerful tool for bootstrapping growth, as I’ve written about before, it is not enough to assume that financial incentives alone are enough to get virality in the early days of a network. There is a science to the supply schedule, pricing, etc., but there is also the art of who holds it, what did they have to do to get it, how ownership is perceived that is the determinant of success for a network ultimately.

Lessons from Web 2.0

Shortly after the private beta of Spiegel and 20 friends in 2011, Snapchat became really popular at a local Orange County high school, which then spread to other local high schools as a “digital way to pass notes during class,” from one popular high school kid to another. The user experience was much more intimate than existing social media platforms and matched well with the communication style of its early adopter base. Another key feature responsible for the app’s success was that someone probably had to show you how to use it, as the average user’s first experience with Snapchat is actually quite confusing. However, this personal tutorial strengthened the app’s differentiation as an intimate social network, an “in-group” you literally had to be invited into.

If we examine the early days of Facebook, we can find a similar pattern. Having been seeded at the top Ivy League school and then subsequently other Ivy leagues and then only .edu emails greatly helped increase the “in-group” appeal of The Facebook. Back then, one of the reasons high schoolers were excited to start college was so that they could finally join Facebook. The key user insight at the time was that for the first time, the web was based on a real-world identity.

At the highest level we find some similarities between the adoption of these two networks they:

1. Satisfied early adopter needs of being a part of an “in-group” as well as a functional benefit that was superior to the existing alternatives
2. Attracted an early adopter community that the mass user wanted to emulate, enough to elicit an emotional response to join
3. Delivered a functional benefit to the mass user that was far superior to the existing technologies, enough for them to overcome the friction and make the decision pragmatically to participate in the network.

What’s Different about Web 3.0

Successful adoption of a crypto network should be very similar in that it should offer the same user benefits as above as well as add a financial incentive to the mix. In addition, there is an economic equation to be satisfied whereby you try to maximize the net present value of each incremental token holder you invite to the network. This is because since earlier adopters will reap disproportionately greater financial rewards (e.g. network value) from the token, they should be also able to contribute the greatest value, otherwise another would be better suited to take her place. This is based on the rationale that for economically sound crypto-networks, tokens are a scarce resource.

Following that train of thought, I’ve collected some principals for token distribution below in no specific order, and I hope projects will find them useful:

Small and Relevant > Large and Diluted

At the early stages, target a small, loyal community of relevant, value-adding members rather than a wide distribution of “randomly-selected” wallets. While there is some ideological appeal to “random”, statisticians can tell you that samples are often biased based on method of distribution. If you announced on Twitter, you’re likely targeting a bunch of VCs; if you announced on Slack or Reddit, you might get a bunch of speculators. If you’re looking to build a network of expert content creators, perhaps they’re not available in the channels everyone’s using because they’re still not super familiar with crypto yet. It’s okay if you’re specifically looking to get the early crypto adopters; however, I’d challenge the assumption that this is, by default, the best early adopter group for every crypto project, since it is such a small and isolated group. It is much harder for a project to cross over into mass user territory based on virality in this group alone since their needs are so different and they may not influence the mass user base. Also, this group might already be invested in several other crypto projects causing their attention and incentives to be spread thin. I urge projects to target value-adding contributors who may be new to the space and to get their attention by abstracting away a lot of the complexity with crypto.

Leverage Existing Reputation Systems as Signal

it is much easier to tap into and grow an existing community/network with established social incentives and reputation signals than to bootstrap one from scratch using only financial incentives.

Projects often struggle with finding the right objective signals for who will be a more valuable community member as it’s shaped significantly by the needs of the network, which are constantly evolving. An interesting model is to use a member’s past performance without a financial incentives, whether that be on Github, Slack or an earlier version of the network. Projects such as Numerai and Props targeted the highest value-adding members from the prior, non-crypto versions of the networks. Although past value isn’t a perfect approximation, it does give some helpful indication. If you are starting from scratch, an opt-in program for your competitor’s top contributors might also be a great place to start, to the extent that their reputation systems are publicly verifiable. Bottom line is that it is much easier to tap into and grow an existing community/network with established social incentives and reputation signals than to bootstrap one from scratch using only financial incentives.

Remember the Non-Financial Incentives

With a targeted group of loyal advocates, it is also much easier to establish trust and reputation/identity which acts as the glue for the network before a lot of the rules are fully fleshed out in the protocol.

While the financial reward is baked into an (hopefully) appreciating token, valuable early adopters are most likely motivated by rewards that are not wholly financial, such as bragging rights, ability to drive revolutionary change, social validation, etc. Early community building efforts should assess and address all of these needs because early adopters are not only the network’s first customers but also its first developers, salespeople, customer support agents, etc., even more so in a decentralized organization than a traditional one.

Specifically, the early adopters group’s affinity could be a value proposition in itself. A user motivated by social incentives is also more likely to demonstrate constructive behavior in growing the network irrespective of financial rewards or utility value from the early network itself. With a targeted group of loyal advocates, it is also much easier to establish trust and reputation/identity which acts as the glue for the network before a lot of the rules are fully fleshed out in the protocol. This gives projects more time to iterate in a safe environment with less governance issues

I’ll go off slightly on a tangent here and argue that in general in building the network, incentivizing trust is key. Much of the goal in crypto is to build trustless systems, which provide the infrastructure for human social scalability. However, when under development, these systems don’t work perfectly and need to be iterated upon, during these times, trust acts as glue that fills the gaps and propels adoption forward. The dissolution of trust too early on can actually significantly impede the network’s ability to evolve, which can be good or bad depending on the use case.

So much of our behavior is governed by social or intrinsic motivators, and it’s important to understand that when we introduce a financial motivator, sometimes the other reward systems are totally disrupted and can be hard to get back. Trade-off between the two systems should be carefully planned when introducing a token. I am becoming more and more fond of networks that have both a financial incentive (token) and a separate reputation system which cannot be bought or sold.

Not “Free” Tokens, “Your” Tokens — Use Endowment Effects

A lot of research has been done by behavioral economists on the psychology of “free” stuff. Giving things away for free is good for generating trial behavior because the notion of “freeness” is exciting beyond the marginal value delivered by that good. However, it has the potential to make people think they should get it for free forever if they didn’t do anything to earn it. Endowment effects, on the other hand, dictate that people would pay more to retain something they already own than to buy that thing in the first place. It also says that the more work you do for something, the more you feel ownership towards it, thus the more you value it. “Work” in this case can literally mean proof-of-work mining, or tweeting 100 times about how awesome this project is, or even the mental turmoil you must withstand during a huge correction in the market (a hodler’s torment if you will). Building endowment effects in the early community is a key part of building value into a token.

Smart distribution mechanisms I’ve observed usually reward a select group of users with tokens for some work that they do that is directly beneficial for the network over some period of time, simulating proof-of-work mining in many ways. The challenge is defining what is the most effective type(s) of work for growing the network as simply as possible, whether that be downloading an app, broadcasting a video, staking your tokens or writing some code. Also important to keep in mind that any work incentive you create can and will be gamed, but the best ones should further improve the network even when they’re gamed.

Lastly, what I haven’t seen but would love to see tested is UI hacks to increasing the feeling of ownership over tokens. A message on the dashboard like “Wendy’s 1000 widget tokens, which took 5 years, 4 months, and 2 days and 55 video broadcasts to earn” or perhaps more creepily, all of my tokens having my face on them might make me feel like they are more mine. I hypothesize that this might be one of the reasons Cryptokitties became so popular; the sheer fact you can customize and think of clever cat-pun names for your kitties makes you want a higher price for your “Naval Navicat.”

Plutocracy is Bad, Let’s Not Build That Future

Small players, in addition to the work they do for the network, contribute decentralization as a positive externality to the network, but often aren’t compensated for it

Plutocracy is bad, for many reasons I won’t get deeper into here. Any incentive mechanism that explicitly rewards economies of scale and rewards contributors proportionately should be counter-balanced by another which disproportionately rewards small players more. As discussed in the design of Casper, this might require an identity system to be implemented, but I won’t digress into that topic here.

Small players, in addition to the work they do for the network, contributes decentralization as a positive externality to the network, but often aren’t compensated for it, so they are disincentivized to join. This is why most reward mechanisms in crypto drive towards centralization eventually.

It is especially worsened by token distribution mechanisms that give out tokens proportionate to stake measured as an absolute level. If the intention is to measure how much skin in the game a community member has and money is the only quantitative measure of this, it should be as a percentage relative to her wealth rather than absolutely. $500K in stake costs way less for a billionaire than the equivalent to an average middle-class person. The latter is far more likely to be a good actor for the network with that amount at stake than the former. That’s the hard bit about incentives; they are not absolute; they are all relative, and financial stake is imperfect and only one data point.

What’s the Cost-Benefit Analysis on Tokenizing?

When it comes to adoption, the trade-off is of having a token vs. not is between bootstrapping financial incentivized network effects vs. the friction of managing a token + the cost of holding the token due to potential price volatility. For some projects, there is some natural breaking point where it no longer makes sense to have a token, where equity might be the better model for accruing value.

It’s All About the People

Overall, I believe we are still at the very beginning of crypto adoption, and the decisions projects make now will significantly impact the pace and extent of mass adoption. I have this fundamental belief that successful technologies should seek to both tap into existing human behavior and change it dramatically on some dimension, but you can’t have the latter without the former, thus overall I would love to see more projects to build for existing communities with existing, potentially socially less-scalable governance mechanisms and to grow them using technology. The true power of crypto-networks is its ability to extend the notion of community within our society as a whole.

I would love to hear your thoughts, critiques, additions to my points above.