We think a great deal about incentives, incentive design, and incentive markets. We invite you to share some of our mindshare — and join our incentive hacking reading list.

1 — Blockchain Background

1.1 The authoritative Guide to Blockchain Development

The authoritative guide to blockchain development

Cryptocurrencies, ICOs, magic internet money — it’s all so damn exciting, and you, the eager developer, want to get in…

medium.freecodecamp.org

1.2 Bitcoin Coursera Course by Princeton

https://www.coursera.org/learn/cryptocurrency

1.3 Stanford Cryptocurrencies and Blockchain Technologies

https://crypto.stanford.edu/cs251/

2 — Whitepapers

2.1 Bitcoin

https://bitcoin.org/bitcoin.pdf

2.2 Ethereum

https://github.com/ethereum/wiki/wiki/White-Paper

2.3 Stellar

https://www.stellar.org/papers/stellar-consensus-protocol.pdf

2.4 Lightning

https://lightning.network/lightning-network-paper.pdf

2.5 Monero

https://whitepaperdatabase.com/monero-xmr-whitepaper/

2.6 zCash

http://zerocash-project.org/media/pdf/zerocash-extended-20140518.pdf

2.7 Casper Whitepaper

https://github.com/ethereum/research/blob/master/papers/casper-basics/casper_basics.pdf

2.8 DFinity whitepaper

https://dfinity.org/tech

3 — Incentive Research

3.1 On Bitcoin and Red Balloons

https://arxiv.org/pdf/1111.2626.pdf

3.2 Reflecting on the DARPA Red Balloon Challenge

https://cacm.acm.org/magazines/2011/4/106587-reflecting-on-the-darpa-red-balloon-challenge/fulltext

3.3 Query Incentive Networks

https://www.cs.cornell.edu/home/kleinber/focs05-qin.pdf

3.4 Do Diffusion Protocols Govern Cascade Growth

https://arxiv.org/abs/1805.07368

3.5 The Strength of Weak Ties

https://sociology.stanford.edu/sites/default/files/publications/the_strength_of_weak_ties_and_exch_w-gans.pdf

3.6 Partisan Bias in Factual Beliefs about Politics

https://www.nowpublishers.com/article/Details/QJPS-14074

4 — Applications

4.1 Time Critical Social Mobilization: The DARPA Network Challenge Winning Strategy

https://arxiv.org/pdf/1008.3172.pdf

4.2 Flash Organizations: Crowdsourcing Complex Work By Structuring Crowds as Organizations

http://hci.stanford.edu/publications/2017/flashorgs/flash-orgs-chi-2017.pdf

4.3 Crowdsourcing Contest Dilemma

http://rsif.royalsocietypublishing.org/content/11/99/20140532.short

5 — Cryptoeconomics

5.1 Selfish Mining Paper:

https://www.cs.cornell.edu/~ie53/publications/btcProcFC.pdf

5.2 Asset Flow and Momentum: Deterministic and Stochastic Equations

http://www2.pitt.edu/~caginalp/Paper70.pdf

5.3 Stablecoins: Designing a Price-Stable Cryptocurrency

https://haseebq.com/stablecoins-designing-a-price-stable-cryptocurrency/

6 — Potpourri

6.1 “The Rational Optimist” by Matt Ridley

https://www.amazon.com/Rational-Optimist-Prosperity-Evolves-P-s-ebook/dp/B003QP4BJM

6.2 “Striking Thoughts” by Bruce Lee

https://www.amazon.com/Bruce-Lee-Striking-Thoughts-Library-ebook/dp/B014C57RYK/ref=sr_1_1?s=digital-text&ie=UTF8&qid=1530658575&sr=1-1&keywords=striking+thoughts

6.3 “Business Adventures: Twelve Classic Tales from the World of Wall Street Paperback” by John Brook

Business Adventures: Twelve Classic Tales from the World of Wall Street

"Business Adventures remains the best business book I've ever read." -Bill Gates, The Wall Street Journal What do the…

www.amazon.com

6.4 “Reminiscences of a Stock Operator” by Edwin Lefèvre

Reminiscences of a Stock Operator (A Marketplace Book)

"Although Reminiscences...was first published some seventy years ago, its take on crowd psychology and market timing is…

www.amazon.com

6.5 “The Mythical Man Month” by Frederick Brooks Jr.

https://www.amazon.com/Mythical-Man-Month-Software-Engineering-Anniversary/dp/0201835959/ref=sr_1_1?s=books&ie=UTF8&qid=1530659424&sr=1-1&keywords=Mythical+Man+Month

Can I see some ID please? (Burning bush. Painting from Hermitage Museum, Saint Petersburg)

Blockchain developers aren’t the first humans to face identity verification problems. In the book of Exodus, Moses finds a peculiar bush that is on fire but not engulfed in flames. When it’s revealed that the burning bush is God delivering a message to his people, Moses asks quite reasonably: how will they know it’s really you who sent me, when I tell them?* ***It turns out that not even God is above identity verification.

**“Tell the people of Israel: I AM has sent me to you.” **Exodus 3:14

When it comes to incentive programs, the problem is that **there’s an inherent tradeoff between the virality of the program and its resistance to identity fraud. **Anyone can create an algorithm that distributes a novel crypto-coin to every person in the world. Without any additional layers of security though, such network would easily fall prey to a *Sybil attack. *Put another way, fake identities would be created for the attacker to gain more rewards than they are fairly owed. Designing a system that is both viral and resistant to such attacks is a holy grail.

Do you remember the DARPA red balloon challenge? They sponsored a competition to find red balloons, and a team from MIT solved it in just 8 hours using recursive incentives. Anyone who found a balloon for MIT would win $2000, and anyone who referred a balloon finder for MIT would win $1000. Even if you referred someone who referred someone who referred someone…you’d win *some *amount of money in return. The truth is that MIT was lucky that it solved the challeng in only a few hours, because Sybil attacks were most certainly on the way.

As Babioff, Zohar et al clearly pointed out, the system had the unintended consequence of incentivizing the creation of false identities to extract undue rewards. For example, if Alice knew where a red balloon was located, she could create two false identities, “Bob” and “Carol,” and claim to have been referred by them. She could add to her reward without doing any additional work.

Many of the “needle in a haystack” problems we’re tackling are best solved by an ad hoc viral network. Finding the right employee, for example, is easiest when you have a large group of people (i) applying and (ii) spreading the word about the job opening. **As networks grow, though, it becomes harder to remain resistant to a Sybil attack. **Therefore, one of the fundamental goals is to discover the point at which the virality of a network is maximized relative to acceptable levels of identity fraud. In reality the answer is really a set of answers depending on the properties of the particular incentive program being discussed.

**A common challenge of an incentive program is to maximize network membership while minimizing the opportunity for fraud. **This property is one of the most under-appreciated aspects of Bitcoin — and what makes it such a powerful system.

When a Sybil attack occurs, a large part of the network doesn’t contribute new information. It is instead dominated by one actor who creates numerous nodes simply to ensure that they reap the rewards if someone finds the solution. **The very incentives that can lead a network to rapidly grow can therefore also lead to its demise. **Additionally, the attacker in the red balloon case would be able to receive up to twice the reward they deserve (instead of just $2000, they can collect almost $4000 in added chunks of $1000+$500+$250+$125.5 etc). This can dispirit the efforts of the honest participants, and cause the network to fail. So you see, identity verification ends up being the limit of virality.

There are a number of interesting ways to secure native blockchain protocols for identity and other related problems such as collusion and fraud detection.

ESL Intel Extreme Masters 2014, San Jose (4 Million Online eSports Viewers)

A tribal network has the potential to become a viral network, especially in certain online contexts. That is why the world of blockchain has assigned an aspirational meaning to the word ‘tribalism.’ If you can create or tap into a “tribal” network, then its users (the “tribe”) will keep it alive and growing. If we knew more precisely what tribalism was and how we could re-create it, then we could motivate large groups of people to essentially act as one. Combined with the right incentive market, especially one implemented on a blockchain, the effects of tribalism in our economy could be nothing short of revolutionary. We therefore need to inquire into the tribalism itself. What is tribalism, and how do we create it?

1. Tribes are formed primarily through shared experiences.

Research like this shows that people experience amplified emotions when they go through an experience together. Even something as simple as eating chocolate, for example, feels more pleasant when another person is eating it along with you. The same goes for attending the same sports event, musical performance, or political speech together with other people - happy events feel happier, sad events feel sadder, and irritating events feel more irritating when they are shared with others.

The amplified emotions that shared experiences create have important downstream effects. More extreme emotions mean that shared experiences are remembered better than solitary experiences. Since memories form the foundations of trust, shared experiences literally hardwire us to feel more trust towards the members of our tribe. What’s more, when people experience strong emotions, they generally assign them great significance. As a result, people see shared experiences and the strong emotions they bring as powerful sources of meaning.

For example, consider the powerful shared experience that Tough Mudder provides. Friends and acquaintances endure an arduous obstacle course together as they co-experience strong emotions and physical fatigue. This is a source of strong bonding that strengthens the community. If someone did a challenge like this on their own, without a group of friends, the experience would be much less powerful.

Shared experiences, even those conducted online, cultivate a strong sense of community.

2. Social interaction reigns supreme.

The opportunity to interact with another person is also a precursor for tribalism. Broadly, social connection is a powerful determinant of happiness and well-being. In fact, loneliness (the opposite of social connection) is as big a risk factor for mortality as excessive drinking and daily smoking. For most people, even short interactions can create the feeling of social connection. Networks that can create more social connection between members, whether digitally or in-person, are more likely to create tribalism.

For example, consider the opportunity to sit in an arena and watch popular Esports competitions alongside other fans. Not only do fans get to share the experience of seeing their idols play their favorite game, but they also get to interact and share war stories with like-minded fans. These in-person conversations provide a key bonding function to the gaming network, which has proven to be one of the most tribal and viral networks.

3. Helping one another generates strong bonds.

Helping others is another critical factor for well-being, and one that is especially suitable for creating tribalism in blockchain-driven networks. Helping others makes people happy, makes their lives more meaningful, and makes people feel that they are more effective. Moreover, it creates norms of cooperation and increases others’ desire to reciprocate. As circles of reciprocity grow, the network grows. Thus, networks that allow people to help each other are more likely to create tribalism, which is partly characterized by feelings of obligation to help one another.

For example, the streaming platform Twitch.tv provides ample opportunities for helping. Specifically, streamers can ask fans to donate to charities and fans can donate to the streamers’ channels. Helping generates positive feelings and a sense of community.

Twitch.tv has tapped into the power of charity to cultivate the Esports “tribe”

Blockchain technology can help charities collect donations and might even become a key ingredient in the fight against cancer.

4. A common outgroup or nemesis provides a rallying call.

Any analysis of tribalism cannot ignore people’s tendency to unite when a common adversary is present. Research like this shows that a common enemy creates significant amounts of ingroup unity. Examples of this component of tribalism can be found in the blockchain community itself. Some have claimed, for example, that Ethereum’s Vitalik coined the term ‘Bitcoin maximalist’ to create a tribal identity pitted against an arch nemesis —

But some advocates of Bitcoin graciously accepted the title nevertheless and reclaimed it as a badge of honor. These rivalries are a source of motivation and cohesion. Having a common adversary can be a potent precursor to tribalism and network growth.

Tribalism is a helpful tool for achieving the potential of blockchain-driven networks to solve problems and attain goals. But without knowing its causes (and potential dangers), we can’t use it effectively. In a previous post, for example, we explained how tribalism can be utilized in a way that does not incite violence. It will be important to keep these insights in mind as we move forward and continue to explore the potential uses of tribalism in online (blockchain) communities.

Go to Part 3

**Going viral is the holy grail for most network designers. **Networks that quickly expand and connect scores of members are more likely to achieve their goals. The Ice Bucket Challenge, for example, got 17 million people to either donate or soak themselves in cold water in front of a video camera, raising more than $115 million for ALS.

Many doubt that there is a recipe for virality, probably because it can sometimes seem like what goes viral is totally arbitrary. There are, however, a number of network properties — like tribalism, Sybil-resistance, and weak ties — that network designers must get right for a network to go viral.

1. Be consistent.

Familiarity is one of the strongest causes of liking and attraction. At the most basic level, simply being exposed to something increases our preference for it. Academic research has shown, for example, that the more people see certain words, drawings, polygons, and people, the more they like them. One hypothesis is that people like what they are familiar with because it is easier to process what they already know. We feel more confident interacting with what is familiar, that is, because there is less risk of being blindsided by a negative experience.

“People like consistency. Whether it’s a store or a restaurant, they want to come in and see what you are famous for.” — Millard Drexler

For networks, this means that consistency is key. Consistent networks tend to provide similar experiences over time, so people can digest novel content through familiar channels. Buzzfeed, for example, uses the same formula for every video they post online. This enables viewers to formulate general expectations even though each video features different content.

2. Find the optimal cost to participate.

Conventional wisdom in business says that customers gravitate towards what is easy. A simple of click of a mouse is easier than filling out a webform, which is simpler still than proving your qualifications to join a network. There is good reason for this mantra: The easier it is to use the network, the higher the likelihood that people will try it out.

There is, however, a psychological downside to low cost for participation. The lower the cost of entry, the lower the psychological commitment to the network. For example, to participate in the Ice Bucket Challenge, people had to share a video of themselves pouring a bucket of ice cold water over their heads— a very high cost for entry. Nevertheless, the ice bucket challenge went viral. How?

First, there was a huge social element that fueled the virality of the Ice Bucket Challenge. People were nominated by their friends, who effectively called them out in a public forum, like a Facebook post. This created social pressure to contribute to a good cause, either monetarily or by spreading the word.

Perhaps more importantly, however, tasks that are marketed as “impossible” or “too hard” pique the interests of people who want to prove otherwise. This little bit of reverse psychology, which plays off the absurdity of pouring ice cold water on yourself on video, can motivate the “underdogs” among us to stand out. Take, for example, the so-called Impossible Game, which pits players against an unpredictable and ad hoc series of challenges. High costs of participation, paradoxically, can make something seem much more enticing than it may actually be.

Why would anyone want to play a game designed to be as frustrating as possible?

**There is thus an optimal point for how costly it should be to join a network, and it likely depends on the type of problem the network was created to solve. **Some networks should be very easy to join because long-term commitment does not matter for achieving the network’s goal. An example might be networks that try to quickly identify specialized solutions within a large array of possibilities, like “needle in a haystack” problems. In contrast, some networks should be difficult to join because exclusivity and longevity matter. One example might be professional networks for alumni of an elite college or a distinguished company. In general, getting the cost of participation right is one of the keys for going viral.

3. Use the right incentives.

The potential for virality also depends on the shape of the network graph, which is a depiction of how all of the nodes in the network are connected. Different network graphs require different kinds of incentives to go viral.

Some networks depend heavily on a few central members who are connected to everyone else, while most other nodes are not connected to one another:

Such networks might be a celebrity’s Twitter following or a political campaign centered around the candidate.

Other networks are more evenly distributed, such that most nodes have the same number of connections as most other nodes:

This kind of graph might represent a fair trading network between communities who share resources with one another.

Still some other graphs are hybrids of the two we have previously seen, with multiple clusters connected through weak ties:

This kind of graph approximately represents a normal social media network, with groups of friends connected by some mutuals.

The optimal incentives for a given network will differ depending on the shape of the network. More evenly distributed and decentralized networks, for example, will likely benefit from relatively small incentives for which everyone in the network is eligible. In more centralized networks, however, incentives must be targeted at the central nodes. Decisions about targeting incentives should follow these analyses.

There is a lot of talk about virality in the blockchain world. In general, understanding the components of virality can help us engineer virality instead of chalking things up to luck and circumstance. We think that one of the keys to virality is finding the right incentives, which can only be done once you’ve studied the abstract form of a network’s graph. There is good science being done on this front, and we shouldn’t let it go to waste.

Go to Part 2

**Designing the right incentive system is crucial for optimal system operation, and that this statement applies to domains far and wide. **Today we’ll focus on a system that countless commentators have called a “broken system” — namely American politics.

Saying that the U.S. is divided across political lines is by now a cliché. The inauguration of each new president in recent memory brings forth media hyperbole about the chasm between Democrats and Republicans. And no, Trump is not unprecedented in this regard. In an article published way back in 2003, Yale psychologist Geoffrey Cohen asked people to evaluate two proposals related to welfare policy. One was very lax — more lax, in fact, than any actual policies present at the time. The other was very stringent, more so than any actual policy at the time. Support for these policies had nothing to do with their content. Instead, people supported and opposed these policies solely based on whether they originated by their favored or by the opposing party. Democrats supported the stringent welfare policy if they read that it was proposed by Democratic lawmakers, and Republicans supported the generous welfare policy if they read that it was proposed by Republican lawmakers. These striking findings, along with the current media bellowing about partisan conflict, seem grim — is the never-ending partisan conflict hopeless?

Support for these policies had nothing to do with their content.

We wonder about the incentive system that underlies political attitudes, and a 2015 article by political scientists John Bullock, Alan Gerber, Seth Hill, and Greg Huber caught our eye. Political divisions can have two sources. The first is a legitimate difference of opinion about the best policies of our country. The second, however, is an irrational difference of perceptions of objective facts. Is the economy doing well or poorly? Has unemployment increased or decreased? Is inflation higher or lower now than a decade ago?

These questions have objectively true answers, and yet Democrats and Republicans disagree on their answers as a function of who was president at the time.** **For example, despite a marked decline in inflation in the 1980s during the Reagan administration, Democrats reported that they believed inflation had increased in that time. The same contradiction was true of Republications in their reports of the Clinton presidency.

Democrat and Republicans disagree on their answers as a function of who was president at the time.

What explains this? One possibility is that partisans indeed have distorted perceptions of reality. Another reason, as suggested by Bullock et al., is that **partisans are motivated to act as cheerleaders for their side, and enjoy expressing their side’s superiority and the other side’s inferiority. **This means that partisans may know that they are ignorant about certain issues, but would rather express statements that boost their own side and take down the other side. For example, Democrats may not know whether inflation increased or decreased in the 1980s, but they prefer saying that it increased because they are motivated to cheerlead for their side (and the same logic applies to Republicans).

If this theory is true, then simple incentives can reduce the apparent partisan conflict, because cheerleading for your own side is less appealing when you have to give up money to do it. Bullock tested this in an experiment done in 2008. The experiment asked Democrats and Republicans to answer a number of factual questions such as whether the level of unemployment increased or decreased compared to 2001, when President Bush took office. Without incentives for answering correctly, Democrats were much more likely to state that unemployment has increased than Republicans. The same pattern emerged for factual questions about the Iraq War, the budget deficit, and so on. These responses exhibit a pattern of partisan thinking, with Democrats seemingly believing that the world under Bush was dark and stormy, and Republications believing that same world was rosy and hopeful.

But the introduction of an incentive changed this picture. Some people in that experiment were randomly to receive a chance to win a lottery of $200 for every correct answer. **Incentivized participants reduced the partisan gap by 55%. **Moreover, in another experiment in this article, people were both incentivized for accuracy and offered the possibility of stating that they do not know the answer to a question. Some 48% of responses in that situation were, “I don’t know,” suggesting that people have some self-awareness of their ignorance about policy facts. When there are incentives to being accurate, people do not replace their own ignorance with partisan perceptions. In conclusion, external incentives can reduce the appeal of partisan cheerleading.

…a correctly designed incentive system can align interests among parties in conflict

**This is a good example of the potential of incentive markets — a correctly designed incentive system can align interests among parties in conflict. **People wouldn’t mind going to war with their political opponents so long as the stakes are low to non-existent. And to pundits and observers, the partisan divide seems unsolvable. To us, it’s all a matter of finding the right incentive system.

How do you find the job candidate who’s an exact fit for your organization? How do find the person you are going to marry? How do you locate the right house for you and your family? How about a pet that’s been missing? Worse yet, a child who’s been missing?

As a society,** **we spend a lot of time looking for things. Individuals search for apartments to rent, cars to buy, and romantic partners to date. Businesses search for customers, employees, and suppliers. Although we tend to approach these searches as if they are unique, we believe that they appear distinct on the surface but are often subject to similar dynamics on a meta level. Our goal is to enable people to use incentive programs and blockchain technology in order to solve some important problems facing the world today. We believe, in particular, that we can help people find higher quality things in a fraction of the time and more efficiently than they might expect.

Some search problems, like finding the closest Italian restaurant, are easy to solve. **But other search problems are more difficult and involve higher stakes. **Finding a missing child, for example, is more challenging because it may be unclear where to start and how to efficiently gather information. It’s obviously a high stakes situation, because it involves the welfare of a child — quite a bit different then satisfying a pasta hankering.

Difficult searches are called “needle in haystack” problems. One of our guiding principles is our belief that these “haystack” problems can be optimally solved by combining decentralized networks, blockchain technology, and the right incentives programs.

A ‘proof of concept’ for this approach happened in 2009 when DARPA (USA defense research agency) placed ten red balloons in secret locations around the U.S. and offered $40,000 to the first team to locate them all. The winning team solved this search problem by offering recursive incentives: Anyone who either correctly identified the location of a red balloon or referred someone whose information led to a person who identified the location of a red balloon was rewarded. For example, if Person A referred Person B who referred Person C who correctly identified the location of a red balloon, then all three people in this chain would get rewarded in some capacity.

Even if you don’t know where the needle is in the haystack, you are still incentivized to help by referring others to get closer. If the tail end of your subtree finds the needle, you too will be rewarded.

**Utilizing blockchain technology in this process enables the network to be transparent and to be audited by anyone at any time. **Everyone who contributes to the network can be absolutely sure that if their contribution is valuable, their reward will be commensurate with this value. This means that the person who finds the missing child *and *the person who spread the word to the finder’s social media network will be rewarded. The person who refers a successful candidate for a specialized job and the person who informed the referrer will be rewarded. These programs can reduce their overall cost with recursive payout structures* *(Larger payouts for those who are “closer” in the chain to the final goal). These recursive programs are scalable and when designed correctly, have an upper bound on the cost of sponsoring them.**

The biggest hurdle to solving a haystack problem is growing the search network quickly enough so that the chance of identifying the needle increases exponentially. Recursive incentives motivate people to inform other people of the search problem. When executed properly, these networks grow quickly and organically, increasing the statistical probability of finding the target. In fact, we believe that recursive incentives, where every node of the successful chain is rewarded, become increasingly more valuable as the search problem becomes more difficult. Spam is another common problem that we will address in a forthcoming post: “Cashout: What Does it Mean to be Spammed with Money?”

As the tail end of the network is further removed from the origin of the network, the spread of information will be more efficient than any other search process implemented in the marketplace.

We fully recognize that each case will come with its own idiosyncrasies, ranging from economic constraints of the searcher to the psychological realities of how people respond to incentives. The incentive structures implemented on our blockchain will be modular enough to apply to many types of search problems. The inclusion of “decentralized networks” and “blockchains” might sound like tech jargon, but in reality it’s just the underlying tech that enables you to give “asking your friends for help” a global reach!

If you’re like most people under 40, you know someone who found their current love interest through a dating application like Tinder, OKCupid, or Hinge. Just a few years ago, though, this was the exception instead of the rule. How popular has online dating become? And how can we expect the world of swiping left and right to change in the near future? Is there an even more efficient way for these applications to bring us together?

It’s no secret that most people want a romantic partner when the time is right. (It’s worth mentioning, though, that you can use some dating applications, like Bumble, to find new friends, too!) In our increasingly busy and productive world, it’s also no surprise that we’ve found some shortcuts to help people romantically connect. Before the internet, there were dating coaches and speed dating events to help. Now, it seems, we trust the apps.

The advent of these applications showed us just how untapped the dating market was. Websites like eHarmony and Match.com were among the first that hopeful lovers entrusted with their romantic lives, and since then the results have been impressive. In 2017, 19% of brides met their spouses online. Newer applications like Tinder or Bumble are targeted at younger audiences, and more casual relationships have become increasingly popular as a result. Since Tinder launched in 2012, it has registered billions of matches and continues to climb the App Store charts (residing now as the 26th most popular app). Even Facebook is jumping on the bandwagon.

Dating apps can do a better job of tapping into a person’s online network

Most dating apps ignore the wealth of information stored in a person’s online network. The user gets thrown into a selection process based on photographs and short descriptions, which provides breadth and ease of contact but is also incredibly inefficient. Of the billions of swipes on Tinder each day, only a small percentage results in a match, and an even smaller percentage of those matches results in an actual date.

This is because** today’s dating applications completely disregard your actual network, which contains the people that know and love you the most.** Many of the problems of dating online could be solved, though, if online matches were more frequently formed through a user’s actual network. Matches would be more efficient, for example, because more individualized connections could be crafted based on the data of who’s in your network. There would be no more fear of someone pretending to be someone they’re really not. And the potential danger of meeting a total stranger would be completely removed. But it gets better than that — building a dating app that implements an incentive program that utilizes your network could prove even more advantageous.

Tapping into the potential of incentives on dating apps

Some dating apps like Tinder offer the option of referring a friend, but they don’t provide enough of an incentive for doing so. Incentivizing referrals more strongly could allow the formation of an incentive market that we don’t quite see in the online dating system today.

Suppose, for example, that a dating application offers advanced features like Mega-Likes, Re-dos, Extra Swipes, and preferential display to other users. These features are worth some number of in-app tokens, which can be purchased or acquired through in-app activity. Referring a friend to the application gives tokens to both parties, and the invited friend’s future referrals recursively rewards the referrer as well. This would create a viral referral system that could spread the usage of this application like wildfire.

“Someday we’ll find it. The rainbow connection. The lovers, the dreamers and me…” -Kermit the Frog

Suppose further that you get more “dateCent” tokens through the successful suggestion of a match between two people in your network. When you make a successful match, you’ll receive an amount of dateCent tokens. Recursive incentives are implemented to make the process even more viral. If Bob sends his profile to Alice and asks her to match him with someone in her network, for example, Alice can send Bob’s profile to Carol — who might not be using the app at that point — and ask her to download the app. After joining the application, Alice can either match with Bob *or *use her own network to send Bob’s profile to someone, and so on. Let’s say that through this process, Bob eventually gets a match! Each person in the chain leading back to Bob’s match is rewarded according to a recursive rule. The person who found Bob’s match could be rewarded 16 tokens, for instance, and the person who found *them *could be rewarded 8 tokens; the person who found *them *could be rewarded 4 tokens, and so on. The token could represent “karma” points to build matchmaker social capital — or even be an actual side hustle for super-connector matchmakers, like a gig economy dating service. This kind of dynamic of using your friends’ and family’s knowledge of you to find your match has been successfully employed for millennia in cultures with facilitated matchmaking and marriages. Network referrals was how dating used to work- exploiting weak ties to find your match.

Blockchains could provide the trust layer to make dating referrals work at scale

Some dating apps like Tinder offer the option of referring a friend, but they don’t provide enough of an incentive for doing so. Incentivizing referrals more strongly could allow the formation of an incentive market that we don’t quite see in the online dating system today.

By implementing recursive rewards for using a person’s network potential, dating applications could incentivize an increase in the number of their users, the frequency of their matches, and the effectiveness of those matches. Exchange would happen naturally, allowing users to benefit from a “web of trust” where parties do not need to be in a trust relationship with each other to reap rewards from referrals that originate from multiple parties. Pseudonymity with honesty is a property that blockchains handle natively and is a candidate solution to the “transparency” problem in online dating which acts as a trust inhibitor at scale for many. And with the ease and scalability of a protocol for recording referrals and attributing value to referrers, it’s a natural fit for dating applications to use with an advanced recursive rewards program.

— The authors would like to acknowledge the following people for providing feedback on this article: Seth Hill

How have auctions changed over the past 50 years? Aside from going online in the form of household names like eBay and Christie’s, the basic mechanics of today’s auctions have stayed more or less the same — a seller puts an item on offer and potential buyers place their bids. Whoever places the highest bid at the end of the day purchases the item for that amount. There are limitations, though, on the effectiveness of today’s auctions. Sellers, for instance, often list their items in an auction and can only hope that the most interested parties find them. Buyers, in turn, need to seek out the right auction for whatever it is they’re looking to buy. When these two incentives align, auctions are mutually beneficial for buyers and sellers alike. We think that tapping into the potential of under-appreciated incentives can revolutionize the world of online auctions.

An inefficient feature of auctions today is that** buyers seek out sellers**. When Alice tries to sell her collectable Beanie Baby on eBay, for example, whoever buys it will ultimately come from the pool of people who search for Beanie Babies and related collectables on eBay. But all of the Beanie Baby collectors who aren’t checking eBay at the right time fly under the radar. A more efficient auction would have the market itself find the ideal buyer for each seller. Suppose, for instance, that Alice lists her item on an application called “auctionCent.” She shares her auction page with her immediate network, distributing tokens to an initial group of participants. Bob is in Alice’s network and knows that his cousin Carol is an avid Beanie Baby collector, so Bob refers Carol to Alice’s auctionCent page. Carol buys the Beanie Baby at a discount due to the referral, Bob receives a reward for finding Carol, and Alice has her collectible sold.

Unlike the hit-or-miss method of traditional auctions, incentive markets search out the perfect buyer, and everyone involved in price discovery benefits from the collaboration.

The evaluation of listed items, which is the purpose of an auction, is accelerated. What’s more, the expiration date of the token provides an easy way to enforce a time limit on auctions. Exchange would happen organically, and users would benefit from a “network of trust,” cashing in on rewards from referrals that trace back to different people. The protocol deters scammers and Sybil attacks while retaining scalability and auditability. It’s a natural fit.

Auctions are extremely useful for the fair exchange of items of unknown value. The problem is that auctions are often not as effective as they could be because there isn’t any infrastructure that connects buyers and sellers independently of the buyers’ own efforts. Instead, one should provide a simple scheme to attract the right people to the auction room and return benefits to everyone who participates.

What’s one of the biggest challenges facing nonprofits, social activists, and political campaigns? The answer is simple: finding donations. Since charitable work addresses some of the most important social, political, and ecological crises facing our society, we need to ensure that these organizations can effectively raise the funds they need. But as things stand, the most common recruitment strategies used by charitable organizations are highly inefficient. Cold calls, distributing brochures, hanging posters, waiting for walk-ins, and standing besides a donation table, for example, are slow methods that rarely reach new audiences. More effective methods in practice seem to involve small token incentives, like offering T-shirts, bracelets, and other small items as rewards for donating- believe it or not, they actually work. Another tried and true method is getting public commitments and matching challenges (If you give X I will match). We think that incentives have the potential to play a much larger role in finding charities the funds they need to help make an impact.

Imagine an NGO called *Adorable Kitten Adoptions (AKA) whose stated purpose is to save abandoned kittens. *They need to quickly find donors, incentivize those donors to give as much as possible, and spread awareness for their cause. The solution is that AKA buys kittyCents and gives one kittyCent to each person in an initial group of participants. Anyone holding a kittyCent can donate $10 to receive an appreciation package for saving a kitten’s life. The appreciation package has a bio and images of the kitten, among other memorabilia. Another thing you can do with your kittyCent, though, is pledge $10 and transfer the kittyCent to someone else. If they (or anyone they transfer to) donates $10, then everyone in the transfer chain matches the $10 donation. Suppose Alice gets a kittyCent, for example. Then:

1. Alice can donate $10 and receive the appreciation token, or

2. Alice can pass the kittyCent to Bob and pledge to match the donation upon redemption. If Bob redeems the kittyCent, Bob will donate $10 and Alice will match the $10, bringing the total to $20, or

3. Bob can pass the kittyCent to Carol and pledge to match the final donation. If Carol redeems the KittyCent by donating $10, both Alice and Bob will match Carol, generating a total of $30.

4. The pledges are triggered after three weeks to guarantee that they are not lost.

5. At the end of fundraising, people responsible for the most funds raised by their direct or indirect recruits would be acknowledged.

In theory, chains of pledges can get arbitrarily long as people pass kittyCents, adding pledges as people pass kittyCents forward. This chain acts as seed money to current kittyCent holders, increasing the chance that they donate or pledge to donate. When someone decides to donate rather than pledging a match, it will trigger a domino effect which can magnify that small donation by many times.

In addition to enabling this novel pledging mechanism, a protocol designed to provide infrastructure to document pledges and track fundraising referrals could also reduce fundraising expenses, which in 2014 accounted for more than $19 billion for charities in the U.S. alone. Charitable giving fraud is another challenge that a trust-less protocol may have a role in alleviating, especially for less well known charities. The effects that charitable organizations have on the social, political, and environmental challenges we face today can’t be understated. Yet the funding these organizations receive to perform their much needed work is abysmal. The solution is to fully harness people’s incentives to spread awareness within their network.

Launch of the SpaceX Falcon Heavy, Feb. 6, 2018 (Joe Burbank/Orlando Sentinel via AP)

Space exploration plays a special role in our hearts and minds. Humanity is set upon a great timeline of exploration, and space primordially calls to us next. Our first steps on the moon, for example, symbolize the cumulative technological achievement of humanity. Ever since then, children dream of jettisoning into outer space to explore the unknown.

Our presence in space doesn’t just provide inspiration for dreams, though. Space exploration provides tangible benefits. Satellites track the weather, storms, and wildfires. They also help us plan agricultural and urban land use from a bird’s eye view. Photographs taken from space shuttles teach us about the planets in our solar system, and one day we may be able to travel extraordinary distances in record time.

Space innovation needs to be better, faster and cheaper still. We believe the missing link is building the incentive systems that motivate innovation in space exploration.

The space industry has a unique economic structure. Since designing, building, and launching satellites and rockets requires cutting-edge science and engineering, the industry requires a highly specialized workforce. The time horizon is long, as capital investments made in initial periods can be recouped only years in the future. Governments tend to be heavily involved in this industry as parts of it pertain to national defense and pride. Private companies, however, have entered the industry as U.S. government involvement recedes.

The private space industry received an important boost in 2004 following the Ansari X Prize competition, which awarded $10 million to the first non-government organization that successfully launched a reusable, manned spacecraft. Many take the competition’s success to demonstrate that private enterprises can indeed meaningfully contribute to space exploration technology. The current market structure, however, stunts innovation because it incentivizes private companies to keep their intellectual property secret in order to protect their competitive advantage. These intellectual silos are unhelpful. The trouble lies in finding a way to quickly identify and motivate people who have the right skills and knowledge to collaborate on projects that can advance space technology.

Catalyzing space innovation through incentive-powered crowdsourcing is revolutionary for several reasons:

(1) Fosters intellectual collaboration that is sorely lacking in the space sector; people who have the right skills and knowledge unite to accomplish a specific technological need.

(2) Leads to prototypes that are far less expensive than what the current market could provide; the cash prizes in the contest summed to a few thousand dollars, whereas private companies like Boeing charge several millions of dollars to design a rocket engine.

(3) It changes the incentives of players in the space industry in profound ways.

Public contests are exciting for scientists and engineers to take part in because they marry two dimensions that kindle the human spirit: (1) the challenge of working on cutting-edge technology and (2) the publicity that comes with successful designs. These powerful social incentives lead to high levels of motivation.

Public contests allow scientists and engineers to work for the common good and most people are very motivated to achieve something meaningful that can help others. Public contests tap into the desire to do something meaningful and good in the world, which is a profound source of motivation in human beings.

Space exploration is a public good, because the discoveries realized in space benefit many of us on Earth.

**Public contests allow scientists and engineers to connect with like-minded peers **and work together on meaningful endeavors, which is another basic human drive. These contests could form the basis for to create decentralized space markets that enable scientists and engineers to innovate incredibly efficiently. What’s more, they create ad hoc networks that exist to solve specific technological problems. Due to their goal driven nature, there will be no unnecessary bureaucracy or “legacy” business units that tend to create barriers rather than incentives for collaboration.

**Contests can be more effective with protocol-powered accurate tracking, attribution, and auditing of the contribution of each party — and to permit multiparty coordination without requiring anyone to trust each other. **Different teams that contribute sub-components of winning prototypes can be compensated based on their relative contribution to the technology. This would allow scientists to create multi-layered networks that can solve sub-components of technological problems and provide for more accurate allocation of rewards. There may be many more applications of blockchain technology in the exploration of space, too.

The barrier to making space travel and exploration a reality is not an intellectual one. There are more than enough capable scientists and engineers who can help the space industry take the leap forward. Instead, the barrier is one of finding the right incentive system to get all of these hearts and minds to collaborate on the right goals at the right time.

Markets are a more efficient way to organize work at scale than firms. (For an introduction to how incentive markets can be formulated on a blockchain, see our recent post entitled The Success of Incentive-Centered Markets.) There are three meta-categories of incentive programs: search, crowd, and *social *programs. Each type offers different advantages and disadvantages, but they all increase the efficiency and value of an existing network.

1. “Search” Incentive Programs

Search incentive programs outsource the work of finding a person, place, or thing to a network of individual searchers. In a search problem, the value or solution lies at the leaf or the bottom of the tree. The basic idea is to incentivize people to solve the search problem 1) directly or 2) indirectly by finding someone else to solve it directly. Since only the referral chain that succeeds in locating the relevant thing is compensated, this kind of incentive program is incredibly cost effective. Constructed properly, search incentive programs have also proven to be more efficient than one might initially think, as demonstrated by the quick resolution of the red balloon project. Matching problems such as dating or kidney donations or medical school placements are a type of two sided or dual search problem.

2. “Crowd” Incentive Programs

Crowd incentive programs combine crowdsourcing and incentives by involving as many nodes to contribute to the solution, which lies at the root of the tree. The focus here could be on crowdwork (labor), crowdfunding (capital) or other resources. The trick here lies in incentivizing collaboration. Building a house, for example, requires the coordination of architects, builders, plumbers, electricians, and so on. Getting these groups to collaborate and participate together on a project can sometimes be a logistical nightmare, but all of these professionals *can *be incentivized to seek one another out and collaborate. By reducing the inefficiency and cost of facilitating collaboration between professionals, “crowdwork” may become the path to deconstruct corporations into markets.

3. “Social” Incentive Programs

Social incentive programs extract value out of social networks by mapping the connections of their users and using this information to generate targeted incentives. The value here lies in the edges or connections in the graph. Facebook, for example, could certainly benefit from incentivizing its users to recruit new users and share high quality posts to attract new users. We expect to see novel uses of social incentive programs, for example, in dating applications, entertainment searches, and marketing.

Medical research has made tremendous strides in understanding the causes of cancer and mitigating its devastating effects. This research dates back to 1775, when it was first understood that exposure to chimney soot is a risk factor. Over the past 250 years, we have (i) developed living models of cancer in the lab, (ii) found drugs that kill cancer cells efficiently in a dish, and (iii) even cured mice of this deadly disease. However, if these treatments even make it to a clinical trial, they usually fail to achieve better outcomes than the unsatisfactory standard to which they are compared. So even as significant advances accumulate in the lab, they fall short in the clinic, and we seldom learn why.

“…even as significant advances accumulate in the lab, they fall short in the clinic, and we seldom learn why.”

Sharing data can quicken the pace of scientific discovery by focusing attention on promising signals generated by researchers with shared goals. In cancer research, there’s little debate about sides; we all want to eradicate this disease by preventing it from occurring and getting rid of it once it does. Greater availability of patient data from trials means that (i) researchers can check one another’s work and correct inaccuracies in reporting of medical trial results, (ii) researchers can test hypotheses using larger datasets, which allows them to make more reliable conclusions, and (iii) signals from small subsets of patients in trials can be expanded upon and followed up. These are the central tenets of personalized medicine to which we all aspire.

However, sharing patient data is not so simple. What a treatment does to a patient is hard to report without describing who that patient is. For example, sharing information about cancer mutations might reveal other risks in that patient’s DNA or that of his offspring. In other words, patients’ privacy is at stake. Traditional methods of de-identifying patient data are helpful but not airtight — data can be mishandled or worse, re-identified by malicious actors. In fact, HIPAA (the Health Insurance Portability and Accountability Act of 1996) provides clear barriers to sharing patient data in the United-States. Moreover, patients’ concern for privacy is asymmetric. As the recent data breach involving Cambridge Analytica and Facebook demonstrate, it takes only one breach to nullify many episodes of successful preservation of privacy and destroy trust.

**We think getting valuable patient data for cancer research while maintaining data privacy has a simple solution: just ask the patients to share. **Seem too simple to work? A recent study of patient attitudes about data sharing revealed that 93% of patients were fully willing to allow their data to be shared with scientists. Only 8% of them believed that the potential negative consequences of data sharing outweighed the positive ones. This study included hundreds of patients from diverse backgrounds.

At first glance, these results seem odd. Why would patients be willing to take the chance that their identity would be inadvertently or maliciously linked to their medical records? This is especially perplexing considering the relatively slow pace of scientific research. Patients who share their data today are unlikely to personally benefit from scientific discovery in their lifetimes, especially those with metastatic cancer whose lifespans are realistically anticipated to be weeks to months.

We believe the answer lies in the human drive to leave a positive legacy in the world. Research finds that legacy motivation is especially strong among people who are near the end of their lives and who wish to live on through their positive impact on others. This is the case, for example, with kidney donors. It is heart-breaking to consider the situation of terminal cancer patients who must face the inevitability of their prognosis. Seeking to provide a meaningful contribution to the world is a noble way of dealing with these difficult circumstances, and people naturally view consenting to contribute their data to medical research as a way to help others in the same situation.

“Grave men, near death, who see with blinding sight. Blind eyes could blaze like meteors and be gay. Rage, rage against the dying of the light.” — Dylan Thomas

Thoughtful incentive programs can improve the infrastructure for patient data sharing. The legacy incentives that motivate patients to share their data can be enhanced by creating tokens that can be transferred to future generations along with patient data. These tokens could have some future value in the form of rewards from philanthropic foundations or tax implications for the family of the token’s namesake. They could contain photos, essays, unfinished works, or will instructions for future generations. Different pieces of this information can be made accessible to different people through different levels of encryption. Parents could incorporate as much of themselves as possible into the token before gifting it to their children. It could even be turned into a gift to a stranger who benefitted from the cancer research made possible by your data.

The upshot is that these tokens can serve as an important psychological source of meaning. Knowing that a future patient who would benefit from the scientific discoveries made possible by one’s data would also receive a token that symbolizes one’s contribution can be a tremendous comfort to a current cancer patient. The additional comfort of knowing that the blockchain will be an immutable record that could realistically outlive any central actor is naturally attractive for legacy purposes. More generally, networks that connect current and future generations and motivate them to “pay it forward” can make the current generation’s goal of leaving a positive legacy more likely to happen. There is also a cross-sectional community aspect as well, where a group of patients may provide support and motivation for each other to contribute by sharing their own stories and motivations.

Giving an outlet to patients’ legacy motivation can also sidestep privacy concerns expressed in HIPAA. While a doctor, nurse, or well-wishing pharmaceutical company or medical center is legally liable for leaking data, an educated and technically empowered patient can share her data with the world without repercussions. The law today is clear: patients own their medical data and may do with it what they wish. Speaking more practically, however, most patients are simply not in a position to exercise complete control and distribution over their data. This is due to a number of factors, like monetization by providers, the misunderstanding of data rights, and technical hurdles associated with transferring large imaging and genomic datatypes common in modern cancer care. By addressing patients’ legacy motivation and desire to do good in the world, we can give control of medical data back to the people who rightfully own it — patients themselves.

Suppose you need to assemble a team of people, each with a highly specialized set of skills and knowledge. You may have some general idea of who they might be, but you cannot guarantee they know, much less trust, each other. They will need some coordination and assurance that they will get paid. What is the best way to assemble your crew?

Businesses regularly find themselves in a similar situation. Consider a company, for example, that builds houses from scratch and sells them. They will need to contract engineers, electricians, plumbers, a realtor, and so on. What’s more, these complete strangers will need to coordinate and collaborate if the project is to ultimately succeed. In a traditional self-managed team, project managers direct the communication, teamwork, and workflow between team members. They assign responsibilities, expand or reduce the team’s size as needed, and supervise the handoff of work from one team member to another.

Complete strangers will need to coordinate their work if the project is to ultimately succeed.

Organizational behavior research suggests that traditional methods of organizing people are incredibly inefficient. Communication can become repetitive and unclear. It can be vague where one person’s work ends and another’s begins. Team members can be uncertain about the broader mission of a task and their individual role within it.

Researchers at Stanford have shown that flash teams can be a more efficient way to organize people. Flash teams are groups of people with special skills and knowledge that are assembled, organized, and rewarded through a centralized platform. More concretely, flash teams are efficient for the following reasons:

1. They automate tasks to reduce costs and time to completion.

2. They feature chat boxes, workflow visualizations, and clearly designated tasks for each member.

3. They can automatically recalculate estimated start and completion times as members submit work.

4. They can automatically manage the number of people on a project, allowing teams to expand and shrink with great ease.

Flash teams can be used for any kind of work that requires different people to contribute different skills and knowledge to parts of a project that can be more or less isolated. Some examples include designing a webpage, producing a movie or animation, and running entertainment events. Still, this style of crowdsourcing work has yet to achieve scale.

Flash teams today rely on a centralized system to ensure trust between their members and the fair distribution of rewards. One can imagine centralized flash teams that fail to attribute value to its contributors. They might unfairly allocate work and rewards, for instance, or profit by skimming too much off the top.

Blockchain technology may be a path forward to scale flash teams trustlessly. On a properly designed blockchain, each work transaction is recorded on an immutable ledger. There is no central authority that can exert malicious influence. Even the participants themselves can’t abuse the system, since doing so requires contributing significant amounts of work. A blockchain can securely contain all user and token information in a decentralized way. People can engage in productive behavior without trusting one another before the fact. Everyone verifies the work performed and the distributed rewards, so it is incredibly hard to cheat the system without ramifications.

Finding the right team members is much like a “needle in a haystack” problem, similar to the Red Balloon Challenge. These problems can solved by incentive programs, including recursive incentives. Blockchain inherently affords security, transparency, and scalability. Paired with the right incentives, flash teams can be more effectively formed and recruit members.

Flash teams represent an idea that is much larger than outsourcing. The two main ways to organize human work at scale are firms and markets. By building the correct incentive markets to enable and scale trustless flashteams, we are moving towards a future where we can deconstruct the firm model. By allowing market forces to enter into the ossified and inefficient domains of firms in the form of flash teams and other incentive markets, we should aim to build a future of work that better unlocks our common potential.

How do you find the best person for a job in the fastest way possible? The answer to this question is something of a holy grail for startups. A recent study has shown, for example, that companies with effective recruitment strategies double the profit margins of their competition.

The flip side of this is many people don’t have the opportunity to be in fulfilling careers— instead merely clocking in and clocking out of their jobs for a paycheck. The result is a broken labor market that operates with a large output gap from the efficient case where everyone has the best job for them and companies with the best ideas are properly resourced.

The most popular recruitment strategy involves storing applications from past job postings and sorting through them each time a position opens. This requires retaining and rereading thousands of resumes for each new hire, which is simply a waste of time and energy. 

Our idea is to treat recruitment as a needle in a haystack problem, which essentially calls for an optimized search to find the best candidate for the job.

Suppose, for example, that a startup company called CrystalAI is developing a novel virtual assistant. They need to hire ten developers, but they can’t exceed a fixed budget. They want to recruit a diverse group of the highest-quality candidates, avoid being buried in unsuitable resumes, and leverage an external network because they don’t have the critical mass to rely on internal referrals. Their solution is to purchase tokens in the form of 100 CrystalCents, which they distribute to an initial group of graduating computer science majors at a research university nearby. Alice is a student who receives a devCent, and she reasons as follows:

1. Alice can apply to CrystalAI, and if she’s hired while holding a CrystalCent, she gets a $10k token bonus.

2. Alice could also send the CrystalCent to Bob. If Bob gets hired while holding a CrystalCent, he gets a 10k token bonus while Alice gets a $5k token bonus as the direct referrer.

3. Bob could also send the CrystalCent to Carol, and if Carol gets hired, she gets a 10k token bonus, while Bob and Alice each get a $2.5k token bonus.

4. If no one applies to CrystalAI holding Alice’s initial CrystalCent, it expires and can be reseeded to another student. Once all ten positions are filled, all of the tokens expire.

Promotes better recruiting through external referrals

Recall that CrystalAI will hire at most ten developers, and each hire through the CrystalCent program costs them no more than $15k of token. This puts an upper bound (of $150k of token) on the cost to run the program — which is quite competitive with traditional headhunting strategies. The expiration date of the tokens also limits the amount of time that the program can run — accelerating the recruiting process and allowing the company to discharge its financial responsibility for the program after a predetermined time.

Furthermore, if Alice knows Bob is not interested in CrystalAI, but is fascinated with FoodAI, she can exchange her CrystalCent for FoodCent to give to Bob — all via distributed exchange and without CrystalAI and FoodAI being in a trust relationship with each other!

The recruiting mechanisms in place today are absurdly inefficient. Fixing these inefficiencies could dramatically decrease unemployment, wasted time, energy, and money. This solution would enable companies to find the best candidate in record time.

Go to Part 2

What does all this mean for incentivizing decentralized networks?

As discussed in Parts 1 and 2 of “Money Can’t Buy Happiness (or Incentives)”, monetary incentives are not as straightforward as is typically assumed.

To be effective, monetary incentives must not be too high or too low, cannot replace an existing social norm or social obligation, and cannot be understood as a signal that the task is not worthwhile. In certain situations, non-monetary incentives may be more effective than money in motivating action within decentralized networks.

In contrast, money can be an effective incentive when people interpret it as a signal that their work is valued. Money can also be effective when the problem requires people to increase their current level of effort because people are sensitive to the magnitude of a monetary incentive (so long as it’s not exceedingly high or low).

Beyond these points, the psychology of incentives highlights the importance of understanding the context of each and every situation in which decentralized networks are employed.

Decentralized networks enable customized applications to be developed to solve a wide variety of problems in business, government, social activism, space exploration, education, and health.

Each of these areas, and each of the specific problems in these areas, will have different social norms, conventions, and idiosyncrasies that would change the psychological meaning that people impute unto monetary incentives, the standards for what is considered “high” and “low” amounts of money for a given level of effort, and the degree to which people closely monitor the effort they put into a task in exchange for money. Designing monetary incentives effectively requires knowing the context well. Experimenting on a small scale with different incentives can also be helpful, if the situation allows.

Below are some interesting reads from the academic world — we made sure that these are accessible and insightful for non-academics. In future articles in this series, we’ll explore non-monetary incentives, the uncertainty inherent in incentives in decentralized networks, and the types of motivations that incentives can tap into.

Heyman, J., & Ariely, D. (2004). Effort for payment: A tale of two markets. Psychological Science, 15, 787–793.

Gneezy, U., & Rustichini, A. (2000). A fine is a price. Journal of Legal Studies, 29, 1–17.

Ariely, D., Gneezy, U., & Loewenstein, G. Mazar, N. (2009). Large stakes and big mistakes. Reviews of Economic Studies, 76, 451–469.

Liu, P. J., Lamberton, C., & Haws, K. L. (2015) Should Firms Use Small Financial Benefits to Express Appreciation to Consumers? Understanding and Avoiding Trivialization Effects. Journal of Marketing, 79, 74–90.

In a previous article, we made the point that people’s behavior as a result of monetary incentives is not linear. In this article, we’ll discuss how non-monetary incentives affect motivation and how they can be used at scale in decentralized incentive markets. 

Non-monetary incentives are used less often than money in business and government, and the reasons are not hard to understand. Money is easy to quantify and describe, and the assumption that people respond to money is intuitive and appealing. But it might surprise many to learn that this is an exaggerated assumption. Yes, people respond to money, but not as powerfully as we think. A starting point to correcting this over-emphasis on money is to get a better understanding of non-monetary incentives. 

1. Consider a study done in in 1999 by psychologist Chip Heath. The study asked managers in a Citibank call center to assess what motivates their employees. Compared to how their employees described themselves, bank managers systematically overestimated the employees’ motivations related to money and compensation, such as salary and fringe benefits. Managers also underestimated the power of non-monetary incentives, such as the desire to learn new skills, doing something that makes you feel good about yourself, and accomplishing something deemed worthwhile.

2. The same result is found in studies involving much larger samples. Gallup repeatedly finds in its national surveys that the main reason people quit their jobs is not money, but rather that they feel unengaged in the work. Likewise, the General Social Survey (a massive nationally representative survey done biannually since 1972) finds that people rank “doing important work” as most important for them but believe that money is required to get others to work hard.

A strong source of non-monetary incentives are group affiliation, tribalism and cultural buy-in. To understand how deeply the desire to be part of a group runs, consider a classic field experiment done by psychologist Muzafer Sherif and his colleagues. In the halcyon summer days of 1954, Sherif’s research team recruited 24 twelve-year old boys to participate in a study in the Robbers Cave State Park in Oklahoma. The boys were all from similar backgrounds, socio-economic, and age groups. Initially, the boys were allowed to spend time as one large community and befriend one another. All the researchers had to do to create group affiliation — and the accompanying intergroup conflict — was to simply randomly assign half of the boys to one cabin and the other half to another cabin and cut off communication between the two groups. Almost immediately, the two groups developed their own nicknames, rituals, jargon, and values.

The tribal intergroup conflict took enormous effort to overcome, despite its absolutely random beginning.After a few days, the researchers asked the boys to name their friends in the camp. Only 8% the boys named outgroup members as their friends, even though just a few days earlier the boys were all friends before the random groups were created. Fierce competitions, slurs against the outgroup, and even a late-night cabin raid to turn the rival boys’ stolen jeans as a trophy ensued.

Another example. In a classic experiment, participants who were otherwise oblivious to art were shown a number of paintings made by two artists, Paul Klee and Wassily Kandinsky.

Participants then picked the favored artist of these two, and were asked to divide money between themselves and other participants.

Simply knowing that another person chose a different artist led participants to take money from this other person and give it to the people who chose the same artist as them. Tribalism happens very quickly and does not need a rational reason to occur.

What are the implications for decentralized networks?

When tribalism can be used for good, it can be a powerful incentive — much more powerful than mere money. For example, tribalism can be especially useful to motivate sports fans, music fans, and members of small towns to join networks to solve specific problems. If completely random assignments in experiments can create loyalty to the ingroup, just imagine what how motivating belonging to actual groups is.

Go to Part 3

The Goldilocks problem: Monetary incentives

The most obvious way of incentivizing people is promising them money in exchange for their contributions. Money is overused and over relied upon as a proxy for value since it’s easy to measure, compare, and describe. However, academic research and everyday life suggest that people’s response to monetary incentives is nonlinear. In other words, the effort people invest in a task does not always increase in proportion to how much money they make. So, using solely monetary incentives to bootstrap abundance and financial utility is right road wrong direction. 

1. Perceived-trivial monetary incentives disincentivizes: Trivial monetary incentives can backfire by insulting people and reducing the amount of effort they’re willing to put into a task. A tiny monetary incentive could be construed as a message that a person’s work is worth very little. The reason is that people not only observe the magnitude of the incentive, but also ponder its social meaning. To illustrate this point, consider the case of daycare centers in Haifa, Israel that were trying to reduce the number of parents who showed up late to pick up their children. Parents’ tardiness was costly to these daycare centers because employees had to stay late with the children. So, the centers instituted a monetary fine for tardy parents with the logical expectation that this fine would deter late pickups. To the chagrin of the daycare centers, the proportion of late pickups increased! The reason this monetary (dis)incentive backfired was because it replaced a social norm. Before the fine, parents tried to get to the daycare center on time to pick up their kids as a way of honoring their commitment to the daycare center owners. The fine suddenly put a price tag on being late, and it wasn’t a very high price. Instead of disincentivizing tardiness, the moderate fine conveyed that being late really wasn’t all that bad. Parents’ late arrivals in daycare centers increased after introducing a monetary incentive to show up on time (Gneezy & Rustichini, 2000). Trivial monetary incentives can result in less effort than giving no monetary incentive at all.

2. Perceived-over ascribed monetary incentives **backfire. **When money is seen as a 1:1 to value, the receivers feel obligated to reciprocate but are unable to, resulting in stress and dispiritng. For example, the behavioral economists Dan Ariely, George Loewenstein, and their colleagues gave different amounts of money to people in rural towns in India for performing a variety of tasks that involved creativity, motor skills, and memory. In most cases, people who received large incentives (as high as full monthly incomes in that region) performed worse than those who received less money. In general, high payments create more stress to perform well, and at extreme levels this can hurt performance.

3. **Money cannot substitute virtue **when value is human-centric social capital, reputational, or bound by social obligation/social norms. For example, you are more likely to offend a friend offering trivial or non-trivial monetary incentives in exchange for a favor.  Money can change the psychological meaning of an interaction. Money can be detrimental in situations where people were already doing something out of social obligation or social norms.

Getting money in exchange for effort creates a mindset of costs and benefits. To illustrate what this means, consider the psychological difference between the way we act in our social relationships and in our work relationships. We usually don’t nickel-and-dime our friends for every favor we do or every meal we buy for them. But in our work, we closely monitor how much effort we put into every task and we frequently ponder whether we are paid enough for our toil and sweat. At times, we do less because we do not believe that we are being justly compensated.

This point is illustrated in an experiment done by Dan Ariely and his colleague, in which people were given either candy or cash for doing a task that required effort and patience (using a mouse to drag digital gray balls from one side of the computer screen to the other side of the screen). Participants’ effort was sensitive to the amount of cash they received, with more money leading to more effort. But participants worked just as hard for a few pieces of candy as they did for a lot of candy. Participants who received money closely monitored their effort relative to how much money they got whereas participants who received candy did not engage in this cost-benefit monitoring.

In summary: People take things too literally in general - in particular with respect for matters of money. By fixating on money as a proxy for value, humans are unable to compute the real intrinsic value of the underlying thing when it has a price tag. For example, studies find that introducing money (the quintessential extrinsic incentive) to an activity that people have been doing for enjoyment reduces people’s willingness to continue engaging in that activity. In other words, introducing extrinsic incentives can reduce intrinsic motivation. Introducing a monetary component to a task that people are already doing for enjoyment can replace their intrinsic motivation with anticipation for monetary rewards, and dispirit participants in a network.

Money can be effective in getting people to do something they do not find meaningful or that they are not already doing — but not every unsolved problem is characterized by these obstacles. Sometimes people are highly motivated to work on a problem, but they cannot organize a network efficiently to solve it. For example, consider sports fans who are very intrinsically motivated to support their team. Introducing a monetary incentive in such an environment might backfire because it would replace the fans’ existing intrinsic motivation. However, symbolic incentives that build on fans’ interests can work much better — for example, the chance to win a signed jersey or meet with star players on the team. Fans may be more likely to join a network for these symbolic incentives than for money.

Designing the optimal incentive system matters.

Go to Part 2

Infinite Incentive markets secret sauce: Weak ties

Weak ties are totally underutilized, especially on the internet. What’s more, we think there is a simple way to systematically find and exploit these connections, which is why we need Infinite Incentive markets. The combination of incentives and blockchain technology can be an effective way of cultivating weak ties. Incentives make it worthwhile for people to join a network. Blockchain technology creates the security, scalability, and auditability of the network — and it does this in a completely trustless way. A blockchain-driven incentive market connects everyone through multiple paths, reducing the reliance on specific bridge nodes and increasing the effectiveness of the entire network as a result.

Restore value attribution.

Everyone will reconstruct the same chain of referrals by querying the smart ledger associated with blockchains, so value attribution is unambiguous. This is important because people will contribute if they believe they will be rewarded accordingly, and blockchain makes this possible.

Provide for trustless exchange.

Blockchain allows the formation of distributed exchange, where parties don’t have to trust each other to exchange assets. Furthermore, sponsoring organizations can deliver value to a user without being in a trust relationship with each other.

Scalable at Internet size.

Incentive markets and recursive incentives to date have not reached internet scale. Part of our thesis is that blockchain properties are unique and facilitate a complex contour of trust forming among a spirited community, that facilitates the scaling of these types of networks. We also address certain problems such as Sybil attacks generally, facilitating creation of ad hoc incentive programs that benefit the world.

The idea that “weak ties” are more critical for successful networks than “strong ties” has persisted over the half a century since sociologist Mark Granovetter published the original academic paper that established this idea. In a world where technology is creating tremendous advances in our ability to create and sustain social networks, it helps to remind ourselves of the original discoveries in network theory. These discoveries also help highlight the potential power of tapping into our “weak ties” with blockchain technology.

What are “weak ties” and why are they important?

Imagine that you recently found out you will be let go from your job, so you must find a new one. You want to spread the word to people in case anyone knows of an opening. Who will you choose to tell: close colleagues you worked with in the past or people in your profession you’ve met once or twice?

The weak ties theory proposes that you tell the latter group. At first, this might seem odd because close colleagues are more motivated to help you than people you barely know. However, your “weak ties” are much more likely to know people outside of your close network and therefore have access to information that your close network lacks. Information about the right job opening for you is more likely to come from your weak ties than from your strong ties.

Information about the right job opening for you is more likely to come from your weak ties than from your strong ties.

This, in essence, is the weak ties theory. Your strong connections are likely to be siloed: All of your close contacts are likely to know each other and spend most of their time with each other. Very little new information is going to come into your close network. In contrast, your weak connections likely travel in different circles and so will have access to new information that your close network lacks. The strength of your network is proportional to the number of your weak ties.

Weak ties are more likely to be bridges to new networks

Granovetter’s original article distinguishes between regular network paths and “bridges.” All networks contain paths that connect between nodes. A path becomes a bridge when it is the only one that connects a specific node to another. Bridges are particularly important for networks because they connect spheres of information that would otherwise not speak to each other. For example, a bridge might be a medical doctor whose brother is a lawyer and who can connect doctors to the legal world when the need arises. Or a bridge might be a U.S. government official with personal contacts in a foreign country and who can facilitate cross-country communication on sensitive policy issues.

Bridges are key to information transfer in networks. Without them, networks become debilitated. The more bridges a network has, the faster information can flow through it. Moreover, a network with many bridges is likely to have access to more diverse information.

Granovetter suggests that weak ties are far more likely to be bridges in a network than strong ties. This is because your strong ties — the people you know well — are also likely to know each other well. This means that there would be numerous paths from you to any one of your close contacts. Thus, none of your strong ties is likely to be a bridge to another network. In contrast, weak ties are more likely to be connected to networks you do not know, and so they are much more likely to be bridges. Weak ties are more likely to be connected to networks you do not know, and so they are much more likely to be bridges.

The power of weak ties is well settled with data

The data Granovetter brings forth in his 1973 article is illustrative in helping us see the importance of weak ties. He cites the ingenious studies done by psychologist Stanley Milgram that tested whether we live in a “small world.” A participant in the study was given a booklet that he or she must then pass on to some randomly chosen person in the U.S. through referrals and mutual acquaintances. The study found that the proportion of successful deliveries ranged between 12% and 33%, and the number of links in the chain ranged from 2 to 10. But more importantly, the study also found that sending the booklet to an acquaintance made it more likely that it would eventually be delivered to the right recipient compared to sending the booklet to a friend. This teaches us that connecting with weak ties enabled networks to be more successful.

Another illustrative study that Granovetter cites was done by sociologists Rapoport and Horvath in 1961. Students in a Michigan junior high school were asked to rank their 8 best friends. The researchers created a network from the persons listed at the highest rankings and another network from the persons listed at the lowest rankings. The results were clear: The network created from all the best friends was much smaller than the network created from the acquaintances. The aggregation of weak ties led to a much larger network than the aggregation of small ties.

Go to Part 2

Given how effective the strategy proved to be in the case of the Red Balloon Challenge, why haven’t similar recursive incentive programs become more popular in the last ten years? The most general answer to this question is that incentive markets have been unable to achieve large (i.e. internet) scale because they require more trust than prevailing technology affords.

More specifically, it is prohibitively difficult at large scales to:

1. keep track of large amounts of data, like all of the information about who is participating and who referred who;

2. keep large amounts of data safe from being tampered with or manipulated by third parties;

3. verify that referral chains are authentic and not fraudulent, especially when there are complex rewards attached to them.

4. allow people to participate without collecting irrelevant information about them that they may want to keep private (i.e. anonymous credentialing); and

5. get people who may not trust each other to jointly participate.

Trust is the missing ingredient.

Enter Blockchain

I was trying to solve the problem of scale for years before I came across blockchain technology. My eureka-moment was the realization that a blockchain could provide the substrate for coordinating incentive markets at scale, without requiring anybody to trust anybody else.

But what exactly is a blockchain, and how does it overcome the problems listed above? A blockchain is, most simply, a ledger of every single economic transaction that occurs with a digital currency, arranged by account type. The ledger is updated at regular intervals of time called ‘blocks.’ The ledger also contains within itself a history of these updates, making it quite literally a chain of blocks, or blockchain.

“The interesting thing about blockchain is that it has made it possible for humanity to reach a consensus about a piece of data without having any authority to dictate it.”

— Jaan Tallinn

The basic unit of any cryptocurrency is a coin, or token. These tokens can actually have data attached to them, which we call their metadata. The metadata of a coin can contain its history — where and when it was created, where it is now, and who held it along the way. It would be like if a penny carried around its own autobiography with it everywhere it went.

Blockchains have a unique set of properties that enable incentive markets to form trust-lessly and, therefore, at scale. In particular, blockchains are auditable, secure, and identity-friendly. Let’s evaluate each of these in turn.

Blockchains are Auditable

A key property of a blockchain is that it allows people to participate in an incentive market trustlessly. Blockchains also solve the problem of keeping track of large amounts of data by recording the life of every token on its ledger. When a person gets invited to help solve a problem (like find a “red balloon”), they’ll receive a token. When they recruit someone to join the search for a solution, they do this by transferring their token to who they think can help. When the problem is solved, rewards are distributed. All of these transactions are recorded on the ledger of the blockchain. This makes tracking large amounts of referral chains more manageable. This gives people the confidence that rewards will be allocated properly without having to trust anyone.

Blockchains are Secure

What’s more, there are multiple parties responsible for maintaining a copy of the ledger on their computer. In fact, the ledger doesn’t actually exist outside the sum of all of these copies. (That’s when we say a blockchain is decentralized.) There are sophisticated programs that allow each of these copies to be in sync and be robust to faulty and even malicious copies. The fact that the ledger is stored in many different places and verified communally helps ensure safety, security and immutability of the ledger. A decentralized network doesn’t require as many trust relationships as a centralized authority.

Blockchains are Identity-Friendly

But what’s to stop someone from creating multiple fake accounts and referring them to each other in order to reap the benefits of a referral program? The risk of fraudulent behavior in the form of these so-called Sybil attacks can be mitigated in a number of ways.

(i) One option is to require the members of a successful referral chain to prove the relevant parts of their identities before receiving any rewards. Companies like Civic and Uport, working together in organizations like Identity Foundation and Consensys, are working on providing these identity verification services to blockchain companies. This is one way to verify that referral chains are authentic.

What’s more, these services are capable of verifying only the relevant parts of someone’s identity, unlike a bouncer who verifies more than someone’s age when handed a driver’s license. This means that people can participate in programs on a blockchain without providing irrelevant or unnecessarily personal information. In some cases, they can even participate anonymously, for all intents and purposes.

(ii) Another option is to require people, over the course of participating in a referral program, to perform a task that is not easily automated but is still easy to verify. This is called requiring proof of work, and it can make Sybil attacks prohibitively costly.

(iii) Since rewards becomes smaller as the degrees of separation from the “red balloon” increase, every participant of a referral program wants to avoid transferring their tokens to someone who might perform a Sybil attack. As a result, people will be motivated to screen the recipients of their transfers, mitigating the risk of fraud.

How to reduce the probability of a Sybil attack is an ongoing research program for all blockchain companies, and the threat worsens when networks get very large. Still, deploying all of the tools listed above can mitigate the risk.

Incentive markets are the key to motivating people to collaboratively solve complex problems, like the Red Balloon Challenge, but they have been unable to achieve internet scale given the amount of trust they require. Blockchains are auditable, secure and identity-friendly, which makes them the perfect fit for enabling trust-less incentive markets to form.