If you’re looking for new content from me, chances are, it won’t be here as I’m finding it hard to publish on both platforms.

Thank you for your following and support — more to come on Substack!

Wendy Xiao Schadeck

The VC game seems to be changing. It’s become a big enough industry with enough widely observed idiosyncrasies that it’s possible to raise an entire fund based on VC memes. One common meme is the mythical VC “value add”— seems like founders and VCs alike are trying to figure out what it actually is.

One reason is probably that increasing competition between too many funds makes it hard for founders to differentiate (however, there is plenty of competition amongst investment bankers, and the value-add to companies is definitely debatable, but they still don’t receive nearly as much meme attention). Another is high expectations, overselling, and bruised egos on all sides of the VC dealmaking process in the past, present, and future.

Memes aside, data also show that there is a discrepancy between founders and VCs on the latter’s perceived level of “helpfulness”:

Five years into venture capital, this is a topic I reflect on a lot because it actually defines the journey I will be on in my career. I try to routinely map out the kind of skills I have to optimize for to be successful in my career (e.g. to be picked by founders) against my goals as a human being and see if they align. Therefore, the answer matters a lot to me, personally.

Here, I will attempt to provide a more nuanced perspective on VC value-add than: “our platform, our network, and our experience,” and I also hope to also build a narrative against the one we seem to be converging to right now: “our money, the fastest, and more of it for less of your company.” I write this with the full understanding that most of it is stuff savvy founders probably already know, but sometimes it’s just helpful to have a framework to think things through, outside of the superficial stuff. This may be especially true during this unique time in the market with an overabundance of venture capital, very limited human interactions, the shortened length of a deal cycle (I’ve heard as little as 9 days end-to-end, to pick a partner for 10+ years potentially?!).

This post, admittedly, may also be my biased way of rationalizing against the evolution of early stage investing into something much more transactional (but hopefully not). Because if the main takeaway from an early stage fundraising process is “how much capital and how fast”, then surely there must be other ways to distribute capital than to tie up hours and hours of talented founders’ time with calls, coffee chats, and dinners.

So what can a VC give your company?

Early stage VC is probably one of the most expensive things you will procure in the lifetime of your company, and there is no stack-ranking of best to worst because it largely depends on what you need from your VC and what kind of company you want to build, but there are some ways to more or less objectively qualify what they bring to the table:

1. Access to capital is first and foremost. It is the lifeblood for the company, and when in famine, it might be the only criteria that matters. I’m not here to convince any founder they shouldn’t set a fair price for their company because that is really important. However, access to capital doesn’t stop at this particular term sheet:

• How real are the terms? Have they done the work to understand your company or do they just want a seat at the table? Have they backed out of or renegotiated deals before?
• What is the investor and his/her partnership’s approach to thinking about you de-risking the business such that they’re willing to support your company if the market falls out or you’re just shy of your milestones for going to the market?
• What is your investors’ conviction in you based on? Is it borrowed? and how fervently will they fight for you when their conviction is tested?
• How much support does your investor have within his/her partnership to build a case for allocating more capital to your company?
• How do you rank against the other portfolio companies in that particular fund in terms of importance and size of position?
• How much dry powder is left in the fund which invested in you?
• What kind of co-investors or follow-on investors will your investor attract for subsequent rounds?
• What kind of signals does it send to the market if they invest or don’t invest into your next round? What kind of signals do you send by accepting their money?

Acknowledging that you most likely won’t get the most helpful answer by asking these questions directly and the assessment will largely have to be done by triangulation. However, by framing the questions around past investments scenarios, you can build towards a better understanding of your partner.

2. & 3. Access to knowledge and access to network are important badges worn on many VC’s websites and are pretty well understood at this point. Different entrepreneurs value different things here, and evidence is mixed on how value-adding it really is, but usually evaluating criteria revolves around:

• Does this person REALLY understand your industry, your customer, and will care enough about the problem you’re trying to solve to keep their understanding current over the next 10 years?
• Does this person have very specific operational insights that can help you shortcut your decision-making process through trial and error, and do you trust that they are still relevant?

(I want to harp on this one for a minute because I think this is overemphasized in investor conversations. Very rarely do we come across the exact same situation twice, and most of the time the entrepreneur will go with her instinct rather than follow someone else’s rubric exactly. What is perhaps more helpful is using these experiences as sounding boards to test the boundaries of experimentation, or method advice to help transfer tried-and-true processes for finding the answer yourself. These operational heuristics can be incredibly expedient, but they’re extremely imperfect as well and can be dangerous when we forget that.)

• Does this person have enough context to help you frame your strategic choices and discuss tradeoffs with you in a way that helps you come to a better decision?
• Is there some inherent network effect of being in this VC’s community? How do they encourage this?
• Does this person help you expand your network in an unobvious way? Do they have also some sort of information advantage on their network (e.g. Talent tracker tools, etc.)?
• How willing is their network to go the extra mile for them? and how much does their word mean (e.g. are they all tapped out)?
• Does your partner offer professional services that your org can get leverage from? And for how long will you need them?

Perception of Helpfulness of VC Operating Teams

4. Access to human. I may be naive in thinking that this is perhaps the most important dimension for picking an investor, especially if you’re thinking about bringing this person onto your board. Ultimately upon signing the docs, you’re picking a partner who now has a major role in your future and the future of your company for a decade. Arguably, the capital, institution, and access to knowledge and network can be commoditized over time — they are essentially table stakes for any modern VC fund — but the human element remains super subjective, as it usually does. Since there are now so many talented humans working as venture investors, and the success of the VC-founder relationship has so much to do with achieving fit:

• What will they be like in 10 years based on what you know they’re maximizing for now? Will you still want to work together then?
• How much of themselves are they willing to dedicate to you? How might that change over time? Sometimes that is directly impacted by how many other portfolio companies they have.
• Do you trust this person? There will be ultimately be times when your incentives will be at odds. Do you trust they will stick to a set of values that still makes it possible for you to collaborate during those times?
• Do they have a different enough perspective but good enough communication with you to help you expand yours? Do you get energy from looking at the world through their eyes?
• How does this person fit into your existing board dynamics?
• Is this investor serious about a long term career with their current fund or are they likely to hand you over to someone else?
• How much skin in the game do they have for you to be successful? How much does their wealth or (perhaps more importantly) their reputation depends on your success?
• How does the person make you feel when you interact, and how does that help you achieve your goals as founder?
• Do they care more about your success more than they care about their ego? (very easy for a VC to be self-unaware about this)
• Does your investor actually know you? and invest into knowing how to better work with your personality / biases / style?

I fully realize that founders often don’t have the time or perhaps the interest to sit and dwell on their VC’s personalities, but it makes me upset when they structure the process such that it basically discourages any form of mutual discovery of human traits. And maybe the sad reality is that founders are converging towards treating their early stage investors as they do with public investors, and the human element isn’t as necessary, especially if the control rights are also structured as such. While this may be true for some founders, but I truly hope this is not the case for all.

Some of the most fruitful VC-founder relationships I’ve seen leading to amazing outcomes have come from growing and developing with one another over the course of decades, and some of the worst have led to a total time waste for everyone involved to energy-draining, or toxic boardrooms.

Feels like an easy win would be to apply more rigor to the process of picking and getting more out of your investor relationship in exchange, rather than taking the human element out of it altogether.

Additionally, VCs should also invest more into developing themselves in this way to better serve the founder (a good article on that from our friends at The Generalist). Some of the best board members I’ve observed have developed it through years of experience, and others have personalities that are naturally more suited to being a good advisor. However, there is a lot of skill-building that VCs can be proactive about and a lot of literature around what good advisors should do. I don’t see this discussed enough in VC circles:

Personally, as I reflect on my path in venture, I think about how I can objectively check the boxes on #1–3 and work to improve my skills in #4. Ultimately, there is still this element of “fit” that will always remain subjective, which means that I won’t be a good choice for every founder, just as not every company’s mission will speak to me personally. I am very curious in the founder perspective after reading since this has largely been a thought exercise for me in order to develop a set of hypothesis against which I am mapping out my own career milestones.

So founders — what do you think? What have you gotten from the best VCs you’ve worked with? How can we optimize the early stage VC process for value-discovery for you — whether that value is the personal relationship, the network, the knowledge, or just the money (and more of it for less of your time)?

How Can I Be Helpful, Really? was originally published in The Startup on Medium, where people are continuing the conversation by highlighting and responding to this story.

As Ballmer put it so energetically in the past, it always comes down to developers. Developers are always at the forefront of technology, but they haven’t always been served directly by technology. The previous paradigm of tech unicorns built their fortune on the backs of developers by convincing them to build within their walled gardens and then taking profits off the top. These tech giants subsidized the tooling as a public good since they reaped most of the value created within their gardens. Developers had plenty free tools to build with, but the experience was always what you’d expect with a public good — sufficiently working, but never mindblowingly good. VCs also hesitated to invest into the space because it’s hard to compete with free, and the value accrued to platforms.

However, more recently, as developers began to take power back from the walled gardens, some of the most valuable tech companies have won by building directly for developers — from public giants like Atlassian, MongoDB, Splunk, Datadog, Elastic, and Twilio to private unicorns like Stripe, Auth0, HashiCorp, Elastic, Snyk, and Unity. These companies and their peers have already created hundreds of millions of dollars in enterprise value and are still growing quickly. We are seeing an acceleration in the funding and adoption of developer-serving technologies which not only gives developer more superpowers but also has huge implications for businesses and product in the next iteration of the web.

Huge seismic shifts creating new building blocks

One of the reasons we’re seeing more innovation than ever in the developer space is because the existing landscape of software development is undergoing seismic shifts, at a faster pace than ever before. This not creates more opportunities for new products to solve new problems, carving out both wallet-share and mind-share from developers, it also creates new building blocks and dynamic infrastructures for developers to experiment with:

1.Breaking down of monolithic applications into microservices adds a lot of devops complexity but also entry & exit points for new software and data feeds. 84% of surveyed mid-to-large organizations have embraced microservices, on average, deploying 180+ microservices each.

2. There are now way more composable building units for developers to build on top of as APIs proliferate, partially due to microservices & popularity of GraphQL. There are more than 50,000 public APIs, and >60% of companies invested in a public-facing API.

3. Crypto protocols with native capabilities like payments, compute, identity, and liquidity allows developers to easily build business models into their projects without expending too much capital upfront to build out entire infrastructures to support these.

4. Accelerated adoption of containers and Kubernetes means applications are portable across different infrastructures, with ~90% of companies running containers to maintain flexibility, save on infrastructure, and increase developer efficiency.

5. Great front-end tools maturing to help make software development more accessible for front-end devs and even business users: Javascript is now the most popular programming language in the world, used by 70% of devs. They can access a suite of APIs using GraphQL, used by almost 40% of JS devs. Low/no-code tools like Webflow democratize web development even further to non-devs — more on that later.

6. Process innovations accelerating as well. Increased CI/CD adoption across enterprises requires new tools and processes across the organization.

7. How companies work with data is changing rapidly with advances in ML, with increasing focus on operationalizing big data and data engineering. 62% of companies are getting ML models deployed into production or have deployed them for already for several years.

These shifting landscapes create new complexities but also new building blocks for developers, and the core value prop of the next web is developers building with powerful web functionalities at a fraction of the risk. This will undoubtedly lead to a proliferation of experimentation and innovation.

From a “support” to a “build” mindset

Meanwhile, as all companies start to become technology companies, business priorities converge with developer priorities, so developers are empowered to build business solutions directly:

There used to be a bifurcation between technology needs and business needs, because technology was seen as efficiency tooling (thus a cost center) for the business functions, but a whole lot was lost in translation between the builders and the users, and you’d end up with systems that looked like this:

This is increasingly no longer the case. As all companies transform into technology companies, technology becomes a core part of the product. It only makes sense that the people building that technology are empowered to make the most important decisions about the products and services the companies use.

Alignment

This is extremely good for enterprises because business priorities can be integrated directly into the developer workflow rather than outsourced to separate, siloed teams. For example, as system performance became increasingly important to software companies, these metrics became integral to the developer workflow rather than outsourced to a separate ops team. Datadog emerged as a unicorn by helping dev teams directly monitor performance for the full stack as enterprise infrastructures shifted to cloud and the processes shifted towards DevOps. Synk similarly helped developers build open source code security into the process itself, allowing its customers to address issues at the root rather than rely on separate processes to find and fix them. We think there is room for many more business priorities to be integrated directly into the development workflow, such as authentication and privacy (Magic.Link, Evervault, Metomic) and cloud optimization (Northflank, Env0, Cloudskiff).

Productivity

As technology shifts further from a cost center to a profit center, developers can help the whole organization become superusers of technology to ruthlessly optimize productivity. Developers are naturally power users, and translating their processes to the rest of the organization can mean huge productivity gains. This A16Z talk from David Ulevich describes some compelling examples of developers’ typical ways of working which can seem eons ahead of our business processes. For example, developers are used to collaborating asynchronously in a seamless way using tools like Github, and products like Figma and Jam.dev are helping to extend this concepts to the design and product teams, but many business users are still stuck using inefficient naming conventions (v1, v2, vFinal, vFinalFinal, etc.) on static files to keep track of versions and sending them back and forth via email. Developers also prefer to use the most simple method of communication as that is basically what code is, and business users are also starting to catch on to command line communication in tasks like email (Superhuman), notes (Notion), and messaging (Slack).

R&D

As advances in ML continues to perfect machines’ ability to synthesize what already exists, being on the forefront of what’s new will become one of the most marketable skills. Developers don’t just write code, they hang out on the bleeding edge of technology for fun. While many have enterprise day jobs, they also have nights and weekend projects which are fertile grounds for testing the newest advances in technology, essentially serving as a free R&D product pipeline for enterprises. This is also a sweet spot for many developer-facing companies to address with free or low-cost product. Whereas it may be difficult to convince a fortune 100 to shift to a completely new user authentication tool, it is much easier to convince a developer to use it for his/her nights and weekend project, until there is enough critical mass for the many superusers together to champion the adoption of these new technologies into larger enterprises.

This kind of technology adoption has two GTM motions — the bottoms up product-driven growth of a consumer-facing product with the potential of achieving virality — and the top down enterprise sales approach with the potential of achieving high ACVs and loyalty. Perfecting this growth strategy is a powerful opportunity for developer facing companies. This leads us to us to our next topic.

New ways to think about sales and product

Developers at the center of businesses have huge implications on marketing and sales, product, and business models. Companies like Asana, Slack, and Figma have already begun to refine these methodologies, but the science area is still nascent compared to traditional enterprise sales. These are just a few non-exhaustive examples:

1. “Consumery” customer acquisition channels such as digital communities, developer evangelism, conferences / hackathons, great documentation, agency partnerships, etc. with implications on how to measure and refine the funnel and ROI for each.
2. A single stakeholder (or champion) sale with multiple “checkpoints” baked into the product journey, signaling depth of enterprise penetration, rather than a multi-stakeholder sign-off, all-or-nothing sale. This has implications on which initial stakeholder to target and how to help your champion sell into the enterprise (tiered pricing, documentation, case studies, business cases, etc.).
3. More potential for virality thus a need to better harness it when it happens but also how to precisely layer in the more cadenced, predictable enterprise sales approach once “landed”.
4. Developers are the most critical customers, so product experience is really important, much earlier than a traditional enterprise product. The “buttery-smoothness” of a user’s experience, which we often ascribe to consumer products applies much more here, with emphasis on documentation and onboarding.
5. More informed and vocal customers as well requiring innovative ways to channel product feedback. Many developer-facing companies include their customers in the product development process as well as roadmap prioritization decisions as a core product feature. This can sometimes be a key reason for adoption because devs, unlike any other enterprise customers, have the option of potentially building it themselves.
6. Sales needs to support product. You cannot over-sell and under-deliver for developer-facing products, but there are important applications of the science of enterprise sales to increase ACV once the product has gained a foothold within an organization.
7. Different monetization paths paves the way for more flexibility in the business models and a reduction in risks overall. From paid APIs, licensing agreements, to crypto tokens, developer-facing companies can price their services in a plethora of ways to carve out value. Companies can also use this to lengthen or shorten the timeline of value realization to make better decisions about how to capitalize the company.

The next level of abstraction in software development: doing more with more

Some may say that the low/no-code movement one day will eliminate the need for developers, and while it is true we are already seeing a blurring of the lines between developer tools and business productivity tools, I highly doubt this will be the case. At the highest level, developer are just experts at abstracting technology tools up a level to help other users use them. The better the tools get, the less work the developer has to do each time. When a business user can leverage basic technology building blocks themselves to achieve their outcomes, the entire organization benefits. Low/no-code platforms help fill the gap in developer resources within an enterprise.

Meanwhile, the components making up these building blocks are also getting more complex and powerful, requiring more developers to abstract them up into useful business applications. For example, the level of developer experimentation we’ve already seen with extremely complex open source technologies such as ML (e.g. GPT-3) and Crypto-protocols (e.g. Ethereum) is already staggering, but it will still take a while before an average business user will touch these technologies using low/no-code tools.

Also, if you think about how much of the world is still unserved by technology, it’s easier to imagine that the no/low-code movement will only create more need for developers. Developers will be the shepherds of technology to the rest of the organization, and these no/low-code platforms will be their collaboration ground.

Ultimately the venn diagram of creators technology, consumers of technology, and the buyers of technology are converging on developers, a population that is growing in its definition day-by-day. This means the pace of technological innovation should accelerate dramatically in the coming years, in ways we can’t yet envision.

We have been diving deeply into this space at Northzone and are excited for what’s to come. The future looks bright for devs, and through them, all of us.

The Future Belongs to Devs was originally published in Northzone on Medium, where people are continuing the conversation by highlighting and responding to this story.

The verbiage “Web 3.0” used to bother me because it implied the existence of an upgraded version of the web that we are already building towards, and as if there was a master plan to convert all users to this new web once the building was complete. This felt like predicting the future, which is not something I’m comfortable doing as a venture capitalist. Instead, to do our jobs, we have to listen closely to the builders and the users, constantly searching for signals, and based on these inputs, we use collective pattern recognition to build hypotheses about the types of changes we might see in the world. This is a bottoms up and an iterative process, and the resulting hypotheses can often be conflicting and dynamic. For me, there is no single version of Web 3.0, and this is what makes the idea more palatable.

When I wrote our investment thesis into this emerging theme, I focused on the economic rationale for the gradual evolution of underlying infrastructures (protocols) powering the shift into the next web paradigm. I argued that they were more efficient for coordinating certain types of work while extracting less value than centralized companies. It was too early then to trust any signals on end user value prop.

In fact, this has long been the billion dollar question for Web 3.0 — what does it do for end users better than the products of Web 2.0?

Many builders in this space were originally attracted to the early ideologies — privacy, decentralization, and self-sovereignty —and some mistake them for value propositions. However, the average end-user does not adopt technology for these reasons. Throughout the past few years, we have seen more and more user-focused builders enter the ecosystem, and the most powerful feature of this ecosystem, and technological innovation in general, is that it has the ability to connect the dots by incremental composability — talented builders tinkering, and more talented builders tinkering on top.

I have followed the Web 3.0 movement closely for several years while also studying signals and building hypotheses in Web 2.0 sectors such as ecommerce, health, fintech, and developer tools. It was only in the past six months that I started to see semblances of conversion between the value propositions of the fastest growing Web 2.0 startups and innate abilities of the early Web 3.0 rails. These signals indicate that internet users are looking for value propositions that Web 3.0 protocols are better equipped to build. This is not enough to say we are crossing the chasm yet per se, but at least we are beginning see where the bridges can be built. Here are the examples that I have identified so far:

Products with Community

We are learning that many of the most successful companies in Web 2.0 have been those who are being picked by their users because of their community. For example, Peloton’s community has been referred to as “cultish” by journalists, not unsubstantiated by some of its hardcore fans who proudly sport Peloton tattoos. The enthusiastic, supportive, yet competitive community of users, led by an attractive entourage of celebrity instructors, creates a powerful product benefit that far surpassed the experiences offered by previous generations of stationary bikes. This allows the company command $2245 plus a $58 monthly subscription from its base of 1.4M users.

Like Peloton, companies like Glossier, MongoDB, and our portfolio company Na-Kd (amongst many others) also rely heavily on their communities to enhance their product in many ways. Communities help companies because they can generate functional value such as content, product development, marketing, customer support, open-source code, as well as more subjective values such as a sense of belonging, purpose and social status. After winning them over, community is also a great retention mechanism for users. It gives users skin in the game, paving the way for the company to have more interactions with them and more ways to monetize them in the future. Rebecca Kaden from USV wrote a great piece on this, and Jeff Bussgang from Flybridge also wrote about how community can as act a moat. These positive effects of community are compounding and therefore hard to copy for a newcomer.

Web 3.0 protocols can power communities in a way that Web 2.0 companies cannot. I wrote about this before in context of protocol design. One of the reasons communities can be hard to scale is because they often rely on a set of obscure rules, which can be hard to define, express, and even harder to enforce. The community can be led astray by a few bad actors, and if it doesn’t course correct quickly, the product experience can be ruined forever. Videochat pioneer Chatroulette’s product was ruined by a small minority of “exhibitionist” users because they didn’t moderate fast enough. Communities face a version of the tragedy of the commons problem since they are a public good. Protocols make communities stronger and scale faster by codifying and enforcing rules to properly manage a common resource.

More specifically, Web 3.0 protocols can give participants skin in the game, align the community’s incentives, clearly define the group’s boundaries, create transparency of behavior, and provide for a means for dispute resolution.

Perhaps the most powerful example of protocols coordinating community is the Bitcoin community itself, which has also been described as cultish and even religious by some. Depending on the evolution of the network and the extent of adoption, the Bitcoin network can be said to manage the common resource of value storage or global money supply. Everything that a community member does to support Bitcoin directly benefits the entire community, defined as anyone who owns Bitcoin. The Bitcoin community, to say the least, has done a lot to support the protocol; they write code directly for the protocol — Bitcoin core has a comparable level of Github engagement as some of the biggest open source projects, sponsored by the biggest companies in the world. They run mining hardware, providing the network enough power to process 100K PH/s and consuming as much electricity as Australia. Some community members have also started companies evangelizing the adoption of the technology. The community of Bitcoiners have developed memes like “hodling” and “maximalism” to reinforce their culture, and let’s also not forget about the marketing:

Products with Finance

We are starting to see in Web 2.0 companies build in things like payments, working capital financing, and other more complex financial functions into their product. These financial services have been traditionally a separate, forgotten part of the product experience, imposing huge headaches for the user.

When you take a taxi for work, you need to first apply for a credit card, then swipe the card, and manually retrieve the receipt and take a picture and enter it into your company’s expense system. The driver, on the other hand, calculates the fee using the taxi meter, then types in the total amount into a credit card machine, swipes your payment, saves the other half of the paper receipt to account for this revenue, and then waits several days before the money processes through the taxi’s systems and finally to his bank account, with a substantial percentage taken out for processing fees. This fragmented user experience is a result of an incredibly complex payments processing stack:

When Uber created on-demand taxi rides, we often don’t think about seamless experience of stepping in and out of a taxi without ever having to think about how much to pay and how to pay, and that drivers get paid instantly into their bank account. However, this is one of the core puzzle pieces to ridesharing companies winning the market; payments has become one of the core aspects of the product.

Owning payments allows you to build a winning product. Your customers have different needs, pain points, and goals relating to the flow of money, such as tip pooling for restaurant staff (Toast), paying down business loans with daily card sales (Square), tying payments to rewards (Starbucks), streamlining collections (DocuSign), automating accounting reconciliation (Airbase), and paying out drivers instantly (Uber). Owning payments allows you to tailor the flow of money to solve your customers’ specific needs. In a world where the best product wins, mastering the payments experience is a strategic imperative. — Pat Grady, Sequoia

Our portfolio company Klarna was one of the first to help ecommerce companies build point-of-sale financing and try-before-you-buy directly into the checkout experience. This has helped the company win over 80 million shoppers globally and become one of the biggest tech unicorns in Europe. Not only does this help the customer to skip the hassle of having to find a loan provider based on some form of personal collateral in order to make expenditures match their paychecks, it also ensures that the merchant receives the cash upfront, risk free, while also making more money.

Faire, a wholesale marketplace for small boutiques, offers short-term working capital financing for its customers in the form of risk-free inventory. This is one of the key value props that has helped Faire grow from a YC upstart to a tech unicorn in just 2 years. Due to its 60-day payment terms, merchants can see if the product will sell through before coming up with the upfront capital to pay for it, and if it doesn’t, they can return the merchandise for free. This sounds simple, but coming up with the working capital for huge wholesale orders is one of the biggest hurdles for small shops whose job it is to carry a large and dynamic variety of merchandise. They would have to go to a traditional bank, who usually can’t underwrite small merchant risk, so the shop owner probably has to secure the loan somehow with their own assets. Ecommerce platforms, like Faire, have a better grasp on the risks associated with their customers and sell-through ability because they have all the data. They also have platform power over their suppliers and can force them to share some of this risk, so in many ways, they may be better equipped to provide credit for their customers and offer this as a core part of the product. Similarly, Amazon and Shopify have also been experimenting with working capital financing for its merchants.

However, to provide financial services within your Web 2.0 product, you need deep pockets or to work with banks with deep pockets, which introduces a lot of friction into the process.

Web 3.0 protocols have finance built-in, and they can build on top of other protocols to provide more complex services, drawing from native pools of capital earning yield natively from each protocol.

There are technical implementations for all types of payment services outside of the typical on-chain transaction, from micropayments, which allow video streaming companies to pay for each unit of video transcoded, to state channels that allow two people to agree on a value exchange that only gets executed if something happens, and paid subscriptions that you can trade while sharing a percentage of the revenue with the issuer— all without a central party to take any risk or extract any value from the transaction.

Outside of payments, Web 3.0 protocols can support collateralized loans, algorithmic exchanges, working capital financing, each protocol with its own natively-supplied pools of capital, and these capital pools are getting deep. In its first year, MakerDAO issued roughly $200Min in collateralized crypto-to-USD loans, and it took Lending Club 5 years to issue $250M; Maker uses this to maintain a stablecoin so that other apps can build dollar on-ramps for their users. dYdX and Compound have issued ~$270M in 50K collateralized crypto-to-crypto loans the past 30 days; they provide the backbone for margin trading and yield-earning fintech apps.

To their users, many of these protocols will have business models that look similar to Web 2.0 APIs (as explained in this post by Jesse Walden from a16z). These protocols abstract away many of the complexities to building a product with finance and empowers smaller upstarts to offer complex financial services without having to raise a ton of capital. We are still in the early days of product innovation using these financial primitives, but one can imagine potentially a fan merch site that accept payment in the form of minutes watched of the a vlogger’s livestream, and that same ecommerce site paying for its web services as a percentage of its sales in realtime.

Products with (portable) Data

Web 2.0 companies are starting to both consume and produce more portable data stores that work across apps.

As companies transition to microservices architecture, APIs proliferated because breaking down monolithic applications into smaller components produces more data feeds and integration points. It also means more ways for companies to monetize their data streams and more ways for new companies to build on pre-existing data stores. For example, Plaid and our portfolio company Truelayer help banks share customers’ data with other apps with the users’ permission. Whereas previously, a fintech company would have had to build custom integrations into each bank their customers use, collect their data, and manage all of that sensitive data, now it can easily plug into a secure data feed and build value-adding products on top. We are seeing similar kinds of shift in healthcare data and employment data as well. Today there are more than 50,000 public APIs, and >60% of surveyed companies have invested in a public-facing API and surrounding infrastructure.

The movement has only been accelerated by enterprises trying to comply with data privacy regulation and shifting consumer preferences. Consumers need to be able to “disappear” from enterprise data stores and to give consent for their data to be used; companies like OneTrust, BigID and our portfolio company Sourcepoint are helping enterprises make consumer data more portable as well.

Simultaneously, companies have to build for a wide variety of end-user displays, resulting in a decoupling of backend data stores from front end development. Multitude of front end displays such as IOT devices, mobile, and web all need to access the same data sources on the backend to provide a unified customer experience. Tools like Firebase, Heroku, and Amplify help companies build pluggable backend data stores that speak via API to front end platforms. Companies like Gatsbyjs and Hugo help front end developers quickly build webpages that plug into these backends that can also be more performant and secure.

Just as the networking layer of software development was abstracted away after Web 1.0, the backend layer seems to be getting abstracted away as well with the rise of tools like cloud, containerization, serverless, and Jamstack.

While more data components empowers more developers to build customized, full-stack app, navigating tens of thousands of public and unstandardized data feeds can be daunting and expensive for a developer.

Web 3.0 protocols are built on open and interoperable data. Blockchain data itself is open, and projects like the Graph help developers query blockchains in a standard way. The Ceramic Network helps developers define data schemas (among other things) across data feeds to make them interoperable, and Arweave allows apps to permanently store data across a community-owned web. Our portfolio company 3Box also securely stores a user’s data across all of their apps. Using it, a developer can built a customized, dynamic app, where users bring their own data.

Outside of data services, Web 3.0 protocols can provide other pluggable backend functions like compute, content management, file storage, content distribution, location services, and authentication. Many of these are still early in terms of performance, but they should be considered as options by companies building backend abstractions for the current web. This post by Joel Monegro from Placeholder explains more in detail why.

The Beginning

Along with the scams and controversies, the capital injection into crypto in 2017 has fueled a lot of bottoms up building and tinkering of important Web 3.0 rails these past few years, which is often less well-publicized. It’s exciting to say that after years of closely observing these movements as well as investment themes across Web 2.0 that these two are beginning to converge (maybe we are already operating in Web 2.5).

Huge successful companies have already been built using community, finance, and data as competitive advantages. I am hopeful that we will see many more because Web 3.0 rails enable builders to build the same features with less capital.

However, these are still only the early signals of where the bridges could be built across the chasm of adoption, and there is still a lot of work to actually build them so that average users can cross. I am hopeful that the same forces that have fueled the earlier growth will only be accelerated in the next few years because the talent inflow into the space has not slowed down.

Bridges across the Chasm to Web 3.0 was originally published in Northzone on Medium, where people are continuing the conversation by highlighting and responding to this story.

A personal story

I have been lucky enough most of my life to have only had to marginally interact with the U.S. healthcare system — the occasional accident when I was young, but the rest of it is pretty routine. It wasn’t until I decided to have kids that I became a real “customer” in the system. I have two kids, one was born in one of the leading hospitals in Manhattan — the epitome of a private, high intervention health system, and the other was born in the NHS in London— the epitome of a public, low intervention health system. I had pretty serious complications in both instances, and luckily the babies and I are all fine in the end. The experience between my two births were polar opposites, yet both equally terrible. I didn’t feel like a customer at all, I felt like an object, having healthcare “done onto me” by many different parties, each with their own incentives and broken pieces of data, and no coherence overall. This next-level horrid customer experience, delivered by some of the best practitioners in the world, is what sent me diving down the healthcare rabbit hole.

The U.S. system was characterized by loads of highly invasive interventions, prescribed by experts doing what was best for me but often without a clear delineation of what would be my choice vs. what is necessary nor clearly explaining what the risks were:

Getting my “buy-in” was a burden to the doctors and nurses, and that was clear in the way they interacted with me.

After 37 weeks of a happy and healthy pregnancy, one of the doctors in my OBGYN rotation decided to begin a series of unnecessary interventions that escalated to some spectacularly bad outcomes. The first time I met her, she decided to dilate my cervix, without asking for my consent. I might have given consent but the big issues is that I didn’t even know if what she was doing was something I needed to give consent over or if it was a part of the standard practice. In addition to it being extremely painful, that day after I came home I began to feel very uncomfortable and then, my water broke, but my body was obviously not ready to go into labor yet. Then at the hospital, when my body wanted to take more time, the interventions escalated from hormonal induction to more manual dilation, nothing happened for 12 hours until my body basically went crazy and dilated from 10% to 100% in 10 minutes. For those of you unfamiliar with the process of labor, this is not normal. All of this lead to a massive hemorrhage that required me to get two blood transfusions and two plasma transfusions and might have ended in a hysterectomy if they couldn’t get the bleeding to stop. Luckily, none of this affected the baby directly and the bleeding eventually stopped after several more interventions. After it all, I remember distinctly hearing the doctor responsible say “thank god you delivered in a hospital, or we don’t know what would have happened”, somehow with no sense of irony.

Several years later, after I forgot all about the trauma, I was pregnant again with my second son. This time, for reasons no one could explain, I went into labor two months early while I was on a business trip to London. There, I got to learn all about the NHS.

The U.K system provided zero room for personalization, and the need to adhere to protocol greatly outweighed my need to understand the risks and the tradeoffs made on me and my baby.

I went into an NHS with some abnormal bleeding after 31 weeks of another happy and no-complication pregnancy, but I was sent home with nothing after they did a couple of routine exams. Then I went back with light, regular contractions and bleeding and was kept overnight with some acetaminophen and sent home again. The next day I came back with regular intense contractions and was told they couldn’t give me anything to slow my contractions until the baby was basically already coming (???). I delivered my second son at 31 weeks, and pretty shortly afterwards had another massive postpartum hemorrhage, but this time I wasn’t given a blood transfusion because I was below the NHS threshold for needing one, so I spent a couple of weeks feeling extremely weak and dizzy.

In both cases, I had no champion to gather all the pieces and fill in the gaps between the countless handoffs from care team to care team, and I wasn’t in the position to be able to do it myself. In both cases, it was hard to cleanly separate the positives vs. the negatives of each system as well as the causes vs. the effects of the intervention / non-interventions. I ultimately needed to trust in the expertise of my care teams, who I know was trying their best, but they are still constrained to and operating within a system that I know is so broken.

There was one microcosm where my experience was different, where the system felt like it was working and our incentives were aligned. My youngest was born prematurely, so he spent a few weeks in the NICU in the NHS. In there, there is so much data being constantly generated and observed, and as obsessive parent of a premature infant, I became a walking data repository for him. (I don’t recommend this level of data obsession for any parent by the way). The mean, median and standard deviation for his vitals, the names and schedule of all his medicines, when and how much he was feeding, the color and frequency of his bodily functions, and his every tiny developmental milestone — I had internalized it all. I joined the daily handoffs between the nurses and the doctor teams, and they began to rely on me for my input in gaging his progress.

It felt empowering to be the owner of all this data and understand exactly when and how to escalate anomalies; it also felt good to not freak out every time the heart rate/oxygen monitor beeped.

We addressed his acid reflux early on, dealt with his high blood-calcium levels, transitioned him off of oxygen very quickly, and when he learned how to feed independently, he was ready to be discharged. He only spent a record of 4 weeks in the NICU (which I’m told is very rare for a 31 week preemie). Overall, I felt that I helped significantly improve the standard of care for Dylan by proactively participating in the data stewardship, and the NICU specifically had a good system such that it allowed me to play that role. The same can’t be said about many other medical experiences. The NICU was this tiny little microcosm of data where I spent 8–12 hours per day talking to the doctors and nurses, this was the cost of maintaining all this data for my son. Still, disappointingly, at the end of his stay, all of the data was just thrown away, buried deeply within the vaults of the NHS in handwritten notes. We were cut off from the microcosm as soon as we were discharged, and my current pediatrician has only a slight glimpse into that entire journey, in the form of a discharge report, which took the NICU doctor hours to put together, that she maybe looked at once for 5 minutes. Other than that, the most important data repository for my baby was me.

What’s the problem

I always knew healthcare was so broken, but my experiences with childbirth made it that much more vivid. I spent some time trying to understand how my experiences applied to the broader topic of healthcare, and many of the issues are the same. If you’ve ever had any kind of health issue, you’ve probably been both tempted to and explicitly told not to google your symptoms. Today, there seems to be a conflict between a consumer’s role in her own health versus an expert’s.

Why does this conflict exist? Why is there such a thin line between “thoughtful consumer” and “hypochondriac” — digging deeper, it comes down to data and incentives.

Consumer-generated health data is lacking and poorly curated. Living with a health condition is an incredibly lonely experience. Although we know that there are probably thousands of people in the world who have experienced similar health issues as us, it’s still impossible to find them in our “connected” world. General social media channels are not the best forums for sharing health issues, and Google, though it has done wonders for general education, falls extremely short when it comes to health data. What exists currently are fragmented anecdotes from outdated forums posted by other desperate consumers, spread out across time and space, looking for answers but mostly getting emotional support at best and hypochondria at worst. (It was definitely hard to fight the urge to Google everything with my two kids). The curation algorithm that runs Web 2.0 fails miserably in highly specialized genres such as health.

More reliable publishers exist to offer basic guidance, but they are still stuck in the Web 1.0 format of 1-to-many broadcasts, offering user experiences reminiscent of Craigslist. There are some companies like The Mighty and Patients Like Me innovating in this space; I am very excited to see more.

In addition, good individual health data just frankly doesn’t exist in the U.S., digitally at least. Medical experts are only able to access thin slivers of your entire health data trail — in the form of limited paper medical records, labs/tests/scans done on you at that point in time, and what you can remember to communicate to them on the day of your visit. Most of the data management is analog and broken, so even when it does exists, it’s not where it needs to be when it needs to be there.

Based on only thin slices of data, only population-wide comparisons are possible, but we know that health is highly personal. It always seemed odd to me that some doctors advise against comprehensive testing like full-body scans because it triggers a lot of “false positives” (how can more data be bad???), but then it occurred to me that without an understanding of the linear progression of your own baseline data through time (longitudinal data), the results can be useless at best and falsely alarming at worst. Longitudinal health data is currently prohibitively expensive to build and impossible to store. Many hospitals now are adopting electronic health records (EHR); companies like Ciitizen, Particle, and ZocDoc are pioneering consumer-led approaches for data collection; and increasingly more people are using wearables and storing their data with Apple Health. However, when you think about how much data we have for systems as complex as our bodies and our mind vs. something like our browsing behavior or shopping activities, I’d say we are only at the very very beginning of the data collection exercise.

Last, in the practice of healthcare, the responsibility handoff isn’t clearly delineated between the consumer and the various healthcare providers. Consumers are not empowered to be proactive champions of their own care, in fact, sometimes there is a visible conflict. For example, our healthcare system is set up such that the doctor gets very little leverage, and most of their work is done with each patient in person. As the doctor is incentivized to see more and more patients, the visits themselves don’t change. Each interaction is composed of patient education, data collection, diagnosis, and treatment, and the time spent educating the patient often comes at the expense of the other tasks or the doctors ability to treat all her other patients on time. Some OBGYNs actually discourage the use of doulas whose sole job is taking the patient’s perspective because it makes the doctor’s job harder. The spread of medical misinformation thanks to Dr. Google probably makes this conflict even more visible.

Health data is complex, and much of the pattern matching should probably be managed by experts. However, there are roles within the data value chain that are best performed by consumers, who have the most incentive to do it right (and also happen to spend all their time with the data subject). Consumers can be stewards and aggregators their own health data. They can also be taught to recognize signals for escalation. They are not empowered to do that now, and this handoff is so broken. It’s mostly done in the format of a few vanilla health history questions that care providers spends a maximum of a few minutes interpreting at the beginning of a visit. Escalation happens only at severe malfunctions — this is why we have a “sickcare” system vs a healthcare system.

“Wellness” vs. “Sickcare”

I am not a proponent of self-diagnosis, and I think swearing off of the advances of modern medicine is harmful and stupid. But I also think some of these health conspiracies arising from distrust of the system are direct results of our lack of self-sovereignty in relation to our health. On top of the growing skepticism, the reactive attitude about our health forces us into being overly reliant on reactive interventions. The system trains us to be objects rather than empowered consumers.

Consumers in China have long been taking health into their own hands by relying on the principles of Chinese Traditional Medicine (CTM). I was born in China, and I grew up with my mom, aunts, and grandmas always nagging me to moderate the amount of inflammation in my body by avoiding specific foods and to increase my Chi by taking certain herbs. I always used to dismiss the advice as old wives tales.

Regardless of whether or not you believe in the methods of CTM, what I’ve come to realize is that the most important thing about CTM is probably not the specific curative properties of certain herbs or tonics but that it provides a framework for people to constantly observe their own body and a way to proactively manage their own health — It makes them tune in.

Within some healthcare categories, western consumers have already taken the proactive role in their care. Categories such as nutrition, dermatology, optometry, and orthodontics were the first to shift, and companies in these sectors have had a lot of success speaking directly to the consumer, with the help of medical professionals. I believe many more categories will start to shift in this direction, such as physical therapy, fertility, developmental pediatrics, chronic disease management, and mental health. Many startups are already building in this space (Kaia, Natural Cycles, Kindbody, Thirty Madison, Shine), and I’m very excited to see more. Overall, I’m super excited to see many more companies using technology to empower us as healthcare consumers while leveraging our medical professionals, and I will be spending time looking for investments in this space.

I’ll end with this: a pretty cool example of how nuanced the topic of nutrition has become when empowered consumers start to take health into their own hands:

Pratyush on Twitter

The replies to this are awesome. People are done with passively accepting the advice of these charlatans and taking their health into their own hands. https://t.co/BLjzjmKqcY

A shorter meditation on Meditations on Moloch by Slate Star Codex

A metaphor has been making waves in the crypto community: Moloch, a mythical demonic creature that gets people to do bad things as a collective though individual intentions may be good. For more on Moloch, go here: Meditations on Moloch by Slate Star Codex. (The rest of this blog post probably won’t make much sense to you if you haven’t read this post)

Moloch is a good metaphor for systemic failures — many of them classic economic problems (externalities, prisoner’s dilemma, etc.), others first introduced to me by Yuval Noah Harari (in his books Sapiens, Homo Deus, etc.). This is the first time I’ve reflected on them one big agency problem caused by a metaphoric demon — it’s a powerful and compelling abstraction.

Upon reflection, I realized that Moloch is against my core (and perhaps naively optimistic) belief that collective human systems result in overwhelmingly good things because we are more than just rational beings optimizing our own utility functions, following the rules of the system blindly.

Something as simple as why people give their seat to others on the subway gives me hope that this is true every day. There are societal norms that get enforced through trust and other incentives which compel individuals to be “good” and do “good”. Some of these factors have been explored by studies on altruism, social signaling, etc. They are harder to quantify as motivators, but still nonetheless exist and are engrained our society. I wrote a post about trust some time back that talks about its positive effects on markets.

“But not only have we not yet reached the sea, but we also seem to move uphill surprisingly often. Why do things not degenerate more and more until we are back at subsistence level? I can think of three bad reasons — excess resources, physical limitations, and utility maximization — plus one good reason — coordination.” — Meditations on Moloch by Slate Star Codex

The discussion around “racing to the bottom” is interesting, but I don’t think it’s these things per se which keep us from plummeting but also the perception of these things — e.g. perception of excess resources creates more cooperative rather than competitive behavior which leads to more excess resources (e.g. growing the pie); perception that pursuit of happiness is a fundamental right keeps us from enslaving ourselves and others; perception of coordination makes us at least try. We agree to value optimism as a society, this makes it a fulfilling prophecy in some way. It allows us to cooperate, even today, and potentially more in the future with the help of technology.

As I wrote about before, I think we are on the brink of being able to align peoples incentives within a network to the network as a whole (using technology) to control for the agency effect. This, in addition to the societal values, makes the feedback loop even stronger.

Last thought, I don’t like the idea that it must be done by some super intelligent machine — if we humans can’t figure it out, there is no way some machine we build can.

P.S. The title is potentially misleading, Moloch exists whether or not we think he does because these coordination problems absolutely do exist. The more accurate (but certainly less snappy) title should be: the more we believe we can contain Moloch, the easier he is to contain…

I discussed in our investment thesis that we believe much of the value of the web will migrate from enterprises to protocols. This is because: 1. Protocols can finally govern value transfer based on a set of business rules directly deployable on the open web 2. Digital assets can be scarce and 3. Protocols can be both more efficient and less rent-seeking at coordinating humans and machines.

What we couldn’t define at the time was how exactly the value erosion happens, and where the wedges would be created into the existing web businesses, dominated by giants.

One potential answer to this is where there are digital market failures in our society today.

In economic terms, a market failure is any time the free market fails to allocate resources efficiently, usually requiring government intervention. There are a few types of market failures, but I want to first focus specifically on the tragedy of the commons. The commons refer to a pool of scarce resources, shared by all, but not governable by any, so it’s often exploited and depleted. The only solution proposed by economic theory was privatization — letting an enterprise govern it — or nationalization — letting governments govern it.

This was until Elinor Ostrom proposed a solution which won her the Nobel Prize in Economics in 2009. She observed that many communities have successfully governed the commons provided they adhere to a set of rules. These rules are largely encoded into the fabric of trust built within the small communities she studied and never thought to be socially scalable to our large, modern societies… until crypto protocols came along.

For the first time, the rules for governing a common resource can be encoded and enforced digitally, within crypto protocols, without the need for an enterprise or a government. If we can account for the resource and enforce rules around its consumption digitally, Elinor’s findings can extend beyond the small communities where this already works in practice.

Let’s take a deeper look at these rules and how crypto protocols can enforce them:

Elinor Ostrom’s 8 Principles for Governing the Commons Applied to Protocol Governance Design:

1. Define clear group boundaries: a protocol’s tokens can grant access rights to the resource governed by network in the form of a membership token or a utility currency

2. Match rules governing use of common goods to local needs and conditions: crypto protocols provide a market for matching supply with demand

3. Ensure that those affected by the rules can participate in modifying the rules: tokens can be used to grant governance rights

4. Make sure the rule-making rights of community members are respected by outside authorities: protocol must integrate with existing legal/regulatory systems

5. Develop a system, carried out by community members, for monitoring members’ behavior: protocols require a reputation system which makes market participants’ actions transparent and obvious

6. Use graduated sanctions for rule violators: protocols must provide a way to punish bad actors for bad behavior (loss of staked tokens, price of the token decreasing, exclusion from network, etc.)

7. Provide accessible, low-cost means for dispute resolution: tokens can be used to ensure skin in the game when disputes arise or to incentivize game theoretic proofs, but this has to be simple enough for people to understand

8. Build responsibility for governing the common resource in nested tiers from the lowest level up to the entire interconnected system: protocols must be inter-operable and clearly delineated

Each of these principles probably deserves to be analyzed more in detail, which I will perhaps do in a future post. However, I think we are already seeing a trend of protocol design converge towards these principles, but it’s also clear that we still have a long way to go. Overall, it’s pretty amazing now neatly some of the uses we have for protocol tokens fit into her research.

The Digital Economy

Crypto protocols will have a harder time managing the physical commons because rules like access rights are harder to enforce in meat space without some sort of central party intervention (e.g. clean air, oceans, schools, roads, etc.). However, applied to the digital commons where all the value can now be traced and tracked for the first time, real world adoption begins to look more near term. However, the idea of a valuable digital commons is still somewhat new to us because digital scarcity is new.

What makes this really big is that any scarce resource can be thought of as the commons and managed by crypto protocols if producers and consumers are willing to produce into and consume from the “collective pot”. The problem with this line of thinking has always been the corruption of the incentives of the central party required to manage the allocations back to society.

Again, crypto protocols solve this problem.

They are actually far better at solving this problem than even the most uncorrupt central authority because protocols can deliver services at monopoly efficiency while maintaining perfectly competitive prices. This means that any time a protocol or an enterprise is competing to deliver a service, a protocol can deliver it more cheaply by making it a commodity.

The usefulness of this economic model extends thus beyond the commons and into other types of market failures such as monopolies or oligopolies, which sounds more like where we are today.

Real World Examples

So far, the crypto protocols we know have not been deployed explicitly to manage a common resource, but there is an argument to be made that that is exactly what they do. Take Bitcoin for example, depending on the evolution of the network and the extent of adoption, the Bitcoin network can be said to manage the common resource of value storage or global money supply. Ethereum, depending on scaling efforts, can be said to manage anything from digital token standards to global contracting or computing resources. ZeppelinOS aims to manage a smart contract operating system as a common resource; OScoin aims to manage open source code development; Filecoin cloud storage, and Livepeer verifiable computing / video broadcasting infrastructure.

A Multiplayer World

Time will tell the extent of adoption and how many of our resources we’ll be able to manage as commons. However, as I wrote about in my last post, aligning our incentives with the collective using crypto protocols is powerful because it is how our society works today, but it isn’t how our businesses work. Our incentive structures fail in the face of increasing human connectivity because they fail to capture the growing impact of our actions on our networks and vice versa. In the future, our decisions’ impact on others will be perhaps greater than on ourselves, so continuing to operate in single player mode in a multiplayer world is not only wrong but it is dangerous.

Thanks to many who contributed to this post including the incredible and late Elinor Ostrom; for more on her work, visit here. Thanks to Par Jorgen Parson and Albert Wenger for encourage me to hit publish, OScoin for being the first project I know to explocitly build based on these principles.

We are still early down the path to understanding the value of crypto-networks. In fact, this is one of the main arguments for institutional capital other than early-stage funds to be sitting on the sidelines today. Early stage investors benefit from the mandate to invest into the first stages of value creation based on criteria such as market size, entrepreneurial talent, and value proposition rather than cash flows, EBITDA and other financial metrics, criteria I like to think of as being more related to assessing the opportunity rather than measuring the progress. Thus, it makes sense that we are among the first to invest into this movement.

Where is the Utility Value?

However, the underlying substance that all long-term investors need to be able to invest into, irrespective of stage, is the creation of utility value — defined in economics as the total satisfaction received from consuming a good or service. Economic utility directly determines the demand, therefore the price, of a good or service; this fundamentally underlies any valuation model looking to describe a crypto-network. While there are many great efforts underway to assess token value from a macro or cost-based perspectives, the intention of this post is to propose a new way to model the basic unit of utility value in a token, an expansion of the very basic building blocks of microeconomic theory. I’d like to propose a key difference in how an individual fundamentally models her own utility:

Whereas our current models of valuation rely on an individual to maximize only her own utility, irrespective of others, valuing crypto-network requires her to maximize her own utility in context of the collective utility, which she needs to have a perspective on.

A Quick Refresher on Microeconomic Demand Derivation

An individual’s willingness to pay for a good or service is determined by her utility curves, which characterizes all the possible trade-offs between consuming various baskets of goods at given levels of satisfaction. An individual’s utility curves can be mapped against a budget constraint to derive individual demand curves for specific goods. These demand curves are then aggregated to arrive at the market demand curve which illustrates the quantity of a good or service the market is willing to buy at different prices. The intersection of this curve with a market supply curve is how you arrive at the price of a good. The individual is then a price-taker and either transacts at the market price or does not based on her own demand.

At the highest level, an individual only has to know her own preferences in order to trade in an economy for goods or services based on a market price that is determined by everyone’s aggregated preferences and supply. The market price is set regardless of whether or not the individual understands everyone else’s preferences. She can simply choose to buy or not buy based on the price set by the market. This applies to all participants in the market. It’s important to understand this to understand how our currently valuation models are flawed in considering networks.

Emerging Crypto-economic Building Blocks and Early Signs of Value

Now let’s switch gears for a minute and look at a few emerging crypto-economic building blocks and the logic for valuing each:

Meme coins/store of value tokens/NFTs: While the equation of exchange has been proposed as a useful and early model for thinking about these types of tokens, it does not discuss utility value. The individual’s utility, thus willingness to pay, for each of these is only slightly determined by her satisfaction of owning the token, especially today when most NFTs lack full-featured environments where you can actually use the asset. Willingness to pay is largely determined by her expectations of what the token will be worth in the future, as a storage of value, which is determined by her expectations of others’ willingness to pay over time. Assets that have been good stores of value in the past have been traditionally backed by something which anchors the collective belief around value storage such as scarcity (gold), history (art/antiques), trusted governments (currency), monopolies (diamonds), etc. Bitcoin was so fundamental to crypto-networks because it established, for the first time, that the collective belief around value storage can be anchored by technology and our collective digital memory.

Prediction Markets: These have been cited as the only “new” use case to come out of crypto, uniquely enabled by the technology. While we are still in the early days of experimenting with prediction markets, I, along with many other investors, are hugely bullish on them as an important economic building block for web 3.0. The entire premise of prediction markets is relying on collective belief to price probabilities around specific events. Here, again, an individual’s willingness to pay is determined by her expectations about the collective belief in relation to reality.

Curation Markets / Bonding Curves: Many of the recently popular top games in Ethereum rely on these dynamics, and in the simplest form, they can be classified as gambling. An individual has to play based on her beliefs about the others who are participating. However, as easy as it is to dismiss these types of games’ importance due to the gambling classification, they can also be studied as complex game-theoretic experiments. Continuing to rapidly generate data on how we as individuals optimize outcomes based on our expectations about the collective belief is key to understanding complex economic problems.

You’ve probably noticed the reoccurring theme between these three very early prototypes of crypto-economic building blocks:

In crypto-economics, the individual does not determine her willingness to pay based on her preferences alone; she needs to be plugged into a wider “collective” whose preferences she has to take into consideration when determining price.

Difference from Speculation — from a Black Box to the Beginnings of a Model

It’s important to acknowledge the conflation between the idea of speculation — pricing without understanding — and the idea of pricing based on expected collective value. This misunderstanding could be somewhat driven by the individualistic philosophies dominating our consciousness today. It is also driven, in large part by our limited ability to model thus fully understand collective utility in economics currently.

One of the main problems in modeling collective value has been that while the collective influences our utility, our definition of the “collective” is within our control — presenting a catch-22 problem. Also, the definition of the collective varies widely depending on the context, ranging from our family unit to all of humanity. Crypto-networks help solve this problem because they codify ones relationship with the collective into a protocol we can refer back to; they also concretely define the collective as holders of that token. By defining a constraint around the “collective”, they can be useful tools in economic model construction. In addition, they also provide real-time data from repeated, real-world experiments.

What Does This Mean for Society?

The most important implication of defining individual value as a function of the collective value is this — if we are all maximizing our own utility as a function of the collective utility, we work towards increasing the satisfaction of the collective as well as ourselves, because the two are fundamentally linked.

Intuitively, this concept is actually not that foreign. Our enjoyment of many things in life relies on the enjoyment of some sort of collective. Why do we spend so much on designer goods (and actually enjoy them)? Why do we enjoy eating at buzzy, scene restaurants rather than empty ones? (Well, sometimes at least.) Why won’t Ethereum miners collude to censor transactions or Bitcoin miners perform a 51% attack? Because the individual’s utility function is largely dependent on the collective — the expensive designer shoe is more appealing when you know a large group of fashion aficionados would covet them; the buzzy restaurant is more enjoyable because other scene New Yorkers around you are also enjoying themselves; Bitcoin and Ethereum are safe from those who have the most power to harm it because their own wealth is tied to how other network participants value the network.

By tying individual value to the collective, cryptonetworks may help us solve some of the most fundamental problems in economics such as the tragedy of the commons. I would go as far as to say in our increasingly connected world, where very few can afford to live in silos, we can no longer just look at individual value irrespective of the collective value, not because it is immoral but simply because it is inaccurate.

Existing Economic Literature

For those who are more familiar with the study of economics, I propose that this is different than economic studies of altruism, where one maximizes her own utility BY maximizing the collective utility because making others happy makes her happy. A buyer of bitcoin might slightly increase the price of bitcoin by placing a large order, which benefits all bitcoin investors, but it is still for personal gain that she does this, and the increase in the collective utility is a by-product.

It also goes beyond social influence, where information about others changes an individual’s opinions, beliefs, actions (though the designer shoe and restaurant examples might be able to be described by social influence theory). Under the collective model of utility, the individual can be positively or adversely affect by collective utility shifts whether or not she has information from others (e.g. huge price increase/drop in bitcoin).

This post is really meant to open the topic up for discussion from this lens. I really look forward to the perspectives from other thinkers, economists, sociologists, and psychologists!

Thank you so much to Matt Stephenson for helping me think through this in context of behavioral economic research and pointing me to a lot of the academic literature I’m referencing in this post.

The Evolution of the Web

The web grows when more people interact in more ways online, and this is a function of trust and technology working in tandem. The adoption of internet protocols such as TCP/IP provided a trusted basis for transmitting information, which powered the formation of the web. However, these protocols didn’t define much else. Web 1.0 was an anonymous place, where you could trust that content will get from point A to B, but there was very little trust in the content itself. The only business transactions were 1-to-many, for the access of static information, from a few large trusted content producers.

There was very little economic activity in the “read-only web” due to commoditization of information, but it gave us the information age. Starting in Web 2.0, institutions built tech platforms to create trustworthy environments online, forming new markets where many-to-many business transactions could take place. This is what we know as the “read-write web”. Participants were able to interact online for the first time because they trusted the platforms to dictate and enforce the rules of engagement. As a result, these Web 2.0 intermediaries took a large cut of the value and defended their positions using network effects and data monopolies.

In the current era of the web, the beginning of Web 3.0, technology standards are being developed to establish the trusted basis for transmitting value as well as information. These protocols also allow the creation of automatic dependencies between value and information transfer. By extension, they create trusted environments online where participants could interact without the help of an institution. In many circumstances, these protocols will out-compete Web 2.0 platforms because they extract less value from the ecosystem. They will also erode institutions over time by significantly decreasing the cost of coordinating economic activity.

Protocols Powered by Tokens

Tokens are the powerful incentivizing force which allow protocols to coordinate trusted economic activity on the web. They help bootstrap a network by first providing liquidity for future network value. This then attracts early adopters to contribute specific work, data, and capital for the network (in exchange for tokens). Last, as the supply side becomes more robust, the demand-side participants can buy and pay (via token) for the network’s services.

Tokens also create bigger markets by adding financial incentives to the non-financial incentives powering most of the web today. Whereas previously, someone contributes to open source, social media, and search for user utility or social rewards only, crypto protocols can also reward its contributors with financial incentives on top.

Finally, we are just beginning to see how tokens can create powerful economic building blocks for Web 3.0. Examples of these are curation markets, bonding curves, NFTs, and staking mechanisms. These will enable new types of markets to be formed, the likes of which we can’t even imagine yet. Trying to do so would be as fruitful as trying to imagine Google’s current business model when the only use case for the internet was email.

Protocols vs. Institutions

If you reduce the role of an institution or a protocol down to the very core, the aim is to get a bunch of people and machines to work together towards a common goal, and usually that’s the production of some sort of good or service. It then stands to reason that whichever entity can do this cheaper will have an advantage over the other and grow, while the other shrinks (adapted version of Coase’s Theorum). Well-designed tokens allow crypto protocols to coordinate people and machines more cheaply than institutions by cutting out middlemen, simplifying contracts, and reducing political friction.

We haven’t yet seen the line drawn between where companies end and protocols begin, but bolstered by tokens, protocols have a lot of room to grow and can be expected to take significant share from current companies. One hypothesis is that the future of the web could be composed of a few widely-used crypto protocols responsible for much of the transactional, low-value tasks in addition to thousands of highly specialized institutions focusing on user experiences.

Which Protocols Will Win?

Protocols must compete with each other for participants, and the differentiation in the long run will most likely not be technical since much of the tech is open source. One possible area of differentiation between protocols is, again, efficiency in coordination. Because networks must coordinate behavior between a complex group of individuals, threading the adoption string between entrepreneurs, developers, early adopters, and mass adopters can be a real challenge. Each has different motivations, preferences, and behaviors, and each group strongly dictate how incentives are structured for the following groups. The protocol is a living, constantly evolving thing, so both understanding and adapting to its various participants’ requirements throughout its evolution is a competitive advantage. In the most successful protocols, the earlier adopters will help design an incentive mechanism and experience that highly appeals to the later-comers, even though their own incentives differ.

Usually, a network is launched with a specific type of intended network effect, but in order for the network to grow exponentially, the meaning of those network effects needs to be expanded over time to become more appealing to those outside the network, creating a bigger and bigger incentive for them to join. In the the early days of Facebook, the draw to join was basically to scope out other single Harvard students. However, as more people joined, it became a way to communicate with all of your college friends, then a way to share photos with your friends and family, and now a way to consume all types of media content, curated by your social graph. In Web 2.0, the direction and scope of these compounding network effects were shaped by a central firm, but in a protocol, it will be shaped by the participants themselves, according to the order they enter the network.

This process of decentralized decision-making is often referred to as governance. I provide here a few high-level principles we look for in protocol governance.

• The chance for success for either probably depends on what the network advertises hence, the kind of participant it attracts
• People within a network value governance differently, so the system has to account for that or else we will always have suboptimal decision-making
• To avoid a plutocracy, we probably need an identity solution
• The most successful system probably involves several different systems each with its own, conflicting incentives constantly grinding against the others

Where is the Investment Opportunity?

After the founding of Bitcoin, it took a several years for a critical mass of people to get mentally comfortable with the idea of money existing as a digital recollection between its participants. Then a flood of private investments in early 2014 funded the ability to more easily custody and trade this value via exchanges and wallets, providing the on and off ramps into the ecosystem. Next, Ethereum familiarized the idea of the internet of agreements and for the first time enabled deals to be made and executed across a distributed network. ICOs brought more private capital to the ecosystem, but more importantly demonstrated the power of tokens for bootstrapping a decentralized community and marketplace. Then, non-fungible (unique) tokens cemented the idea of digital scarcity into cryptoassets. Now, we are seeing cryptoeconomics primitives such as curation markets being built and incorporated into the toolkit for building the future decentralized business models.

Many of the key infrastructural layers have been established, but currently the gap between these protocols and end users is massive. Other than speculation, there is very little that users, and even to some extent, developers can do on the most robust protocols. Yet, due to network effects and community momentum, it’s likely that many of these base layers, or interoperable replacements, will become the fundamental building blocks of Web 3.0 and that solutions will be built on top to make them more usable. We believe there is a significant opportunity to invest in these layers in between. These layers provide valuable services like identity, off-chain oracles, curation, decentralized exchange, as well as developer tools and interoperability. Most of this layer of infrastructure is being developed right now, with some positioned to become as dominant as a Web 3.0 operating system and perhaps even more valuable than the underlying protocols themselves.

We will also be following the user adoption movement closely as these connecting layers become more robust, monitoring for the first signs of adoption. We have already observed interesting momentum in the gaming sector where many of the crypto-economic primitives can be experimented upon rapidly. In addition to existing layer 1 infrastructures, we believe there are opportunities in function-specific protocol layers that will become more important in Web 3.0 such as file storage, CDN, transcoding, location verification, etc. to provide cheaper, more decentralized alternatives to existing incumbent solutions.

Where Will Value Accrue?

We look at the future protocol stack in many inter-operable layers, similar to value chains, and protocol layers can be even more distinct and numerous compared to firms. Some of these layers will become commoditized, reinterpreted, and their positions may also shift in relation to one another. How much value accrues to the protocol depends on the specific market dynamics — such as competitive advantage, centralization of the surrounding layers — as well as token dynamics around each layer. At the highest level, a protocol accruing value is not too different than a firm monetizing a business in that it relies on the user to pay a fee, based on her utility, for the network’s services. A part of the fee will be distributed directly to the supply side of the protocol, and a part of the fee will be paid by the user incurring a cost to hold the protocol token (due to price volatility and complexity). The total price will be determined by the user’s utility for the network’s services in conjunction with her utility for the token itself.

Our Approach

We are a nimble team with a transparent thesis and process. We hope to be the first partners along the crypto entrepreneurs’ journey and be committed to the success of the project long-term. We will work with our teams to build and refine their crypto-economic models as well as contribute actively to the network and community in the early days. We will also continue to monitor and invest into the seed ecosystems developing on top of the protocols we back. We will leverage our 20+ years of venture experience to help teams think through go to market and product design.

If you’re interested in learning more about our investment process, please read more about our thinking here.

Our Crypto Investment Thesis was originally published in Northzone on Medium, where people are continuing the conversation by highlighting and responding to this story.

Venture capital is also an industry where traditionally you had to learn via a long, humbling apprenticeship. So for the young, ambitious VC, the idea of carving out an ecosystem where none of the old rules apply is extremely enticing. However, before throwing all caution to the wind and completely unshackling from traditional VC wisdoms, I chant a daily mantra that the reality of what’s about to unfold in the short run lies probably somewhere in between conservative thinking and the most aspirational, and in the long run, it will probably surpass them both. Applying quantum thinking to all of this, the conflict zone in between these two versions of reality and the alternate realities that arise from that conflict is probably where I will spend most of my time as a venture investor in this era of tech.

“History doesn’t repeat itself, but it often rhymes” — Mark Twain.

Through reflecting on the hundreds of projects I’ve observed in the last two years (still a limited dataset compared to the universe of startups) and speaking to much wiser investors who have spent decades in VC, I’m starting to think about diligence for crypto investments from two different facets.

Facet One: The Art and the Science of Analyzing the Project itself

Why are they doing this? Everything about the team’s vision, incentives, dynamics, skills, and values are all incredibly important to the success of a protocol. I spend most of my time trying to understand this aspect of each project I work with. This question is important because of both the skills component of protocol/founder fit and the fact that there are many decisions in the design of the network that may mean a direct trade-off between the short term profit of the founding team and the longevity of the network. This ranges from things as obvious as pump-and-dump schemes to how to capitalize the network and distribute tokens. There aren’t enough robust governance mechanisms and best practices in this early ecosystem, but as these develop over time as they have in traditional tech startups, this aspect of diligence becomes more formalized into the funding process. In the meantime, I believe that founders who are more committed to the 10-year outcome of the network than their own gain/egos are more likely to make the right decisions along the way.

Can this team lead a community of thousands or tens of thousands from behind the scenes? Can they do this by understanding and tapping into other people’s own motivations? Incentives structure between a company and a protocol are vastly different, and just as companies can claim their organizations as competitive differentiation, the community around a protocol is perhaps as important as the technology itself. Whereas entrepreneurs scaling equity organizations need to be able to scale pyramidally, ruling by dictatorship, protocol entrepreneurs need to lead by inspiration, influence, and incentive design, often taking themselves out of the story line in the long run. Perhaps that means Jim Collin’s concept of Level 5 Leadership is more relevant in this paradigm.

Is the market potential for this project 100X larger than what I would consider sufficient for a traditional venture business? There is significantly more risk in an early stage investment in this sector because the adoption story is still being written and the many regulatory risks of an early ecosystem. To quantify the risk of investing, we typically try to uncover all of the known unknowns and then add another layer of risk to account for the unknown unknowns. In enterprise software for example, this layer is pretty thin, in crypto, this layer is much thicker because much more experimentation is needed in the sector before we can start to understand them. Therefore, to compensate for this added risk, the market potential for a protocol has to be much larger.

Important to note that many of these protocols are building where incumbent players can’t compete, directly commoditizing their profit centers. This is an advantage in general for the space because some may argue that GAFA poses a risk to traditional startups in this era almost as significant as the risk faced by the crypto space as a whole. Another note on market size is that in crypto protocols, we see valuation in terms of network value (market cap if at scale) instead of enterprise value, which means we’re looking at potential value of monetary supply as well as demand for the goods and services provided by the protocol rather than demand for goods and services minus cost of providing those services. This is an argument for crypto networks to be valued much higher than enterprises in this era and why we are investing in the former rather than the latter.

What is the wedge that leads to real adoption? Projects have to be the right mix of skeuomorphic and game-changing in order to create an entry point into existing user behavior to ultimately change it dramatically. Some crypto concepts are so foreign to the average person that not understanding the product and value prop is going to pose the biggest first hurdle to adoption. Also, usability, most would agree, is still very poor in this current generation of products. This includes everything from custodianship, key management, crypto-to-crypto exchanges, etc.

Some may argue we are still in the infrastructure building phase and not ready for prime time with the end user, but it’s important to keep in mind that there are many groups of users who will need different things (I go into this later). I sometimes see projects anchor around value propositions that are idealistic rather than useful. This might appeal to the first few user groups (early adopters, some developers, etc.), but won’t answer to the needs of the rest. I expect the first iteration of products to get adopted will have to look and feel at least somewhat similar to web 2.0 and address an existing user need for the immediate next group of adopters. However, the vision for the protocol still has to be big enough that it can anticipate the web 3.0 stack and evolve alongside. Finding the right balance of this is one of the things I struggle with the most in evaluating crypto protocols. I will go more into depth on this in Part Two of this post.

How does the thread of adoption string together all the different groups of adopters? Since crypto protocols must coordinate behavior between a complex group of individuals, I look for reasons to believe that there will be a strong thread of adoption between the various groups — from entrepreneurs, developers, early adopters, to mass and late adopters. Each has different motivations, preferences, and behaviors, and each shapes the incentives and user experience for the following. The protocol is a living, constantly evolving thing, so both understanding and adapting to its various participants’ requirements throughout its evolution is a competitive advantage. In the most successful protocols, the earlier adopters will design and continue to shape an incentive structure and user experience that appeals to the later-comers, even though their own may differ.

How will the protocol achieve compounding network effects? This concept is related to the last point. Usually, a network is launched with a specific type of intended network effect, but in order for the network to grow exponentially, the meaning of those network effects needs to be expanded over time to become more appealing to those outside the network, creating a bigger and bigger incentive for them to join. In the the early days of Facebook, the draw to join was basically to scope out other single Harvard students. However, as more people joined, it became a way to communicate with all of your college friends, then a way to share photos with your friends and family, and now a way to consume all types of media content, curated by your social graph. In Web 2.0, the direction and scope of these compounding network effects were shaped by a central firm, but in a protocol, it will be shaped by the participants themselves, according to the order they enter the network.

How dynamic is the protocol, and how dynamic does it need to be? Lastly, it appears that the protocols that have developed thus far have only been able to adapt by evolution, by consensus upgrading and forking. This is a slower process than the traditional tech “pivot” and potentially leaves the protocol vulnerable to competitors with even slightly better tech and a faster-growing community. It will be interesting to compare the pace of an established protocol’s evolution vs. the pace of community adoption for a new, competing protocol as this will be the ultimate determinant of who wins the race. This is the argument for having the right kind of governance model for a protocol and will largely depend on the intended use for the network. There are many more elements of governance design we could get into in diligence, as well as how the team plans to test and learn from real people during the process. However, the higher level question I always ask is how it fits the community and the network’s intended purpose.

We are still in the early phases of protocol evolution, but here are three early hypotheses about what balances the equation between governance and function: 1. A really homogeneous user base, to minimize conflict of incentives. Example could be local / regional networks or those allied around a common ideology. or 2. Really minimalist design, so basis for disagreement is minimized. Example would be single use-case protocol around something highly function-specific. or 3. A highly adaptive protocol that is constantly renewing its community of highly engaged participants.

There are several more important elements of deal analysis such as token economics, product/community fit, development pace, etc. I won’t have time to get into any of these in detail here, but they are all also very key parts of analyzing the project itself.

Facet Two: Understanding Yourself and Your Own Biases

What do I think? What do I think about what I think? A huge part of successful early stage investing is really knowing yourself, when to trust yourself, and where you are prone to bias. After I’ve processed all the answers I could dig up from the analysis part of diligence from above, I always ask myself, “Why do I now like or dislike this investment opportunity; which biases may have led me there, and how do I correct for them?” Ours is an industry that relies heavily on heuristics, but we often leave those heuristics unchecked over long periods of time. They are an interesting place to start and can get you 90% of the way there. This is really helpful for prioritization and gives the experienced investor a significant head start compared to an inexperienced one. However, I think what separates top quartile investor from the average investor is her ability to examine and refine her own heuristics actively over time. Just as a single strategy doesn’t serve a company well across tech paradigms, stagnant heuristics won’t either. Also, I believe our unexamined biases are probably the most dangerous motivators because the human tendency to rationalize is so powerful.

After getting to know someone well enough, you usually can quickly spot their biases, and when you get external references, you probably take the info you receive with a healthy dose of salt. It’s unreasonable to expect that a single person possesses the absolute perspective on reality (or that this really exists at all). The more objective truth is most likely formed by a collection of perceptions that are all relative to each other, and these perceptions are completely shaped by our brains’ flawed shortcuts. In important decision-making, I try to treat myself like one of those external references and step outside of my shoes and take a bird’s eye view on the collection of everyone’s perceptions including my own, flavored each with its own grain of salt. Even better when the collection is from all different angles of a problem and they vastly disagree with one another.

Am I thinking exponentially? This goes back to my earlier question around finding the wedge. Once I am convinced a project can get early market traction and I’ve developed a level of comfort with the known unknown risks, I have to answer to my beliefs about the unknown unknowns. The unknown unknowns are produced by the multiplication of risks, which many VCs want to avoid, but it is also the underlying key formula for exponential growth. It is difficult for humans to think exponentially due to how our brains have evolved. The min and max ranges of specific outcomes are so wide apart that it is intellectually impossible to model. However, the reality of tech progress in our era is that it is almost always exponential.

This extremely oversimplified chart demonstrates a collection of perspectives about the future vs. reality.

Lacking better information, taking an average of all of these expectations seems like the natural thing to do. However, in its linear form, this only works for speculative investing with short term horizons because you only need to match other investors expectations about the market and the state of adoption. For example, in hype markets, the average expectations slope will be much higher than the adoption curve but it doesn’t really matter, and if you invest along those lines, you will probably get a good short term return because other short term traders will validate you. The same goes for bear markets. These speculative signals can crowd out utility signals, which makes the long term investors job harder.

Long term investors investing based on utility value have to invest along several adoption hypotheses, constantly read the signals of utility, and iterate and reevaluate. Such utility signals can be the flow of talent into the sector, development momentum and quality, consumer interest and understanding, regulatory direction, and institutional adoption. I am constantly and obsessively looking out for these. Long term investors also have to ask questions like “Do I think the entrepreneurial team can iterate on their adoption hypothesis using real world signals” rather than “Can I accurately predict the future”.

Conclusions

Any investor spending a significant amount of time in crypto right now is making a personal investment of time and reputation in addition to the financial capital they deploy into the sector. Many do so with the expectations of potential exponential outcomes. Many also believe that we are at the beginnings of a whole new investment paradigm, which is why it’s good to remember no one is really an expert yet, and it’s best look for those who are thoughtful instead. We may find that wisdom we draw from traditional venture may only be slightly applicable, but I think it’s a good place to start. I have to conclude by thanking those who have helped me think through these starting points, especially Par Jorgen Parson, who has been a great mentor and taught me most of what I know about investing.