Thesis Driven Investing
What this covers
The conversation is between Patrick and Rebecca Kaden of Union Square Ventures. It starts with how thesis-driven investing works in practice. Kaden says that when a fast market overwhelms an investor, the remedy is to stop looking at companies for a couple of weeks and think about what they believe, even at the cost of missing some deals. She also describes USV's earlier thesis on horizontal networks. Each new one offered a smaller opportunity because network effects produced dominant platforms. That led to the decentralization thesis, which treats blockchain as the first real chance to decentralize the internet. She calls this contested, and the claim is a forecast.
The middle of the discussion is sector-specific. Kaden argues that venture has mostly worked on goods that got cheaper while education and healthcare went unaddressed. She says the gap between college curricula and job requirements is widening, citing Google dropping college requirements and General Assembly's enterprise training. USV's education approach is outside-in, building consumer demand rather than selling to school districts. In healthcare she points to B2B tools such as Ribbon Health's real-time referral network, and in finance to a fee model that structurally excludes middle-income Americans.
The last stretch is about distribution. Kaden calls it harder than product and says a founder needs a plausible plan beyond Facebook ads, ideally a hook built into the product. She notes that Facebook customer acquisition costs have risen. She separates networks from community, using Dia&Co's Facebook group as the example. She also predicts that the best digital marketers will be the fastest channel testers. The episode ends with Bill Campbell as a model of mentorship: honest feedback and real follow-through.
Rebecca Kaden argues USV's thesis-driven investing works because it aggregates the partners' shared beliefs into a strategy, and that technology should target the expensive categories of education, healthcare, and financial access by making them better and cheaper, with distribution and community as the hard problems.
- Thesis-driven venture means building a strategy from partners' shared beliefs rather than from a top-down market scan
- Education and healthcare costs have risen exponentially while tech-driven goods got cheaper, so technology should target them
- Distribution is harder than product, and products with built-in sharing or community mechanics solve it better
USV's investing aggregates shared beliefs into evolving theses, and venture success requires a clear firm-level thesis.
- USV's first thesis in 2003 and 2004 held that equity value would accumulate where large networks of engaged users formed, and that platforms would grow from those networks; this produced investments in Twitter, Indeed, and Etsy.
“the belief was that equity value would amass towards where there were large networks of engaged users”
- USV is a thesis-driven firm, meaning rather than treating venture as a coverage game of how many companies exist and how many it can see, the partnership aggregates shared beliefs into a thesis that shapes its investing strategy.
“rather than think of venture as a cover-driven model, of how many companies are out there and how many we can see”
- Thesis 3.0 seeks to broaden access by using platforms and protocols to systematically drive up value and drive down cost around major customer spending buckets, focused on three categories: access to knowledge, access to capital, and access to wellbeing.
“How can you systematically drive up value and down cost around major buckets of customer spend?”
- Thesis 2.0 asked whether the same network dynamics appear in vertical networks and in the underlying infrastructure for building them, leading to investments in Twilio, MongoDB, Stripe, and Cloudflare.
“could you see those same tendencies in vertical networks, as well as the underlying infrastructure to build and create those networks”
- Infrastructure companies like Plaid and Stripe succeeded by supporting an emerging ecosystem of startups early, and their infrastructure scales with those customers, a model USV believes can work in digital health.
“companies like Plaid, or Stripe, started by this idea of supporting an emerging ecosystem of companies”
- There are many ways to be a good venture investor, but the best firms have a strategy they stick to and a point of view, and success requires commitment to one's own ideas in a risky asset class.
“The ones I admire most are the ones that have a strategy and stick to it, and have a point of view.”
- When overwhelmed by a fast market, an investor should stop looking at companies for a couple of weeks and instead think about their own beliefs, because focusing on what they believe leads to the right investments despite missing some deals.
“Don't look at companies for a couple weeks, just think about your ideas, think about what you believe in”
- USV publishes its theses after the ideas are already guiding its investing, because the thesis forms organically through conversations among partners and the firm believes publishing ideas helps everyone, including itself.
“it's as much a reflection of the ideas that we've already been investing in for some time as it is about where we're going”
- Thesis 2.0's decentralization element arose because centralized platforms controlled users' data and activity, which posed risks to consumers and stifled innovation, and blockchain, first applied through crypto, is seen as the technology with a real shot at decentralizing the internet.
“blockchain is the first technology that really has a shot at decentralizing the internet, crypto being probably the first application”
- Telemedicine companies that need to refer patients to in-person doctors rely on static databases that quickly go out of date, so real-time referral networks like Ribbon Health are needed.
“any information that's static, we know, is not very high quality at all”
Education is underserved by venture, and sectors like it are open to technology-delivered consumer alternatives.
- USV only invests in things it believes technology can make better, defining better as higher quality at less money.
“we define better by better quality and less money”
- Venture capital and technology have almost exclusively focused on the bottom half of the price graph, the goods that got cheaper, leaving education and healthcare, the categories that became exponentially more expensive, largely unaddressed.
“venture capital and technology has almost exclusively focused on the bottom half of the graph”
- How Americans learn has not been structurally changed in roughly 50 years, teacher salaries are flat, and public school success and appreciation rates are flat, even though technology has transformed nearly every other part of life.
“how we learn structurally in America has basically not been touched for 50 years, maybe more”
- Over recent decades prices for technology-driven goods like TVs, toys, and software have fallen systematically, housing has stayed roughly mid-range, while education and healthcare have exponentially outpaced inflation and become far more expensive.
“education and healthcare are, by far, the outliers and they exponentially shoot up”
- The gap between education and job readiness is bigger than ever, and if the education system keeps failing to get people to their goals, it will eventually break, although institutionalized systems are very hard to break.
“eventually, the system's got to break”
- There is a large business opportunity in the education-to-workforce gap because companies, students, parents, and even universities are all willing to pay to close it, despite complexity such as misaligned incentives between school and student.
“Companies are willing to pay for that gap. Students are willing to pay for that gap. Parents are often willing to pay”
- The gap between education and job readiness is widening because college curricula reflect what was learned decades ago while job requirements have changed, and rising student debt makes the cost of that education harder to justify, pushing people toward alternatives.
“what you learn is effectively what you learn decades ago”
- Trusted brands like Google can compete with Harvard, Stanford, and Princeton for education credentials, because traditional education is built on long-term brands and consumers may increasingly trust new brands as technology penetrates their lives.
“you could see traditional education getting a run for its money by new brands that enter the space”
- A healthcare company built on Zoom delivers addiction recovery and medication-assisted treatment remotely, improving adherence and access by removing the need to drive to inaccessible clinics, which mirrors USV's education thesis of improving the product through its delivery mode.
“that saves them driving to clinics that are really inaccessible, leaving their jobs to be able to go there”
- The banking system is built on an AUM-based fee model that charges fees to the two ends of the spectrum, low-balance customers and very high net worth clients, so most middle-income Americans are structurally excluded.
“the banking system as a whole was built on an AUM model that really... favors fees off of the two ends of the spectrum”
- Financial services are hard to understand and use, built around a fear-driven design, and have not been productized, creating an opening for products that are more convenient, cheaper, easier to understand, and better experienced.
“it's built kind of with a fear module when you think about the emotion of the bank”
- Access to education, finance, and water are not surprising unlocks; they have repeatedly been the core drivers of economic movement and success in history, but technology has not focused on them.
“These things aren't surprising. They've been repeated throughout history as the core unlocks of economic movement and success”
- Consumer healthcare innovations are built on legacy rails and lack speed, convenience, and value, so the increasing opportunity is B2B software that supports them, such as referral networks like Ribbon Health that solve the problem of static provider databases.
“any information that's static, we know, is not very high quality at all”
- Consumer health businesses like Nurx and Modern Fertility give customers more control, autonomy, and access to their information at a fraction of the cost of traditional healthcare, letting people buy tests and treatments privately on an app.
“Modern Fertility, you can buy it yourself, you can have that information upfront and it's a fraction of the cost”
- Wellness and healthcare are increasingly blending because as healthcare becomes more consumerized, customers think about how they feel and how to feel better rather than whether something is healthcare or wellness.
“a customer doesn't think, is this healthcare or is this wellness? They think of how do I feel?”
Online learning approached from outside the system can change what learning means and make education better.
- USV's education strategy is outside-in: it targets products that create mass customer demand and behavior that penetrates the system from outside, rather than selling through school districts, school boards, or traditional schools.
“we're not focused on products and services that are selling through school districts or traditional schooling or school boards”
- Learning online is not only more convenient and broader in reach but can make the education itself better, because the modality makes it easier to learn in some ways.
“if you learn it online, you actually learn it better”
- The internet has taken offline education and put it online, which is valuable but uncertain in quality; the more important opportunity is to change what learning means rather than just move offline learning onto the web.
“what I'm interested in is how can the internet work to change what learning means”
- Duolingo's free language learning on a phone works better than classroom learning because it integrates into daily life, is bite-sized, is fun, and can be habit-forming in ways that are hard to replicate in a classroom.
“it integrates into your life, it's bite-sized, it's fun, it can be a little bit addicting”
- Zoom made synchronous video cheap and easy to build on, lowering the cost structure and making synchronous learning a why-now for the education category.
“Zoom made it very easy to build on top of it, and it's also way more affordable than it was”
- Outschool is a live, synchronous K-12 learning marketplace where teachers post classes, enabling children across the country and world to learn with teachers and peers outside their districts, including unconventional subjects like Spanish through Taylor Swift songs.
“teachers post classes, it's a marketplace”
Investing around personal interests and focus areas filters out hot-deal noise and market overwhelm.
- Following your own interests and passions is a good enough reason to invest, and a fund designed around specific focus areas lets partners decline hot deals outside their interests, which helps them block noise and have fun.
“following your interests and passions in investing is a good enough reason or as good as any”
- When Kaden, a first-year business student and former journalist, said she wanted to join Google or Facebook, Bill Campbell gave honest feedback that she should consider venture capital and Maveron, and he personally called her before and after the Maveron interview.
“I think you should think about venture capital... go do Maveron”
- Series A rounds that used to be 5 to 7 million dollars were coming in at around 20 million dollars, which USV does not typically do, adding to market overwhelm.
“series A's that maybe used to be 5 to 7, were coming in and were saying they're raising 20”
Consumer fintech distribution is easier than expected because customer intent and product-embedded hooks drive adoption.
- In investing and fintech, distribution is harder to solve than product, because product is something a startup can take many shots at, while distribution is very competitive, but customer acquisition has nonetheless been easier because customers with intent are looking to improve their finances.
“distribution is very challenging. And so... acquisition has not been that hard here”
- Stash built its consumer investing product around a $5 minimum and interest-based investing, organizing investments around what people believe in rather than specific stock picks, which led 80 to 85 percent of its customers to invest for the first time.
“so, 80%, 85% of the customers on Stash have never invested before”
- A good sign of distribution potential is a hook inside the product itself, such as a natural sharing moment, rather than relying on paid ads, word of mouth, or hoping for organic scale.
“is there a hook in the product itself that lends itself to growth?”
- Investing is complicated, and no one answers the basic questions of what to invest in, how much, and what to expect unless you have a lot of money, so productizing those simple answers creates huge value for lots of Americans who should be investing.
“those answers aren't that complicated. But if you've never heard them before, how would you know?”
- Customers today have intent to improve their financial health and no longer see local banks or distrusted big platforms as their only options, so this intent makes it the best time to acquire them.
“you have people with intent, and that's always the best time to acquire them”
- Facebook aggregated eyeballs that businesses could pick off cheaply for a long time, but the platform has grown more efficient at maximizing its own margin, so customer acquisition costs have risen and engagement is fragmenting.
“Facebook's job is to maximize its own margin, and so the customers have gotten more expensive”
Digital marketing success depends on rapidly testing and iterating across many channels and creatives.
- Dia&Co's clothing boxes naturally lead women to post their try-ons to friends to decide what to keep, creating a built-in sharing rhythm that doubles as a distribution channel within the core product experience.
“a woman gets a box of the clothes and they post it for their friends to see”
- Facebook is a giant that is not going anywhere, and core areas continue to grow strongly despite market criticism.
“I don't think Facebook is going anywhere at all”
- Digital marketing is shifting from optimizing one platform to quickly iterating and testing across many channels, including offline and new ones, so the best digital marketers over the next decade will be the fastest testers who spot and lean into new channels.
“the best digital marketers of the next 10 years are going to look like the fastest testers”
- The faster a company changes creative and copy on Facebook, the better the ads perform, so products that lend themselves to rapid creative iteration tend to work better in digital marketing.
“the faster you change creative and copy on Facebook, the better it works”
- Platforms like Roblox and Twitch can be acquisition channels for brands, but they are fragmented, so the arbitrage window opens and closes quickly, requiring marketers to lean in, test, and pull back as the opportunity moves.
“the arbitrage moment probably happens faster. So you're going to have to lean into them, see what works. Go fast there.”
Network effects centralized horizontal platforms, so blockchain decentralization matters for the internet.
- The opportunity in horizontal networks got smaller with each new one built, because network effects are very strong and produced centralized platforms that were very hard to compete with.
“the opportunity got less and less with every horizontal network built because these network effects are very strong”
- A horizontal platform like Facebook has a core network effect that is the business itself, while a vertical platform like Duolingo is a specific-industry business such as education or finance where the network effect strengthens the product but the network is not the product.
“horizontal is something like Facebook, where the core network effect is the business. Vertical is something like a financial services platform or an education platform”
- Thesis 2.0's decentralization element held that concentrating users' data and activity in a few platforms is dangerous for consumers and stifles innovation, because dominant horizontal platforms with monopolistic tendencies are very hard to compete with.
“all of your activity online, all of your data, were controlled by a very few number of players”
- USV's decentralization thesis holds that blockchain is the first technology with a real chance to decentralize the internet, with crypto probably the first application of blockchain.
“the belief is that blockchain is the first technology that really has a shot at decentralizing the internet”
Durable companies grow by building community and unifying brand through digital marketing.
- Disney is a masterful marketer because it combines product marketing, brand marketing, and digital marketing, and that integration is the goal for modern marketers.
“It is a product marketer as well as a brand marketer as well as a digital marketer. That's the goal.”
- Brand marketing requires consistency while digital platform marketing requires speed, and the two must share the same voice, so skilled marketers must reconcile these opposite tendencies.
“Brand marketing, I think that's true, you need a lot of consistency. Digital marketing on platforms... you need a lot of speed”
- A sharp transactional product like Stash or Modern Fertility can become a long-term valuable company by building community around it, turning a transactional brand into a sticky, trusted brand.
“to become a long-term valuable company, they're going to build a community on the back of it”
- A product can create a community that the customer did not previously have, such as Dia&Co's Facebook group where women who felt unprioritized build confidence and eventually plan vacations together, reversing the usual model of selling a product into an existing community.
“the reason women like it is because it introduces them to this really valuable community that they have lacked in their lives”
Training and hiring non-traditional talent for non-engineering careers is an underserved opportunity.
- Google and other companies have relaxed hiring rules by dropping preferred-college lists and college graduation requirements, and they run internal training programs, while General Assembly sells enterprise programs that prepare non-ready applicants inside companies.
“They took out their kind of preferred colleges list. They took out college graduation as a requirement.”
- Technology-enabled job training is a large opportunity outside tech, for example for the American nursing shortage, because those careers are valuable and underserved even though they do not look like engineering careers.
“the shortage in nursing in America. A whole bunch of categories that don't look like people who want to be engineers”
Technology creates connectivity without belonging, which drives loneliness and undermines wellbeing.
- Community and belonging may be the core of wellbeing and loneliness is an epidemic, yet technology has created connectivity without cracking belonging or community.
“there may be nothing more core to our wellbeing than community and belonging. Loneliness is an epidemic.”
- Networks are top-of-funnel growth mechanisms whose value comes from scale and structural growth, while real engagement and belonging come from how important the community becomes to a person and how it changes their interactions.
“networks maybe top of funnel, right? ... But real engagement, I think that's belonging.”
Worthwhile mentorship comes from honest feedback, deep thought, and followed-through commitments.
- The kindest acts Kaden has received were from mentors and peers who showed interest in her career, thought carefully about what would help her, and followed through, which she models for others.
“that level of interest in helping others' careers is something that I've been shown a lot and tried to show a lot”
- Lasting mentor relationships often begin when the mentee has more to receive than to give, and they can be fundamental even though the word mentor feels formal.
“the relationships at least started where I had more to take than to give have been really fundamental for me”
- A community of women venture investors who navigate careers, families, and co-investing together has been one of the biggest catalysts of Kaden's venture career, providing both practical and personal guidance.
“how to be really thoughtful as both friends and mentors and co-investors has been one of the biggest catalysts”
- Bill Campbell modeled a standard of mentorship in which he gave honest feedback, thought deeply about what would help the mentee, and followed through on commitments, rather than making a perfunctory email introduction.
“he was honest, he was thoughtful, and he followed through. And I think about the follow through a lot.”