My First Million Wiki

Network Effects

8 min read

Network Effects

Five or six of the seven most valuable companies in the world have network effects at their core. That was not true twenty years ago. It was not true thirty years ago. Something fundamental changed.

James Currier, the founder of NFX and perhaps the most systematic thinker on this topic, has spent decades studying this shift. His fund manages 110 million to Monster. When he talks about networks, he is not theorizing. He is reporting from the field.

The basic definition is deceptively simple: a network effect exists when each new user makes a product more valuable for existing users. Twitter is the canonical example. More people tweeting makes Twitter more useful for everyone already there. But this simplicity masks enormous complexity. NFX has catalogued seventeen distinct mathematical types of network effects. Most people never get past understanding the first one.

Why Network Effects Matter More Than They Used To

The shift happened because software ate the world. In the physical economy, scale was the dominant advantage. Walmart could buy cheaper because it bought more. Amazon could ship cheaper because it shipped more. These advantages were real but incremental.

Digital products work differently. When Microsoft dominated operating systems, it was not because Windows was better. It was because more Windows users meant more developers building for Windows, which meant more software, which attracted more users. The loop compounds. Word Perfect learned this the hard way.

This explains why profitable businesses often plateau while seemingly unprofitable ones explode. Shaan Puri described a conversation where the distinction became clear: many founders build profitable companies but never understand the “shape” of a billion-dollar business. That shape almost always includes network effects or deep defensibility. Without it, you have a good business. With it, you might have a great one.

Watch: Investing Wisdom from Nassim Taleb - Discussion of billion-dollar company “shape”

The Four Defensibilities

Currier argues that in the digital world, there are only four ways to build lasting protection against competitors. He ranks them:

1. Network Effects - The most powerful. Each new user increases value for all users. Facebook, Instagram, WhatsApp, Twitter. The compounding is almost impossible to overcome once established.

2. Embedding - Deep integration into customer operations that makes switching painful. Oracle and Salesforce customers do not leave because leaving would require rebuilding years of workflows. This is strong but static.

3. Brand - Emotional and status associations that create loyalty beyond product features. Nike. Louis Vuitton. Ford. Brand defensibility is real but takes decades to build and can evaporate surprisingly fast.

4. Scale - Size advantages in buying power, distribution, or production. This is the weakest defensibility. Currier uses MrBeast as an example: more revenue funds bigger productions which generate more views which generate more revenue. It sounds powerful. But this is the same loop BuzzFeed had. Scale effects without network effects are vulnerable.

Watch: James Currier Full Interview - Four Defensibilities framework

The distinction matters because people conflate these constantly. Having lots of customers is not the same as having network effects. Being big is not the same as being defensible. Many large companies are one innovation away from irrelevance precisely because their size created complacency, not compounding value.

Your Life on Network Effects

Here is where Currier’s thinking takes an unexpected turn. His most popular essay applies network thinking not to startups but to personal decisions. The framework is startling in its simplicity once you see it.

Every major life decision is actually a decision about which network to join or leave. College is not primarily about education. It is about joining an alumni network. A city is not just a place to live. It is a network with career and lifestyle implications. Marriage is not just about two people. It is about merging two family networks. You will have Christmas with her parents forever.

Currier puts it directly: “Don’t think of yourself as choosing a job or choosing an industry - think of yourself as choosing a network. My company is a network. Who I hire into my network. Which journalists I get to write about me, I bond that person into my network. Which investors I bond into my network.”

Watch: Your Life on Network Effects - Personal application

The framework extends further than most people are comfortable with:

  • Language - Speaking English versus Mandarin is not a skill difference. It is a network difference. Each language opens and closes different opportunity networks.
  • Religion - A church is a community network with specific values, relationships, and access patterns.
  • Citizenship - A country is a network with particular obligations and access rights.

The Tax Migration Mistake

Sam Parr recounted a conversation with Currier that stuck with him. When someone mentions they moved to avoid state taxes, Currier thinks they are making a mistake. The math seems obvious: save thirteen percent on taxes by leaving California for Texas or Florida. The actual math is different.

San Francisco is a network. The investors are there. The founders are there. The information flow happens in coffee shops and dinners. When you leave that network to save thirteen percent, you might make thirteen times less money on your next opportunity because you were not in the room when it came together. The tax savings are visible. The missed opportunities are invisible.

This is not an argument that everyone should live in San Francisco. It is an argument that people dramatically underestimate the economic value of being embedded in the right network. The calculation is not “cost of living in expensive city versus cheaper city.” The calculation is “network I am joining versus network I am leaving.”

Watch: Sam on Tax Migration - Network effects of city choice

Network Effects and AI

Currier makes a prediction about artificial intelligence that runs counter to most venture capital thinking. He believes OpenAI and Nvidia are making bad bets. His reasoning: AI and large language models will become free like water. The technology itself will commoditize to zero.

The only AI companies that will survive are those that build platform network effects into their offerings. Nvidia has this partially through CUDA - developers who learn CUDA are less likely to switch, and more CUDA developers means more software optimized for Nvidia chips. But the underlying AI technology? That is racing toward commodity status.

The implication for entrepreneurs is significant. Building an AI company is not enough. Building an AI company with network effects - where users create value for other users - might be.

Building Network Effects: Practical Notes

Dharmesh Shah of HubSpot made an observation about community that illustrates how network effects work in practice. Once a community reaches critical mass, it becomes very difficult to disrupt. The reason is circular: people stay where other people already are. The content improves because more people contribute. The connections become more valuable because more people participate.

This is why community businesses like Skool have inherent advantages. Members of one community often start their own, creating an organic growth loop that looks like network effects even when the product itself is relatively simple.

Watch: Dharmesh Shah on Community Network Effects

The vampire attack is the inverse strategy: draining value from an existing network to build your own. LIV Golf attempted this by poaching PGA stars. The theory is that certain nodes in a network carry disproportionate value. Remove those nodes and the network weakens. Attract those nodes and your network strengthens. Tiger Woods reportedly adds $1.5 billion per year to PGA viewership when he plays. One person. One and a half billion dollars.

The Quiet Insight

The deeper point in all of this is uncomfortable. Most of us think of our decisions as individual choices. We choose a job. We choose a city. We choose a spouse. We choose where to invest our time.

Currier suggests we are actually choosing networks. And networks have their own logic, their own compounding effects, their own gravity. The person who joins a great network and contributes to it will likely outperform the person who makes better individual decisions in a weaker network.

This is not determinism. You can choose your networks. But the choice of network might matter more than the choices you make within it. That is a different way of thinking about almost everything.

Sources

Graph View

Backlinks