Network Effects: The Moat That Turns Startups Into Giants

Key Takeaways

The 70% Rule: According to NFX, network effects are responsible for 70% of all value created by tech companies since 1994, yet only 35% of $1B+ companies actually had them.
Unicorn Correlation: According to Morgan Stanley research, more than 50% of unicorns relied on network effects in their business models at peak.
Not Virality: Network effects are about defensibility and retention, not user acquisition. Confusing the two is one of the most common VC mistakes.
The Cold Start Trap: Every network-effects business faces a bootstrapping crisis before it gets powerful. How a startup solves it predicts whether the moat ever materializes.
16 Types Exist: NFX has identified 16 distinct types of network effects. Most founders only think about one. The strongest companies layer several simultaneously.

Most investors think they understand network effects. Most of them are wrong.

They see rapid user growth and call it a network effect. They confuse virality with defensibility. They back platforms that look like Airbnb but have the structural moat of a restaurant directory. Then they wonder why the market didn't tip the way they expected.

Here's the real story: network effects are responsible for approximately 70% of the value created by tech companies since the Internet became a thing in 1994.

That's the finding from NFX, the venture firm built entirely around this thesis. The number is staggering. And the implication is even starker: if you're evaluating startups without a firm grasp of network effects, you're flying blind on the single most important variable.

A network effect is the phenomenon where each additional user makes a product more valuable for every existing user. Simple enough. But the execution and the evaluation are anything but.

Why Network Effects Beat Every Other Moat

Brand is expensive to build and easy to erode. IP gets challenged. Economies of scale get competed away by a better-capitalized rival.

Network effects remain the most durable form of competitive moat in technology-enabled markets, particularly where value scales with participant counts, data depth, and ecosystem richness. In platforms, every new user or transaction can amplify value for all participants, creating a self-reinforcing loop that can push incumbents into winner-takes-most dynamics.

That "winner-takes-most" phrase matters.

The underlying principles of network effects imply that the business, website, or platform with the highest market share will be more successful in the long run. Its market share is likely to grow more substantially. Markets in which network effects play a major role are often referred to as winner-takes-all markets.

By looking at each of the $1B+ companies' business models and comparing them to known network effects, an estimated 35% of those companies had network effects at their core. They were, however, typically much more valuable than companies without network effects, so they added up to 68% of the total value.

That's the asymmetry. A minority of companies capturing the majority of value.

Companies relying on other defensibilities like embedding, scale, and brand typically top out in the $1-$2B range.

Companies with strong network effects are the ones that become $100B outliers.

For VCs chasing the power law, this isn't a nice-to-have. It's the whole game.

Direct vs. Indirect: The Two Architectures That Matter Most

Not all network effects are created equal. Understanding the difference is the first filter any serious investor needs.

Direct network effects occur when more users directly add value to all other users.

A real-life example of this type of effect would be social networks and messaging apps like WhatsApp, where the service becomes more valuable as more contacts join the network.

The value is symmetric and immediate. Every new node connects to every existing node.

Indirect network effects are more subtle and often more powerful.

An indirect network effect occurs when the value of the network increases for one type of user as another type of user joins, who then increases the network's value further, attracting more users and creating a flywheel.

Think of the App Store: more iPhone users attract more developers, more developers attract more users, repeat indefinitely.

The math behind this is described by Metcalfe's Law.

Metcalfe's Law states that the value of a telecommunications network is proportional to the square of the number of connected users. In simple terms, the more users a network has, the more valuable it becomes.

Double the users, quadruple the value. That's the curve every VC is hunting for.

But here's the honest version: the fundamental flaw underlying Metcalfe's Law is in the assignment of equal value to all connections or all groups.

Not every new user creates equal value. A marketplace for enterprise software buyers adds vastly more value with user 1,000 than user 1,000,000 does. The law is a mental model, not a formula.

The 5 Questions Every VC Should Ask Before Calling It a Moat

Most investors call it a network effect when they mean "the product is social" or "it has more users than last year." That's not a moat. Here's a tighter framework.

1. Does each additional user make the product more valuable for existing users?

This is the only question that actually defines a network effect. Not "does the product grow?" Not "do users share it?" Does user N+1 create tangible value for users 1 through N? If the answer is no, it's not a network effect. It's a distribution story.

2. Who is the "hard side" of the network?

A common challenge within the cold start problem is how to acquire and retain the hard side of networks. The hard side is the side of the network that is either less incentivised to participate or the side that has more options.

For Uber it was drivers in new markets. For Airbnb it was hosts. For OpenTable it was restaurants. The startup that can crack the hard side first builds a moat. The one that can't stalls before the flywheel spins.

3. Has the startup reached its atomic network?

Andrew Chen, General Partner at Andreessen Horowitz and former head of rider growth at Uber, introduced this concept in his book The Cold Start Problem.

The atomic network is the smallest possible network that is stable and can grow on its own — where there are enough people that everyone will stick around. For Zoom's videoconferencing network, this can work with just two people, whereas Airbnb's requires hundreds of active rental listings in a market to become stable.

If a startup hasn't hit its atomic network, the whole thesis is theoretical. Ask: "Where is this working today, completely, in a single geography or segment?" If there's no clean answer, there's no moat yet.

4. Are switching costs amplifying the network?

Once a platform crosses a critical mass threshold, retention and engagement often become self-reinforcing through habit formation, network-driven product improvements, and the costliness of moving to an alternative ecosystem.

The best moats stack network effects on top of switching costs. Slack is the canonical example: its value increases with every new integration and every message history archived, making leaving progressively more painful even if a competitor offers parity features.

5. Is it a real network effect or just virality?

This is the most common misread.

Network effects are not viral effects. Network effects are about creating defensibility, and viral effects are about getting new users for free. They have totally different objectives and playbooks.

A product that spreads fast but doesn't become more valuable for each user isn't building a moat. It's building a mailing list. Angry Birds spread fast. It had no network effect.

Angry Birds grew quickly in popularity but didn't necessarily increase in value with each new user.

The Types VCs Miss Most Often

NFX has identified 16 different kinds of network effects, listed in order of strength, starting with Physical (e.g. landline telephones), Protocol (e.g. Ethernet), Personal Utility (e.g. iMessage, WhatsApp), and Personal (e.g. Facebook).

Most founders and investors focus on the personal and marketplace varieties. Here are two they systematically underweight:

Data Network Effects. Your product collects data with each interaction, that data improves the product, and the improvement attracts more users who generate more data. This is Google Search's actual moat.

Google's search engine platform is one of the best examples of a durable moat created by network effects, as far more accurate search results are provided because of more user data collection.

It's also the moat being built by every AI company that has proprietary usage data.

Protocol Network Effects. When a startup becomes the standard that an entire ecosystem builds around, it captures value from every participant in that ecosystem without ever competing with them directly. Ethereum, Stripe's payment rails, and Twilio all have elements of this. It's rare, but when you find it early, it's the most durable moat that exists.

What This Means If You're Screening Deals

In 2021, more than 50% of unicorns relied on network effects in their business models.

That's not a coincidence. It's selection pressure in action: the startups most likely to escape gravity are the ones where growth makes the product better, not just bigger.

When you're evaluating a startup with a potential network-effects moat, a research-backed tool like Unicorn Screener can help you systematically score the dimensions that matter: traction velocity, competitive defensibility, and whether the market structure favors a winner-takes-most outcome. The highest-scoring companies on our live leaderboard tend to show at least one compounding network dynamic embedded in their core product.

But no score replaces the five questions above. Quantitative signals confirm what qualitative analysis should surface first.

One honest caveat: network effects change more quickly than ever. Instead of winner-take-all markets where early movers may have once had a lasting advantage, we're seeing all kinds of network effects companies splitting markets among multiple players.

A moat built on network effects is powerful, but it isn't permanent. Governance, product quality, and strategic timing all still matter.

The Cold Start Is the Real Filter

Here's the counterintuitive truth.

Most new networks fail. If a new video-sharing app launches and doesn't have a wide selection of content early on, users won't stick around. The same is true for marketplaces, social networks, and all the other variations of consumer and B2B products — if users don't find who or what they want, they'll churn.

The cold start problem is where most network-effects startups die. They have the architecture right. They have the market right. They can't get the first dense cluster of users to create enough value to retain the second wave.

How a founding team solves their cold start is one of the best signals of execution quality you'll ever get. Tinder seeded college campuses party by party. LinkedIn front-loaded its network with Silicon Valley executives before opening broadly. Reddit's founders famously created fake accounts to simulate activity. Each hack was different. Each revealed founder creativity under pressure.

Having a network effect is the single most predictable attribute of the highest-value technology companies. And yet surprisingly, only 20 percent of the business plans reviewed before the founding of NFX had network effects in them. Most founders fail to design network effects into their businesses because they don't understand them well enough.

If founders aren't designing for network effects from day one, the moat almost never materializes after the fact. It's not a feature you bolt on at Series B. It's an architectural decision made at the whiteboard stage.

What This Means for You

Screen for the flywheel first. Ask whether user N+1 creates value for users 1 through N. If not, it's not a network effect, regardless of growth rate.
Probe the cold start solution. How the team bootstrapped their first dense network cluster tells you more about execution quality than any deck slide.
Look for stacked moats. The best companies layer network effects on top of switching costs and data advantages. Each layer compounds the others.
Distinguish winner-takes-most markets from fragmented ones. Strong direct network effects create winner-takes-most dynamics. Weak or indirect ones often result in multi-player markets. Know which one you're underwriting.
Score the defensibility, not just the growth. Try Unicorn Screener to systematically evaluate a startup's moat depth before writing the check.

For a deeper look at what separates companies that reach unicorn status from those that stall, see what makes a startup a unicorn and the 7 product-market fit indicators that VCs actually track. Network effects and PMF aren't the same thing, but the strongest companies have both.

Want to screen startups like a top-tier VC? Score any startup for free with our research-backed evaluation model.