The Most Overused Phrase in Startup Pitches

"We have network effects." It's in countless pitch decks, because network effects are the most powerful moat in tech. When a product gets better for every user as more users join, the leader pulls further ahead and competitors struggle to catch up.

But most products that claim network effects don't have them. They have growth, or brand, or scale. This essay explains what network effects really are, the math behind them, and a simple test for spotting the real thing.

🎯 Quick Answer (30-Second Read)

  • A network effect exists when each new user makes the product more valuable for existing users.
  • Metcalfe's law says a network's value grows with the number of possible connections, roughly n². Real networks usually grow slower than that.
  • Types: direct (WhatsApp), two-sided (Uber, Swiggy), data (Google Maps traffic), and local (food delivery per city).
  • Not network effects: economies of scale, viral growth, and brand. They help, but they're different.
  • Test: if a user's experience would be the same with 10x fewer other users, there's no network effect.

The Math: Why Networks Get Powerful

With n users, the number of possible pairs who can connect is:

connections = n × (n − 1) ÷ 2

Users Possible connections
2 1
10 45
100 4,950
1,000 499,500
xychart-beta title "Users vs possible connections" x-axis ["10", "20", "40", "60", "80", "100"] y-axis "Connections" 0 --> 5000 line [45, 190, 780, 1770, 3160, 4950]

Users grow 10x (10 → 100), connections grow 110x. That's Metcalfe's law: value grows roughly with n².

Reality check: you don't talk to every WhatsApp user. You talk to a few dozen people. Each additional user adds value only if they're someone you care about. That's why real network value grows more slowly than n², and why networks often win city by city or community by community.

Four Kinds of Network Effects

1. Direct
More users of the same kind make it better. WhatsApp is useful because your contacts are on it.
2. Two-sided
More users on one side attract the other. More riders bring more drivers, which means shorter waits, which brings more riders.
3. Data
Usage generates data that improves the product. More phones on Google Maps means better live traffic.
4. Local
The effect only works within an area. Swiggy having more restaurants in Bengaluru doesn't help a user in Pune.

The Two-Sided Flywheel

flowchart LR R["More riders"] --> D["More drivers
join (more earnings)"] D --> W["Shorter waits,
better coverage"] W --> R W --> P["Lower prices
possible"] P --> R

The catch is the start: with no riders, drivers won't join, and with no drivers, riders won't stay. This is the chicken-and-egg problem, and solving it usually means subsidising one side, starting in one small market, or supplying one side yourself at first. (More in Two-Sided Marketplace Lessons.)

What People Mistake for Network Effects

Often called a network effect What it actually is Why it's different
"We're growing fast because users invite friends" Virality Users spreading the product doesn't mean the product gets better for them
"We're bigger, so our costs are lower" Economies of scale Helps margins and prices, not each user's experience
"Everyone knows us" Brand Valuable, but doesn't improve the product per user
"Users have uploaded lots of data" Switching costs Locks users in, but doesn't benefit other users

All four are useful advantages. But they're weaker than a true network effect, and competitors can more easily copy them.

Why Network Effects Fail

Even real network effects can break:

  • Multi-homing: riders keep both Uber and Ola, drivers drive for both, and restaurants list on both Swiggy and Zomato. If switching is free, the network protects less.
  • Saturation: after a point, more restaurants on an app don't improve your experience much.
  • Quality decline: more users can mean more spam and noise (negative network effects).
  • Local fragmentation: winning Delhi doesn't guarantee winning Chennai.

A Simple Test

  • Does each new user make the product better for existing users, not just bigger?
  • Would a user's experience noticeably degrade if 90% of other users left?
  • Is it hard to use two competitors at once?
  • Does the effect keep growing, or does it flatten after a certain size?

If the first two answers are "no," it's not a network effect, whatever the pitch deck says.

Frequently Asked Questions

What is an example of a network effect?

WhatsApp is a classic direct network effect: it becomes more useful as more of your contacts join. Uber and Swiggy have two-sided network effects, where more customers attract more drivers or restaurants and vice versa. Google Maps has a data network effect from live traffic.

What is Metcalfe's law?

Metcalfe's law states that a network's value is proportional to the square of its number of users, because the number of possible connections grows roughly as n squared. In practice, users only interact with a small part of the network, so real value usually grows more slowly.

Is virality a network effect?

No. Virality describes how a product spreads, with users bringing in other users. A network effect describes how the product's value increases as more people use it. A product can be viral without getting better as it grows, and vice versa.

Conclusion

Network effects are powerful because value compounds with every user, and the math of connections explains why leaders pull ahead. But they're rarer than pitch decks suggest. Many products have virality, scale, brand or switching costs instead. Real network effects make the product better for existing users as new ones join, and that's the test worth applying.

Sources

Last verified: 14 October 2026.

Related reads: Two-Sided Marketplace Lessons · WhatsApp: 450 Million Users With a Tiny Team · How Swiggy Built Logistics Without Addresses · Uber System Design

I write about how tech businesses win and why. Follow on LinkedIn or X, or invite me to speak.