Network Effects
A product becomes more valuable as more people use it. One of Peter Thiel’s four characteristics of durable monopoly, and one of the strongest competitive moats a startup can build. But network effects are more nuanced than “more users = more value” — they asymptote, break down, and can even invert.
Types of Network Effects
Direct Network Effects
More users directly increase value for existing users. Each new user adds value to every other user.
- Examples: Phone networks, messaging apps, social networks
- Strength: Very strong when the product IS the network
Two-Sided / Marketplace Effects
Supply and demand reinforce each other. More sellers attract more buyers, which attracts more sellers.
- Examples: Airbnb, Uber, eBay, Etsy
- Challenge: Chicken-and-egg problem at launch — need both sides
Data Network Effects
More users generate more data, which improves the product, which attracts more users.
- Examples: Google Search, Waze, recommendation engines
- Strength: Compounds over time, hard to replicate
Platform / Ecosystem Effects
More developers build on the platform, creating more value for users, attracting more developers.
- Examples: iOS App Store, Shopify, Salesforce AppExchange
- Strength: Creates deep lock-in through ecosystem dependencies
When Network Effects Asymptote
Network effects don’t grow linearly forever:
- Value saturation: In ridesharing, once wait time hits ~5 minutes, more drivers don’t improve the experience. There’s a ceiling.
- Network size limits: Humans have limited trust/attention. Social lending groups >7 people become less effective, not more.
- Value proposition shifts: As Facebook shifted from social networking to media distribution, the original network effect weakened.
Commoditized vs Differentiated Supply
This distinction determines how strong your network effects will be:
| Commoditized | Differentiated | |
|---|---|---|
| Example | Uber drivers, delivery couriers | Airbnb listings, Etsy shops |
| Effect strength | Weaker, shorter-lasting | Stronger, longer-lasting |
| Why | Users indifferent to which provider once liquidity exists | Diversity captures overlapping preferences |
| Moat | Shallow — competitors match at base liquidity | Deep — inventory is unique and non-replicable |
When Network Effects Break Down
- Network overlap: Competitors with similar existing networks enter rapidly (Instagram Stories killed Snapchat’s growth)
- Low switching costs: Easy onboarding reduces stickiness — if you can sign up in 30 seconds, so can a competitor’s users
- Multi-tenanting: Users adopt multiple platforms simultaneously, creating pricing pressure
- Quality degradation: Trolls, spam, and low-quality supply reduce value — curation becomes critical
Building Network Effects
- Start with a tiny niche and achieve critical mass there first (Thiel’s “start small” advice)
- Solve the chicken-and-egg problem by seeding one side (Uber recruited drivers first, Reddit seeded content with fake accounts)
- Use unscalable tactics to bootstrap the network
- Invest in curation and quality control early — network effects are only as strong as the network’s quality
See Also
Sources
Backlinks
- a16z-network-effects
- about-this-wiki
- ai-era-entrepreneurship
- ai-era-entrepreneurship-slides
- balfour-growth-loops
- bootstrapping
- case-study-airbnb
- case-study-shopify
- case-study-slack
- case-study-stitch-fix
- case-study-stripe
- chen-cold-start-problem
- community-building
- distribution
- growth
- growth-loops
- marketplace-dynamics
- moats
- product-development
- product-led-growth
- start-here
- technical-decisions
- where-the-experts-disagree
- where-the-experts-disagree-slides